Available online at www.sciencedirect.

com

Acta Biomaterialia 4 (2008) 808816

www.elsevier.com/locate/actabiomat

A mechanical evaluation of three decellularization methods

in the design of a xenogeneic scaold for tissue engineering
the temporomandibular joint disc
Sarah B. Lumpkins a, Nicolas Pierre b, Peter S. McFetridge b,*
a

Engineering Physics and the School of Chemical, Biological and Materials Engineering, University of Oklahoma,
100 East Boyd Street, Norman, OK 73019-1004, USA
b
University of Oklahoma Bioengineering Center, University of Oklahoma, 100 East Boyd Street, Norman, OK 73019-1004, USA
Received 22 August 2007; received in revised form 11 January 2008; accepted 15 January 2008
Available online 7 February 2008

Abstract
Tissue-engineered temporomandibular joint (TMJ) discs oer a viable treatment option for patients with severe joint internal derangement. To date, only a handful of TMJ tissue engineering studies have been carried out and all have incorporated the use of synthetic
scaold materials. These current scaolds have shown limited success in recapitulating morphological and functional aspects of the
native disc tissue. The present study is the rst to investigate the potential of a xenogeneic scaold for use in tissue engineering the
TMJ disc. The eects of decellularization agents on the discs mechanical properties were assessed using three common decellularization
protocols: Triton X-100, sodium dodecyl sulfate (SDS) and an acetone/ethanol solution. Decellularized scaolds were subsequently characterized through cyclic mechanical testing at physiologically relevant frequencies to determine which chemical agent most accurately
preserved the native tissue properties. Results have shown that porcine discs treated with SDS most closely matched the energy dissipation capabilities and resistance to deformation of the native tissue. Treatments using Triton X-100 caused the resultant tissue to become
relatively softer with inferior energy dissipation capabilities, while treatment using acetone/ethanol led to a signicantly stier and dehydrated material. These ndings support the potential of a porcine-derived scaold decellularized by SDS as a xenograft for TMJ disc
reconstruction.
2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Keywords: Tissue engineering; Temporomandibular joint; Xenogeneic scaold; Mechanical properties; Hysteresis

1. Introduction
The temporomandibular joint is a diarthroidal joint that
joins the mandible, or lower jaw, to the temporal bone of
the skull. One of the most frequently used joints in the
body, the TMJ is also often considered the most complex
joint [1]. The vast complexity of the TMJ yields itself to a
wide variety of pathologies, collectively termed temporomandibular disorders (TMDs). Epidemiological studies
have reported that as many as 2025% of the population

Corresponding author. Tel.: +1 405 325 7193; fax: +1 405 325 5813.
E-mail address: pmcfetridge@ou.edu (P.S. McFetridge).

experience symptoms of TMDs, which can range from

painless clicking and locking of the jaw to debilitating pain
and dysfunction [2]. Of these conditions, 70% of patients
seeking treatment for TMDs suer from a form of internal
derangement in which the brocartilaginous disc exhibits
an abnormal relationship to the condyle, fossa and articular eminence [3]. However, despite the large number of
TMJ related conditions, the TMJ remains one of the least
studied joints in the body.
Current treatment methods for TMDs are widely varied
and are generally regarded as experimental among surgeons due to marginal success rates [4,5]. For example,
alloplastic disc replacements such as ProplastTeon and
Silastic implants were popular in the 1970s and 1980s.

1742-7061/$ - see front matter 2008 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.actbio.2008.01.016

S.B. Lumpkins et al. / Acta Biomaterialia 4 (2008) 808816

However, by the 1990s the literature was replete with

reports of detrimental eects of these implants. Namely,
the implants were prone to fragmentation and tearing,
which led to complications such as bone resorption and
osteoarthritis. Although the FDA issued an advisory to
surgeons to discontinue their use, many patients still suer
from complications caused by the implants [6].
In light of this, the eld of tissue engineering oers a
unique alternative to treating cases of internal derangement
and other complications when the TMJ disc is damaged
beyond repair. Tissue engineering of the TMJ disc is still
in its infancy, however, with as few as 10 comprehensive
studies present in the literature. One of the issues raised
at the worlds rst TMJ Bioengineering Conference in
2006 was that of a need to develop a suitable scaold material for engineering the disc [6]. Suggested bioscaolds have
included photopolymers [7] and alginate hydrogels [8], but
successfully duplicating the complex morphology of the
TMJ disc as well as its widely variant mechanical properties may require other alternatives such as xenogeneic
extracellular matrix scaolds. Unlike scaolds comprised
of synthetic materials, xenogeneic scaolds are inherently
suited to facilitate cell adhesion and native remodeling processes. Promising results with tissue xenografts have been
reported in several animal and clinical studies. For example, scaolds derived from porcine urinary bladder submucosa and small intestinal submucosa have been used as
xenografts to reconstruct musculoskeletal structures, cardiovascular tissues and skin [911].
It is known that the method of decellularization, which
is an important step to reduce the immune impact of allogenic/xenogeneic scaolds, can dramatically alter the
mechanical properties of the resulting scaold [11]. As
such, it is the aim of these investigations to demonstrate
the potential of a porcine TMJ disc xenograft as a scaold
for engineering disc replacements. Accordingly, we have
examined the eects of three decellularization methods on
the biomechanical integrity and energy dissipating capabilities of the derived scaolds. Two common surfactantbased decellularization protocols were chosen for decellularization (sodium dodecyl sulfate (SDS) and Triton X100) as well as an alcohol-based solution comprising 25%
acetone and 75% ethanol. Such alcohol-based solutions

809

have been used previously, albeit in combination with

other chemical treatments [12]. Comparative analysis
between the resultant acellular scaolds has been investigated to assess the feasibility of using one of these decellularization methods to create a TMJ disc substitute.
2. Materials and methods
2.1. Specimen preparation
Fresh TMJ discs were collected from the jaws of pigs
ages 69 months through Animal Technologies Inc. (Tyler,
TX). Dissection of the jaws took place under sterile conditions in a laminar ow hood and required 2030 min per
jaw, as shown in Fig. 1. After rst separating the upper and
lower jaw halves to reveal the joint capsule, the superior
surfaces of the discs were exposed by delicately removing
the connective tissue attaching them to the temporal bone.
Ligaments connecting the disc to the condyle were then
severed using a scalpel to make the nal disc separations.
Any remaining retrodiscal tissue attached to the posterior
disc band was carefully cut away. Discs with perforations
or signs of disc injury were discarded. Baseline data were
then taken o the initial disc dimensions in both the anteroposterior and mediolateral directions using a digital
caliper.
2.2. Decellularization process
The porcine discs were divided into three sample sets of
six TMJ discs each and each set was immersed in 250 ml of
the appropriate decellularization solution: 1% (w/v) SDS,
1% Triton X-100, or a 25% acetone/75% ethanol (vol.%)
combination. The sealed 500 ml capacity glass bottles containing the discs and the three solutions were agitated at
100 rpm. on a horizontal shaker plate for 24 h at 25 C.
After decanting the decellularization solutions, the discs
were washed ve times for 1 h each in phosphate-buered
saline (PBS) to remove the residual chemicals. Measurements of both the anteroposterior and mediolateral disc
dimensions were recorded and subsequently compared to
those of the freshly dissected discs. The discs were then
stored in 0.15 M PBS (pH 7.4) at 4 C until use.

Fig. 1. TMJ disc dissection process. Porcine jaws were separated into the upper and lower joint compartments using a scalpel (A). Connective tissue
attaching the disc to the temporal bone was removed, revealing the disc connections to the condyle (B). Ligaments connecting the disc to the condyle were
severed to separate the disc (C).

810

S.B. Lumpkins et al. / Acta Biomaterialia 4 (2008) 808816

2.3. Mechanical analysis

A dynamic mechanical analyzer (Rheometric Solids
Analyzer II (RSA); Rheometric Scientic, Piscataway,
NJ) was used for all the mechanical testing of the decellularized discs and the control group of six native discs that
did not undergo decellularization. The test setup was modied by adding a chamber lled with PBS (0.15 M, pH 7.4)
to the bottom compression plate (see Fig. 2). This conguration allowed samples to be tested in a hydrated environment analogous to the in vivo environment, since the native
TMJ disc is enclosed in a joint capsule lled with synovial
uid [13]. This is also in accordance with other studies on
TMJ disc mechanics which have utilized PBS testing environments [14,15]. Samples were prepared from the central
section of each decellularized disc using a 4.0 mm diameter
tissue punch (Fig. 3). The central region was chosen for
these analyses, as mechanical testing and nite-element
modeling have shown that this region is subjected to the
most loading during normal physiological function
[16,17]. Before testing, the inferior surface of each sample
was dried and glued using cyanoacrylate adhesive to the
center of the 5 mm diameter bottom compression plate.
This step was necessary to prevent slippage of the disc
due to incurred buoyant forces during testing. Additionally, the native disc in vivo is prevented from slipping via
several collateral ligaments which attach it to the medial
and lateral poles of the condyle [13]. An in vivo simulated

Fig. 2. Cyclic loading test chamber. Schematic drawing of the chamber

developed to maintain a hydrated test environment. A load transducer is
used to measure the deformation force applied to the sample (A). Top and
bottom compression plates, each 5 mm in diameter (B), were used to
deform the sample mounted to the bottom compression plate (C). The
discs were submerged in PBS solution which lled the test chamber (D).
Note that this gure is not to scale; the disc samples were 4.0 mm in
diameter and therefore smaller than the compression plates.

Fig. 3. Sample preparation for mechanical testing. The TMJ disc is

shown, along with the labeled mediolateral and anteroposterior orientations. The retrodiscal tissue is attached to the posterior section of the disc.
A tissue punch with a 4.0 mm inside diameter was used to extract samples
from the central disc region (labeled C) for mechanical testing before
and after decellularization.

temperature of 37 C was achieved using the environmental

control subsystem of the RSA II.
2.4. Stressstrain/cyclic loading
Before testing, samples were allowed to equilibrate
unloaded in PBS for 5 min at 37 1 C. A digital caliper
was used to measure the thickness of each sample in four
locations (three approximately equally spaced around the
periphery and one in the center) and the average thicknesses were used as inputs in the resulting stressstrain
analysis.
In vivo, the TMJ disc is subjected to several loading
modalities, but under healthy conditions the most relevant
is that of dynamic or cyclic compressive loading. Therefore,
testing of the porcine tissue was carried out based on this
loading modality. For all of the cyclic loading tests conducted, deformations of the samples were dened by the
strain e = DL/L0, where DL is the decrease in sample thickness relative to the initial thickness, L0 The resulting stress
on each sample is dened by r = F/A, where F is the compressive force and A is the cross-sectional area of the
sample.
In order to compare the time-dependent tissue behavior
between each type of decellularized sample set, each cyclic
loading test consisted of 15 compression cycles. A compressive strain of 10% was chosen for all samples in accordance
to the maximum amount of joint space reduction during
maximum clenching [18]. Additionally, each sample was
tested under four levels of loading frequency, 0.1, 0.5, 1
and 1.5 Hz, which correspond to the chewing frequency
of humans [19]. Samples were compressed under a sinusoidal strain dened by e = De sin(xt), where De and x are the
respective strain amplitude and frequency. A trial experiment was conducted under each amplitudefrequency combination and it was found that a recovery time of 5 min was
necessary to obtain reproducible stressstrain curves. Upon

S.B. Lumpkins et al. / Acta Biomaterialia 4 (2008) 808816

completion of all frequency test cycles for a particular sample, a control test was conducted at the initial conditions
(10% strain and 0.1 Hz) to ensure that testing order had
no eect on any of the stressstrain relationships.
Results of the cyclic loading tests were analyzed by calculating both the instantaneous and steady-state tangent
moduli, Eint and Ess, which correspond to the peak stresses
at the rst and last cycle relative to the respective strain
amplitudes.
2.5. Hysteresis analysis
Hysteresis relates to the energy dissipating capabilities
of a particular material and is a fundamental property of
all viscoelastic tissues, including the TMJ disc. In order
to determine how the decellularization processes aected
the resulting porcine discs, integrals were carried out over
each hysteresis loop and were recorded as the amount of
energy dissipation per unit volume of material.

811

variances were not equal across groups. Paired t-tests were

used to assess changes in disc dimensions within individual
sample groups before and after decellularization.
3. Results
3.1. Sample morphology
Prior to mechanical testing, disc thicknesses for the Triton X-100, SDS and acetone/ethanol samples were: 2.22
0.20 mm, 2.13 0.18 mm and 2.13 0.11 mm, respectively. Additionally, a digital caliper was used to measure
each disc before and after decellularization in both the
anteroposterior and mediolateral dimensions (Fig. 3). Data
shown in Fig. 4 displayed no statistically signicant dierences in the initial and nal dimensions of the discs treated
with SDS or Triton X-100. However, the samples treated
with acetone/ethanol decreased signicantly in both the
anteroposterior (P < 0.001) and mediolateral (P < 0.005)
directions.

2.6. Histology analysis

3.2. Modulus analysis
Histology was performed on the central sections of discs
prepared using each of the three decellularization methods
as well as the native disc (control) to assess ECM structural
changes. Samples prepared for sectioning were xed in 10%
formal saline for 24 h before proceeding to dehydration
with graded concentrations of ethanol. After paran
embedding and sectioning at 7 lm, the samples were subjected to standard hematoxylin and eosin staining (7 min
and 15 s, respectively) and viewed under a microscope.

The relative stiness of each disc was measured during

both the initial and nal hysteresis loops and these data
correspond to the modulus values shown in Fig. 5.
ANOVA revealed no signicant dierences between

2.7. SEM analysis

In order to determine the eects of the three decellularization methods on the surface structure of the TMJ discs,
four samples underwent SEM analysis. Samples were prepared using a 5 mm stainless-steel tissue punch to extract
tissue from the central regions of each disc. Surfaces of
the discs were distinguished by pulling a needle with a different thread color through the samples in the superior to
inferior direction. The thread ends were knotted to identify
the superior surfaces.
2.8. Statistical analysis
Each set of decellularization data was calculated based
on testing of six samples (n = 6) in comparison also to a
control group (n = 6). SPSS software was used for all statistical analysis. One-way analysis of variance (ANOVA)
testing was used to determine statistical signicance
between strain frequency and modulus values for tissues
decellularized using each of the three methods. One-way
ANOVA was also carried out to determine signicant differences between peak stress, peak hysteresis and hysteresis
percentage of each decellularization method, with a Tamhanes T2-test being included as a correction factor when

Fig. 4. Disc dimensions. The porcine discs were measured in both the
anteroposterior (a) and mediolateral (b) directions before and after each of
the respective decellularization treatments. Results are signicantly
dierent only between the samples treated with acetone/ethanol, showing
that this treatment had a dehydrating eect on the tissue samples (
P < 0.005;  P < 0.001).

812

S.B. Lumpkins et al. / Acta Biomaterialia 4 (2008) 808816

Fig. 6. Representative hysteresis loops. Hysteresis loops during the rst

loading cycle for all three decellularized tissue types as well as the native
porcine tissue. The acetone/ethanol and Triton X-100 samples displayed
the largest and smallest degrees of hysteresis, respectively.

Fig. 5. Modulus values for the central disc sections. Instantaneous and
steady-state modulus values are shown in (a) and (b), respectively. Values
are reported for testing under compressive strains of 10% and frequencies
of 0.1, 0.5, 1 and 1.5 Hz. Frequency dependence on moduli: the Triton X100 and acetone/ethanol samples showed no frequency dependence on
Eint, whereas there was a frequency dependence for both the native
(P < 0.0001) and SDS (P < 0.05) tissues. Additionally, there was also a
frequency dependence on Ess for the native (P < 0.0001), SDS (P < 0.05)
and Triton X-100 (P < 0.05) samples. Decellularization dependence on
moduli: for each frequency, signicance levels are shown as: P < 0.005 ()
and P < 0.0001 () compared to native tissue. Note that at each
frequency, modulus values for Triton X-100 and acetone/ethanol tissues
are signicantly smaller and larger, respectively, than any of the other
tissues (data not shown).

applied frequency and instantaneous modulus values (Eint)

for the Triton X-100 and acetone/ethanol samples. There
was, however, an overall frequency dependence for both
the native (P < 0.0001) and SDS (P < 0.05) tissues. There
was also a frequency dependence on the steady-state modulus (Ess) for the native (P < 0.0001), SDS (P < 0.05) and
Triton X-100 (P < 0.05) samples.
Signicant dierences also existed between decellularization method and resulting modulus values (both Eint and
Ess). Specically, the acetone/ethanol and Triton X-100
samples had signicantly larger and smaller moduli than
the native samples, respectively, for all frequencies tested.
The modulus values for the SDS and native samples were
not signicantly dierent, with the exception of Eint at frequencies of 0.1 (P < 0.005) and 1 Hz (P < 0.0001).

treated with Triton X-100 displayed the smallest peak stresses and hysteresis loops, corresponding to the least amount
of energy dissipation. Comparative peak stresses during the
rst compression cycle are shown in Fig. 7 and peak hysteresis values are shown in Fig. 8. Although the degree of hysteresis did vary widely between samples, Fig. 9 shows that
the percentage of hysteresis of the nal compared to the initial cycles for each sample set was similar, with no statistically signicant dierences.
3.4. Histology analysis
Histological sections of the native and decellularized
disc tissues are shown in Fig. 10. Control sections stained
with hematoxylin revealed the presence of brochondrocytes in the native tissue, with the decellularized tissues
showing no sign of cell nuclei.
3.5. SEM analysis
SEM images were taken of both the native and decellularized disc surfaces (Fig. 11). The native tissue displayed
relatively small collagen bers in comparison with those

3.3. Hysteresis
All the samples tested displayed clear hysteresis loops,
although the size and shape of the loops varied between
decellularization methods. Representative hysteresis loops
corresponding to the rst loading cycle are shown in
Fig. 6. Samples treated with acetone/ethanol exhibited
the largest hysteresis loops and peak stresses, while those

Fig. 7. Peak stresses. Maximum stress values for each tissue type during
the initial compression cycle. Asterisks show signicance of dierence in
the peak stresses ( P < 0.0001;  P < 0.05) compared with those of the
acetone/ethanol tissue.

S.B. Lumpkins et al. / Acta Biomaterialia 4 (2008) 808816

813

of the decellularized tissues and they were oriented primarily in the anteroposterior direction along with smaller bers
oriented in oblique directions. The bers appeared
crimped, or wavy, in accordance with a SEM study by
Minarelli et al. of the central disc region [20].
The decellularized tissues generally retained the native
collagen ber organization, although the bers appeared
to be compressed into larger bundles not evident in the
control set. Additionally, the collagen bundles of the acetone/ethanol samples appeared more attened than those
in the Triton X-100 and SDS samples.
Fig. 8. Peak hysteresis. The hysteresis values during the rst compression
cycle of each decellularized tissue and the native tissue are shown.
Signicant dierences in hysteresis are also presented as follows: 
P < 0.05 compared to acetone/ethanol and Triton X-100;  P < 0.01
compared to Triton X-100.

Fig. 9. Hysteresis analysis. Mean values of hysteresis in the nal

compression cycle as compared to the initial cycle for each type of
decellularized tissue as well as the native tissue. Results are not
signicantly dierent between samples.

4. Discussion
One of the primary challenges in using xenogeneic scaffolds for TMJ disc tissue engineering lies in choosing an
appropriate animal model. The TMJ shows remarkable
functional and morphological variation among species
and TMJ characterization studies have been carried out
on a range of animals including: rabbits [21,22], dogs
[23], rats [24,25], sheep [26] and pigs [14,27,28]. While each
species has been found to have distinctive TMJ adaptations, the pig in general is the only animal studied to show
similar levels of loading to that of the human jaw [29].
Other similarities favoring the selection of the pig animal
model include the size of the articular structures, disc morphology and the fact that pigs are omnivorous [30]. One of
the few dierences in jaw function lies in chewing frequency; pigs chew slightly faster (2.03.0 Hz) than humans
(0.11.5 Hz) [19]. Therefore, due to the many similarities in
jaw morphology and function, the pig is generally regarded

Fig. 10. Histologic images. Hematoxylin and eosin stained images of the native TMJ disc (A) in addition to the tissues decellularized using Triton X-100
(B), SDS (C) and acetone/ethanol (D) showing complete cell removal.

814

S.B. Lumpkins et al. / Acta Biomaterialia 4 (2008) 808816

Fig. 11. SEM images of superior TMJ disc surface. All images are taken from the central regions of the disc. The native tissue is shown (a), along with the
tissues decellularized using Triton X-100 (b), SDS (c) and acetone/ethanol (d). Collagen bundles in (b) and (c) are aligned primarily in a single direction as
in the native tissue, while those in (d) are more disorganized. Arrows indicate the anteroposterior direction.

as the most appropriate animal model for TMJ disc studies

and was the basis for the current investigation.
After a suitable animal model has been chosen, an
appropriate decellularization method must be found to
prepare the xenogeneic scaolds for either direct implantation or, as in the tissue engineering methodology, cell seeding. The goal of any decellularization protocol is to
eectively remove immunogenic cellular material while
maintaining the biological activity and mechanical integrity
of the extracellular matrix. Several methods of decellularization have been cited in the literature, including physical,
enzymatic and chemical treatments. Studies have shown
that the eciency of a particular decellularization protocol
is largely dependent upon the tissue of interest [31].
Because there have been no previous studies to determine
which method is most suitable for decellularizing the
TMJ disc, the present investigation analyzed the eects of
three common chemical agents on the discs.
Triton X-100 was selected as a method of decellularization due to its wide use as a non-ionic surfactant for
decellularization purposes. In general, non-ionic surfactants are used extensively in decellularization protocols
due to their mild eects on tissue structure; they disrupt

lipidprotein and lipidlipid interactions, but generally

leave proteinprotein reactions intact with the result of
maintaining their functional conformations [31]. However,
mixed results on the ecacy of this chemical treatment
have been reported. Regarding extracellular matrix components, Triton X-100 has led to a signicant loss of glycosaminoglycans (GAGs) in heart valve tissues [32],
whereas it has left GAG components in the anterior cruciate ligament unchanged [33]. Alterations in mechanical
integrity of tissue structures as a result of Triton X-100
have also been reported [33,34].
The results of the present study indicate that Triton X100 may not be the most eective decellularization method
for the TMJ disc. Although the disc dimensions were preserved compared to controls, the resulting mechanical
integrity of these discs was found to be far inferior to that
of the native tissue. Based on hysteresis analysis, the energy
dissipating capabilities of the Triton X-100 discs were
approximately one-third that of the native tissue. This
result also implies that these samples have inhibited
shock-absorbing capabilities leading to an inferior disc
construct, as has been previously hypothesized that the
shock-absorbing capabilities of the disc are advantageous

S.B. Lumpkins et al. / Acta Biomaterialia 4 (2008) 808816

during sudden impacts to the joint when rapid energy dissipation is crucial [17,28]. Additionally, the instantaneous
and steady-state modulus values, which represent the ability of the discs to resist deformation, were less than half of
those of the native tissue.
The second set of TMJ disc samples were decellularized
using a 1% SDS treatment. SDS is a commonly used ionic
detergent, and these are generally eective for solubilizing
both nuclear and cytoplasmic cellular membranes. However, results in the literature have also been discrepant
regarding the eciency of this class of chemical agents. A
study by Seddon et al. reported that ionic detergents tend
to denature proteins by disrupting proteinprotein interactions [35]. In contrast, a study by Schaner et al. concluded
that SDS treatment is a feasible option for vascular tissue
engineering eorts among several detergents studied; cell
extraction was eectively achieved without signicant disturbances in extracellular matrix morphology and strength
[36]. Results of the present study also support SDS as a viable option for treating the TMJ disc. The sizes of the discs
were preserved after decellularization and the modulus values were not signicantly dierent between the native tissue
and SDS-treated tissue (except for the instantaneous modulus at frequencies of 0.1 and 1 Hz). Both the peak stress
and peak hysteresis values also showed no statistically signicant dierences.
Additionally, the SDS tissue exhibited similar frequency
dependence on modulus values compared to the native tissue; the modulus values of both tissues increased with
increasing frequency. This trend has been experimentally
veried in studies regarding interstitial uid pressurization
during cyclic loading of articular cartilage [37]. Specically,
the total stress imparted by cartilage during loading is a
combination of the elastic stress due to deformation of
the solid matrix and interstitial uid pressure, the latter
of which has been shown to support greater loads [37].
Therefore, as uid pressure becomes increasingly dominant
(as load frequency increases), the resulting stresses and
modulus values become larger. Threshold frequencies at
which the transition occurs between loading regimes are
highly dependent on material properties. This threshold
may have been made suciently larger for the discs that
underwent acetone/ethanol and Triton X-100 treatments
such that interstitial uid pressure no longer dominated
the total stress imparted by the disc and, consequently,
modulus dierences could not be detected within the particular frequency regime used.
The third set of samples was prepared using a solution
of 25% acetone and 75% ethanol. This treatment method
appeared to have a dehydrating eect since the tissue
became signicantly shortened in both the mediolateral
(P < 0.005) and anteroposterior (P < 0.001) directions.
The result that the tissues were shortened more signicantly in the anteroposterior direction may be related to
ber orientation, as SEM studies have revealed that collagen bers exhibit a strong anteroposterior alignment
through the central disc regions [20].

815

Although the tissue was more ecient in dissipating

energy, as can be seen by large peak hysteresis values, the
alcohol treatment also resulted in a much stier material
with modulus values approximately three times greater
than those of the native tissue. SEM analysis also revealed
that the collagen bundles may have become more disorganized and attened in these tissue samples. Due to the signicant alteration of the mechanical properties and
morphology of the disc, decellularization using acetone/
ethanol is the least attractive option for scaold preparation of those presented in the current investigation.
Methods for attaching the tissue-engineered discs into
the joint space will be an important area for future investigation. In the past, surgeries ranging from simple repositioning of the disc during open joint surgery up
through discectomies have resulted in a wide variety of
techniques to reattach discs. Plication procedures were
popular in the 1980s and 1990s and involve the suturing
of some portions of the disc to either surrounding joint
tissues or to the condyle itself [38]. Modern techniques
of reattaching discs, particularly in the case of disc-repositioning surgeries, include the use of Mitek mini anchors
which are drilled into place in the condyle and serve as
attachment points for two sutures passing through the
disc [39,40]. Most recently, bioresorbable screws have
been utilized to attach the disc directly to the condylar
head [41]. Promising results have been seen during
short-term follow-ups of these latter procedures which
will likely be candidates for attaching future tissue-engineered discs.
5. Conclusions
The choice of scaold material is a crucial step towards
engineering a suitable TMJ disc replacement. This investigation is the rst to introduce the possibility of using a
xenogeneic extracellular matrix scaold as an alternative
to the synthetic materials currently being explored. In
light of the many protocols used in the literature to
remove cellular content from ex vivo tissues, the current
study investigated three popular chemical treatments in
an eort to determine which decellularization method
was most eective in preserving the mechanical integrity
of the native tissue. The results show that SDS was more
eective than either the Triton X-100 or acetone/ethanol
methods in the case of the TMJ disc. The SDS-treated
discs maintained the same range of energy dissipation
characteristics and modulus values as the native porcine
tissue and the general morphology of the discs was preserved, with the exception of the collagen bers being
slightly more compressed. Further studies regarding the
potential of this scaold material for engineering the disc,
including cell seeding on the acellular scaolds, should be
explored in the future. Overall, these results support the
potential of a porcine acellular scaold for the engineering and reconstruction of damaged or diseased TMJ
discs.

816

S.B. Lumpkins et al. / Acta Biomaterialia 4 (2008) 808816

References
[1] Haskin CL, Milam SB, Cameron IL. Pathogenesis of degenerative
joint disease in the human temporomandibular joint. Crit Rev Oral
Biol Med 1995;6(3):24877.
[2] Solberg WK, Woo MW, Houston JB. Prevalence of mandibular
dysfunction in young adults. J Am Dent Assoc 1979;98(1):2534.
[3] Farrar WB, McCarty Jr WL. The TMJ dilemma. J Ala Dent Assoc
1979;63(1):1926.
[4] Mercuri LG, Wolford LM, Sanders B, White RD, Giobbie-Hurder A.
Long-term follow-up of the CAD/CAM patient tted total temporomandibular joint reconstruction system. J Oral Maxillofac Surg
2000;60(12):14408.
[5] Mercuri LG. The use of alloplastic prostheses for temporomandibular
joint reconstruction. J Oral Maxillofac Surg 2000;58(1):705.
[6] Detamore MS, Athanasiou KA, Mao J. A call to action for
bioengineers and dental professionals: directives for the future of
TMJ bioengineering. Ann Biomed Eng 2007;35(8):130111.
[7] Poshusta AK, Anseth KS. Photopolymerized biomaterials for application in the temporomandibular joint. Cells Tissues Organs
2001;169(3):2728.
[8] Almarza AJ, Athanasiou KA. Seeding techniques and scaolding
choice for tissue engineering of the temporomandibular joint disk.
Tissue Eng 2004;10(11-12):178795.
[9] Badylak S et al. Naturally occurring extracellular matrix as a scaold
for musculoskeletal repair. Clin Orthop Relat Res 1999(Suppl. 367):
S33343.
[10] Badylak SF. The extracellular matrix as a scaold for tissue
reconstruction. Semin Cell Dev Biol 2002;13(5):37783.
[11] Badylak SF. Xenogeneic extracellular matrix as a scaold for tissue
reconstruction. Transpl Immunol 2004;12(3-4):36777.
[12] McFetridge PS, Daniel JW, Bodamyali T, Horrocks M, Chaudhuri
JB. Preparation of porcine carotid arteries for vascular tissue
engineering applications. J Biomed Mater Res A 2004;70(2):22434.
[13] Piette E. Anatomy of the human temporomandibular joint. An
updated comprehensive review. Acta Stomatol Belg 1993;90(2):10327.
[14] Allen KD, Athanasiou KA. Viscoelastic characterization of the
porcine temporomandibular joint disc under unconned compression. J Biomech 2006;39(2):31222.
[15] Tanaka E et al. Impulsive compression inuences the viscous
behavior of porcine temporomandibular joint disc. Eur J Oral Sci
2003;111(4):3538.
[16] Perez del Palomar A, Doblare M. The eect of collagen reinforcement
in the behaviour of the temporomandibular joint disc. J Biomech
2006;39(6):107585.
[17] Beek M, Aarnts MP, Koolstra JH, Feilzer AJ, van Eijden TM.
Dynamic properties of the human temporomandibular joint disc. J
Dent Res 2001;80(3):87680.
[18] Kim KW, Wong ME, Helfrick JF, Thomas JB, Athanasiou KA.
Biomechanical tissue characterization of the superior joint space of
the porcine temporomandibular joint. Ann Biomed Eng 2003;31(8):
92430.
[19] Druzinsky RE. The time allometry of mammalian chewing movements: chewing frequency scales with body mass in mammals. J Theor
Biol 1993;160(4):42740.
[20] Minarelli AM, Del Santo Junior M, Liberti EA. The structure of the
human temporomandibular joint disc: a scanning electron microscopy study. J Orofac Pain 1997;11(2):95100.
[21] Weijs WA, de Jongh HJ. Strain in mandibular alveolar bone during
mastication in the rabbit. Arch Oral Biol 1977;22(12):66775.

[22] Widmer CG, English AW, Carrasco DI, Malick CL. Modeling rabbit
temporomandibular joint torques during a power stroke. Angle
Orthod 2002;72(4):3317.
[23] Scapino RP. The third joint of the canine jaw. J Morphol
1965;116:2350.
[24] Byrd KE. Opto-electronic analyses of masticatory mandibular
movements and velocities in the rat. Arch Oral Biol 1988;33(3):
20915.
[25] Weijs WA, Dantuma R. Electromyography and mechanics of
mastication in the albino rat. J Morphol 1975;146(1):133.
[26] Thomason JJ, Grovum LE, Deswysen AG, Bignell WW. In vivo
surface strain and stereology of the frontal and maxillary bones of
sheep: implications for the structural design of the mammalian skull.
Anat Rec 2001;264(4):32538.
[27] Detamore MS et al. Cell type and distribution in the porcine
temporomandibular joint disc. J Oral Maxillofac Surg 2006;64(2):
2438.
[28] Tanaka E et al. Dynamic compressive properties of porcine
temporomandibular joint disc. Eur J Oral Sci 2003;111(5):
4349.
[29] Herring SW, Decker JD, Liu ZJ, Ma T. Temporomandibular joint in
miniature pigs: anatomy, cell replication, and relation to loading.
Anat Rec 2002;266(3):15266.
[30] Detamore MS, Athanasiou KA. Motivation, characterization, and
strategy for tissue engineering the temporomandibular joint disc.
Tissue Eng 2003;9(6):106587.
[31] Gilbert TW, Sellaro TL, Badylak SF. Decellularization of tissues and
organs. Biomaterials 2006;27(19):367583.
[32] Grauss RW, Hazekamp MG, Oppenhuizen F, van Munsteren CJ,
Gittenberger-de Groot AC, DeRuiter MC. Histological evaluation of
decellularised porcine aortic valves: matrix changes due to dierent
decellularisation methods. Eur J Cardiothorac Surg 2005;27(4):
56671.
[33] Woods T, Gratzer PF. Eectiveness of three extraction techniques in
the development of a decellularized boneanterior cruciate ligament
bone graft. Biomaterials 2005;26(35):733949.
[34] Cartmell JS, Dunn MG. Eect of chemical treatments on tendon
cellularity and mechanical properties. J Biomed Mater Res
2000;49(1):13440.
[35] Seddon AM, Curnow P, Booth PJ. Membrane proteins, lipids and
detergents: not just a soap opera. Biochim Biophys Acta
2004;1666(1-2):10517.
[36] Schaner PJ et al. Decellularized vein as a potential scaold for
vascular tissue engineering. J Vasc Surg 2004;40(1):14653.
[37] Soltz MA, Ateshian GA. Interstitial uid pressurization during
conned compression cyclical loading of articular cartilage. Ann
Biomed Eng 2000;28(2):1509.
[38] Weinberg S, Cousens G. Meniscocondylar plication: a modied
operation for surgical repositioning of the ectopic temporomandibular joint meniscus. Rationale and operative technique. Oral Surg
Oral Med Oral Pathol 1987;63(4):393402.
[39] Mehra P, Wolford LM. The Mitek mini anchor for TMJ disc
repositioning: surgical technique and results. Int J Oral Maxillofac
Surg 2001;30(6):497503.
[40] Mehra P, Wolford LM. Use of the Mitek anchor in temporomandibular joint disc-repositioning surgery. Proc (Bayl Univ Med Cent)
2001;14(1):226.
[41] Sembronio S, Robiony M, Politi M. Disc-repositioning surgery of the
temporomandibular joint using bioresorbable screws. Int J Oral
Maxillofac Surg 2006;35(12):114952.