Matrix Biology 23 (2005) 543 – 555
www.elsevier.com/locate/matbio
Intrinsic fibroblast-mediated remodeling of damaged collagenous
matrices in vivo
Paolo P. Provenzanoa,b,*, Adriana L. Alejandro-Osoriob,c, Wilmot B. Valhmub,
Kristina T. Jensena,b, Ray Vanderby Jr.a,b
a
Department of Biomedical Engineering, University of Wisconsin, Madison, WI, United States
Department of Orthopedics and Rehabilitation, University of Wisconsin, Madison, WI, United States
c
Department of Biomolecular Chemistry, University of Wisconsin, Madison, WI, United States
b
Received 1 July 2004; received in revised form 22 September 2004; accepted 23 September 2004
Abstract
Numerous studies have examined wound healing and tissue repair after a complete tissue rupture and reported provisional matrix and scar
tissue formation in the injury gap. The initial phases of the repair are largely mediated by the coagulation response and a principally extrinsic
inflammatory response followed by type III collagen deposition to form scar tissue that may be later remodeled. In this study, we examine
subfailure (Grade II sprain) damage to collagenous matrices in which no gross tissue gap is present and a localized concentration of
provisional matrix or scar tissue does not form. This results in extracellular matrix remodeling that relies heavily upon type I collagen, and
associated proteoglycans, and less heavily on type III scar tissue collagen. For instance, following subfailure tissue damage, collagen I and III
expression was suppressed after 1 day, but by day 7 expression of both genes was significantly increased over controls, with collagen I
expression significantly larger than type III expression. Concurrent with increased collagen expression were significantly increased
expression of the collagen fibrillogenesis supporting proteoglycans fibromodulin, lumican, decorin, the large aggregating proteoglycan
versican, and proteases cathepsin K and L. Interestingly, this remodeling process appears intrinsic with little or no inflammation response as
damaged tissues show no changes in macrophage or neutrophils levels following injury and expression of the inflammatory markers, tumor
necrosis factor-a and tartrate-resistant acid phosphatase were unchanged. Hence, since inflammation plays a large role in wound healing by
inducing cell migration and proliferation, and controlling extracellular matrix scar formation, its absence leaves fibroblasts to principally
direct tissue remodeling. Therefore, following a Grade II subfailure injury to the collagen matrix, we conclude that tissue remodeling is
fibroblast-mediated and occurs without scar tissue formation, but instead with type I collagen fibrillogenesis to repair the tissue. As such, this
system provides unique insight into acute tissue damage and offers a potentially powerful model to examine fibroblast behavior.
D 2004 Elsevier B.V./International Society of Matrix Biology. All rights reserved.
Keywords: Subfailure damage; Sprain; Scar tissue formation; Inflammation; Medial collateral ligament; Real-time quantitative polymerase chain reaction (realtime QPCR)
1. Introduction
Ligament sprains occur more often than complete tissue
rupture and can be chronically debilitating, frequently
causing persistent joint instability, prolonged pain, and
* Corresponding author. Present address: Department of Pharmacology, 3619 MSC, 1300 University Avenue, Madison, WI 53706, United
States. Tel.: +1 608 265 5094.
E-mail address: [email protected] (P.P. Provenzano).
progressive joint degeneration. Although numerous studies
have examined Grade III1 ligament injuries, further study is
1
Grades I, II, and III tissue injuries are described as follows (Andriachi
et al., 1987; Scuderi and Scott, 2001): Grade I sprains are mild stretches
with no discontinuity of the ligament and no clinically detectable increase
in joint laxity. Grade II sprains are moderate stretches in which some
collagen fibers are torn. Enough fibers remain intact so that the ligament has
not completely failed, but detectable abnormal laxity is present at the joint.
Grade III sprains are severe and consist of a complete or nearly complete
ligament disruption and result in significant joint laxity.
0945-053X/$ - see front matter D 2004 Elsevier B.V./International Society of Matrix Biology. All rights reserved.
doi:10.1016/j.matbio.2004.09.008
544
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
Fig. 1. Average stress–strain curves for sham control and subfailure
stretched tissue (meanFS.D.). The averaged data were fit with a microstructural model (Hurschler et al., 1997) to illustrate the shape of the stress–
strain curve. Note that matrix damage due to subfailure stretching increases
the strain-stiffening region of the stress–strain curve and results in
decreased elastic modulus, indicating a significant change in the material
properties of the collagenous matrix.
required to describe the mechanics and biology of the
collagenous extracellular matrix after a subfailure damage
(sprain) injury.
A prospective study of patients with partial tears of the
anterior cruciate ligament (ACL) revealed clinically unstable knees in over 50% of the patients after a mean follow-up
of 17 months (Fruensgaard and Johannsen, 1989). Such
instabilities are likely caused by increased deformations
within the physiologic strain-stiffening region of the loaddeformation (Panjabi et al., 1996) and stress–strain curves
(Provenzano et al., 2002b), and laxity (Provenzano et al.,
2002a) of ligaments after a subfailure injury. In collateral
ligaments, the initial increase in tissue laxity is likely caused
by collagen fiber damage and/or rupture (Provenzano et al.,
2002a; Yahia et al., 1990) to fibers that are first recruited
during tissue loading (Viidik, 1972). In addition, the initial
injury to the tissue results in substantial cellular damage
(Provenzano et al., 2002b) followed by apoptosis (Provenzano et al., 2002c). Hence, it is reasonable to expect a
substantial healing response to remove cell and tissue debris
and repair the damaged extracellular matrix. Yet this repair
is imperfect. Although tissue strength increases with healing
time after a subfailure injury, tissue laxity persists (Provenzano et al., 2002a). An in vivo study of rat ligament healing
showed a significant increase in tissue ultimate stress (to
80% of control) 2 weeks after a subfailure injury; yet there
was no significant improvement in the tissue laxity over 2
weeks (Provenzano et al., 2002a). Further studies of the
causal mechanisms of tissue damage (and the resulting
laxity) as well as the in vivo healing response after
subfailure injury are required.
After a Grade III injury (complete rupture) to ligament, a
region of scar tissue is deposited to provide a provisional
matrix for connecting the two residual ends of the ruptured
tissue (Frank et al., 1983; Jack, 1950). Biochemical analysis
of this scar tissue reveals changes in total collagen and
glycosaminoglycan content within the scar tissue out to 40
weeks post-injury (Frank et al., 1983). More specific
analysis of collagen reveals a strong increase in type III
collagen (Frank et al., 1983; Williams et al., 1984), which
constitutes the majority of newly synthesized collagen in
scar tissue (Williams et al., 1984). These results are in
agreement with gene expression analysis of Grade III
injured ligaments showing large increases in type III
collagen mRNA, with increases in type III collagen
expression exceeding type I expression increases 3 weeks
post-injury (Boykiw et al., 1998). Hence, Grade III injuries
of collagenous extracellular matrices and subsequent scar
formation have received considerable study. However, less
is known about the biology of healing after a Grade II injury
and the effects of subfailure tissue deformations. It is known
that after a Grade II ligament injury, extracellular matrix
damage appears to be distributed and that a localized scar
formation is not detected macro- or microscopically (Laws
and Walton, 1988; Provenzano et al., 2002a,b). This
information, combined with reports of a potential intrinsic,
non-inflammatory, fibroblastic-mediated healing response in
subfailure damaged tissue (Laws and Walton, 1988),
indicate further study of subfailure tissue deformation and
Fig. 2. Ultimate stress (A) and elastic modulus (B) values after a subfailure
ligament injury are significantly decreased ( p=0.03 and p=0.026,
respectively; meanFS.E.M.) following subfailure damage injury to the
tissue indicating significant and consistent tissue damage.
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
545
Fig. 3. Introduction of a subfailure damage injury does not produce an increased macrophage response within the ligament tissue (400). In both control and
damaged ligament tissue, the majority of the tissue area was void of macrophages (a, b). However, both tissues also possess very few regions in which
macrophages were present (c, d), with no difference between control and injured tissues. Alternatively, the scar tissue surrounding the ligament (resulting from
the surgical insult) is strongly positive for macrophage infiltration in both control and damaged tissue (e, f: boundary between ligament (L) tissue and
surrounding scar tissue (S); g, h: macrophages in surgically induced scar tissue surrounding the ligament). Furthermore, macrophages are present near the
bone–ligament insertion (i, j).
546
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
Fig. 4. Introduction of a subfailure damage injury does not produce an increased neutrophil response within the ligament tissue (400). In both control and
damaged ligament tissue, the majority of the tissue area was void of neutrophils (a, b). However, both tissues also possessed very few regions in which
neutrophils were present (c, d), with no difference between control and injured tissues. Alternatively, the scar tissue surrounding the ligament (resulting from
the surgical insult) is strongly positive for neutrophil infiltration in both control and damaged tissue (e, f: boundary between ligament (L) tissue and surrounding
scar tissue (S); g, h: neutrophils in surgically induced scar tissue surrounding the ligament). Furthermore, neutrophils are present near the bone–ligament
insertion (i, j).
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
injury is necessary and may provide unique insight into
fibroblast behavior and the fibroblast–matrix interaction.
The purpose of this study, therefore, was to determine
whether there is an up-regulated inflammatory response
after subfailure injury or whether the remodeling process is
an intrinsic fibroblast-mediated response. Further, gene
expression patterns for extracellular matrix proteins and
matrix-degrading proteinases were examined following a
subfailure injury to elucidate the cellular response to tissue
injury. To address these questions, immunohistochemistry
was utilized to determine whether macrophages and/or
neutrophils were infiltrating the connective tissue after
injury. Real-time quantitative polymerase chain reaction
(QPCR) analysis was employed to determine the expression
levels of collagen, proteoglycan, and proteinase mRNAs at
1, 3, and 7 days post-subfailure damage injury in the rat
medial collateral ligament. In addition, typical mRNA
markers of inflammation were used to lend additional
insight into the intrinsic remodeling hypothesis.
2. Results
Mechanical analysis following the stretch-induced subfailure tissue damage indicated a substantial and consistent
injury. Dramatic changes in the stress–strain behavior of the
tissues were detected (Fig. 1) and ultimate stress ( p=0.03)
and elastic modulus ( p=0.026) values of the subfailure
tissues were significantly decreased to approximately 50%
of the control value (Fig. 2). Strain at failure was not
significantly different. However, tissue length and laxity
values were significantly increased. Initial tissue laxity was
calculated from ligament length values to be 12.9%.
Moreover, the strain at the transition between the nonlinear
strain-stiffening region (dtoe-regionT) and the dlinearT region
of the stress–strain curve was significantly ( p=0.049)
increased in damaged tissues (e TRANS=0.015F0.008 for
control tissues and 0.037F0.007 for damaged tissues).
Immunohistochemical evaluation of damaged ligaments
indicated that injury did not induce an inflammatory
response within the tissue. Macrophage and neutrophil
levels and localization were not different between sham and
damaged tissues (Figs. 3 and 4). Both tissues possessed a
very small number of regions in which macrophages were
present (Fig. 3c,d) although the majority of the tissue area
was void of macrophages (Fig. 3a,b) with one damaged
tissue completely negative for macrophage staining. In
contrast, macrophages were primarily localized to tissue
surrounding the MCLs where tissue was injured due to the
surgical procedure (Fig. 3e–h) and near the bone–ligament
insertion (Fig. 3i–j). Furthermore, neutrophil localization in
control and damaged tissue had a similar pattern to that seen
for macrophages (Fig. 4). In both control and damaged
ligaments, the majority of the tissue area was void of
neutrophils (Fig. 4a,b), yet both groups had very few
regions in which neutrophils were present within the
547
ligament (Fig. 4c,d). Similar to macrophages, neutrophils
were present in the (surgical insult-induced) scar tissue
surrounding the ligament (Fig. 4e–h), and near the bone–
ligament insertion (Fig. 4i–j).
Induction of a subfailure (Grade II) ligament injury
resulted in substantial changes in gene expression. Analysis
of GAPDH mRNA levels did not show any significant
differences between control and injured tissues. Collagen I
expression (Fig. 5A) showed a trend of initial suppression
after 1 day, followed by an increasing trend at 3 days. By
day 7, collagen I gene expression was significantly
increased over the sham side MCL ( p=0.004) with a 48fold increase in mRNA expression in healing tissues. Type
III collagen gene expression (Fig. 5B) also revealed a
suppressed trend at 1 day, which by 3 days was virtually
identical to sham values. After 7 days, type III collagen gene
expression was significantly increased over the sham side
MCL ( p=0.026) with a 42-fold increase in mRNA
expression in healing tissues. Hence, type I collagen
revealed a larger fold increase than type III collagen as
well as significantly larger mRNA expression than type III
( p=0.014) after 7 days. Furthermore, type I collagen
expression in damaged tissue was ~20-fold greater than
type III expression, while in sham tissues this ratio was ~17,
indicating that the increases in both collagen I and III may
Fig. 5. Type I (A) and III (B) collagen gene expression following a
subfailure (Grade II) MCL injury. Both collagen I and III expression in
subfailure damaged ligaments show a trend of initial suppression after 1
day, followed by an increasing trend at 3 days, and a significant increase by
7 days ( p=0.004 and p=0.026, respectively). Note the difference in scale
between A and B.
548
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
Fig. 6. Proteoglycan expression following a subfailure ligament stretch
reveals significantly (*) increased levels of decorin, fibromodulin, lumican,
and versican, but not biglycan, mRNA in stretched ligaments 7 days postinjury. Sham=gray bars; damaged=black bars; F=fibromodulin; D=decorin;
L=lumican; V=versican; B=biglycan.
be the result of an acceleration of the normal remodeling
response that produces much more collagen I. Analysis of
proteoglycan mRNA levels (Fig. 6) indicated significant
increases in fibromodulin ( p=0.017), decorin ( p=0.013),
lumican ( p=0.006), and versican ( p=0.030) 7 days postinjury, corresponding to increases in collagen expression.
Biglycan mRNA levels were low with no significant
changes at any time point. In addition, proteinase expression
(Fig. 7) displayed a trend of initial suppression at 1 day, with
significantly increased levels of cathepsin K ( p=0.048) and
L ( p=0.037), but not MMP-13 ( p=0.19) or MMP-2
( p=0.88), at 7 days. Matrix metalloproteinase-2 and -13
expression were not significantly different at any time point,
while TIMP-1 mRNA levels were significantly increased
after 3 ( p=0.046) and 7 ( p=0.024) days. Tartrate-resistant
acid phosphatase levels were low with no significant
changes at any time point. TNF-a mRNA levels were lower
in damaged tissue when compared to sham tissues:
57.7F15.0%, 57.4F22.1%, and 49.8F21.0% of control
values for 1, 3, and 7 days, respectively, with no statistically
significant between any time point.
Results herein point toward an intrinsic process of collagenous ECM remodeling that relies more heavily on
replacing type I collagen directly than producing a predominantly type III collagen provisional matrix (scar) to be later
replaced by type I collagen. Furthermore, proteoglycan
mRNA levels indicate significant increases in conjunction
with increased collagen expression. Since proteoglycans
such as decorin, lumican, and fibromodulin are associated
with collagen fibrillogenesis (Birk et al., 1995; Chakravarti
et al., 1998; Ezura et al., 2000; Jepsen et al., 2002; Svensson
et al., 1999; Vogel and Trotter, 1987), as well as cell
behavior (see Kresse and Schonherr, 2001), it appears likely
that they are, at least in part, up-regulated to help modulate
the fibrillogenesis of the newly synthesized collagen.
During embryonic tendon development in vivo, type I
collagen fibrillogenesis is a multi-step fibroblast-mediated
process (Birk and Trelstad, 1984, 1986; Birk et al., 1989)
during which fibril segments polymerize within deep
channels in the fibroblast surface and form fibril bundles
within a peripheral extracytoplasmic compartment of the
fibroblast (Birk and Trelstad, 1986). These collagen fibril
segments then form intermediate structures that assemble
into collagen fibrils and undergo post-depositional growth
during embryonic development (Birk et al., 1989, 1995,
1997), a process that is, in part, mediated by small leucinerich proteoglycans. For instance, studies have shown that
decorin plays a role in regulating collagen fibril organization
and maturation by delaying fibril assembly (Vogel and
Trotter, 1987), inhibiting fibril maturation (Birk et al., 1995;
Scott et al., 1981), and regulating fibril tip-to-tip fusion
(Graham et al., 2000). Furthermore, both fibromodulin and
lumican have also been implicated in regulating collagen
fibrillogenesis. When collagen fibrils from mice deficient in
fibromodulin and/or lumican are examined, they exhibit
abnormal development in terms of diameter and shape,
resulting in abnormal tissue function (Chakravarti et al.,
1998; Ezura et al., 2000; Jepsen et al., 2002; Svensson et al.,
1999). Therefore, proteoglycans, such as decorin, fibromodulin, and lumican, strongly regulate multiple aspects of
3. Discussion
Mechanical evaluation of ligaments after subfailure
injury revealed a substantial strength deficit and increased
deformations within the strain-stiffening toe-region of the
load-deformation, consistent with previous reports (Provenzano et al., 2002a,b) and indicating a significant tissue
injury. This early portion of the stress–strain curve is
physiologically relevant. Changes in mechanical behavior in
this region indicate a longer functional length prior to load
bearing, which in turn can result in joint laxity and abnormal
motion leading to degenerative joint disease. As such, tissue
remodeling following this type of injury and the biologic
attempts to resolve this abnormal behavior require study.
Fig. 7. Proteinase expression following a subfailure ligament stretch reveals
significantly (*) increased cathepsin K and L, but not MMP-13, -2, or
TRAP in stretched ligaments 7 days post-injury. Sham=gray bars;
damaged=black bars; 2=MMP-2; 13=MMP-13; CK=cathepsin K; CL=cathepsin L; T=TRAP.
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
collagen fibrillogenesis during embryonic development,
while versican can promote fibroblast proliferation (Zhang
et al., 1998). Consequently, the up-regulation of these
proteoglycans in concurrence with type I collagen after
subfailure injury implies fibrillogenesis is occurring to
remodel the tissue and restore functional mechanical
properties.
In addition to changes in expression levels of extracellular matrix molecules, cysteine proteinase mRNA levels
were also up-regulated following subfailure injury. While
these mRNA levels may only be reflective of protease
activity, this level of regulation may indicate increases in a
non-MMP protease family to clear damaged tissue and help
facilitate tissue remodeling For instance, frequently three
main classes of enzymes are strongly implicated in matrix
(particularly collagen) degradation: the MMP family (e.g.
MMP-2 and -13), the cysteine proteinase family (e.g.
cathepsin K and L), and the acid phosphatase family (e.g.
TRAP). Traditionally, the matrix metalloproteinase family
has received the most study and MMP-2 and -13 have been
implicated in matrix remodeling and increased expression
during wound healing (Creemers et al., 1998; Hellio Le
Graverand et al., 2000). However, the cysteine proteinase
family is emerging as an equally important class of enzymes
(Chapman et al., 1997). Cathepsin K is a member of the
cysteine proteinase family that possesses strong collagenolytic activity, at pH range of 4.0–7.0, which exceeds the
activity of MMP-1, -9, and -13 (Bromme et al., 1996;
Garnero et al., 1998; Hou et al., 2001; Li et al., 2000).
Moreover, in addition to known intracellular lysosomal roles
for cathepsin K and L (Kirschke et al., 1998), cathepsin L is
known to act as a potent extracellular protease (see Kirschke
et al., 1998) while there is increasing evidence of an
extracellular role for cathepsin K (Parak et al., 2000; Tepel
and Bromme, 2000) in cell types other than osteoclasts
(Saftig et al., 1998; Xia et al., 1999). Another collagenase
examined in this study, TRAP, has been implicated in
collagen degradation in bone (Henthorn et al., 1999; Oddie
and Schenk, 2000), identified in fibroblasts (Dwyer et al.,
2004, Muir et al., 2002), and is associated with macrophages and inflammation (Bune et al., 2001; Lord et al.,
1990; Schindelmeiser et al., 1989). Interestingly in this
study, expression levels for the MMPs did not change at any
time point while cysteine proteinases cathepsin K and L
were both significantly increased 7 days post-injury,
suggesting a role for these proteinases in ligament tissue
remodeling after injury. Their specific roles, however,
remain to be elucidated, particularly their potential intracellular versus extracellular functions. Furthermore, TRAP
levels did not increase, implying that a significant macrophage infiltration did not take place over the time period
examined, consistent with immunohistochemical data. This
conclusion is also supported by the constant MMP levels,
which typically show rapid increases with inflammation (see
Everts et al., 1996) and the relatively constant, but
decreased, TNF-a levels (a typical marker for inflammation)
549
in injured tissues. Hence, these data support the conclusions
of Laws and Walton (1988) that Grade II injuries in
ligaments may remodel intrinsically and differ from
complete tissue tears that involve a coagulation/inflammation/provisional-matrix response.
After subfailure tissue damage, one would intuitively
expect an up-regulation of debris-clearing proteinases to
precede collagen up-regulation. The lack of inflammatory
markers, however, suggests anabolic and catabolic events
are more intrinsically regulated and occur simultaneously.
As such, it is possible that matrix secretion and resorption
are taking place in different locations within the tissue or
they are acting in synergy to remove damaged tissue and
replace it. Although further analyses, such as in situ
hybridization and histological protein localization, are
required to lend insight into the relationship between the
anabolic and catabolic processes, current information does
raise interesting questions about the mechanisms of tissue
laxity in light of previously reported microstructural data
(Provenzano et al., 2002a,b). For instance, examination of
collagen microstructure with scanning electron microscopy
after subfailure damage revealed ruptured collagen fibers
and fibrils, and fiber and fibril dretractionT appeared to have
taken place (Fig. 8). If proteinases are to clear debris and
remove damaged tissue after injury, it is unlikely that the
initial gap between ruptured fibers and/or fibrils would, at
that point, decrease. As such, if the tissue gap is present
when new tissue is dfilled inT, the repaired fiber or fibril
would be longer than its pre-injury length and as such more
lax. It is unclear whether long-term remodeling would
resolve this issue or not, although examination of remodeling ligaments following subfailure injury revealed substantial laxity remained 2 weeks post-injury (Provenzano et al.,
2002a). Hence, further study is required of long-term
remodeling, the presence of myofibroblasts (which may
pull the fibril/fiber ends together), the location of newly
synthesized collagen, and the anabolic/catabolic processes
and their interaction.
Lastly, numerous studies have examined healing after a
complete tissue rupture or transection. However, few studies
have examined the biologic response after a subfailure
injury in which a localized concentration of scar tissue does
not appear to form. In contrast to Grade III injuries in which
scar tissue forms at the rupture site (Frank et al., 1983), via
mediation from the coagulation system and a largely
extrinsic inflammatory response (Frank et al., 1983; Riches,
1996; Williams et al., 1984), Grade II injury appears to be a
largely intrinsically mediated response. Therefore, since
inflammation (as a source of growth factors and cytokines)
plays a large role in wound healing by inducing cell
migration and proliferation and controlling extracellular
matrix scar formation (Eckes et al., 1996; Riches, 1996), its
absence leaves fibroblasts to direct tissue healing. This
difference in extrinsically versus intrinsically dominated
healing response may account for the relatively low ratio of
type III to type I collagen expression after subfailure injury
550
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
Fig. 8. Collagen fiber and fibril damage above the damage threshold. After axial tissue strains were induced in medial collateral ligaments, control tissues (0%
strain; A, C, D) possess normal microstructure, while tissues stretched above the defined damaged threshold for rat MCL (~5% or 40–60% of the failure strain)
show collagen fiber and fibril damage, disorganized fibril organization, and ruptured fibrils (B, E–H). (A) Control tissue. (B) Tissue stretched to 7.7% strain
showing collagen fiber damage. (C, D) Control ligaments displaying normal fibril organization and morphology. (E) Tissue stretched to 7.7% strain revealing a
region in which collagen fibril morphology and organization appear normal (the collagen fiber that contains this region appeared intact). (F) Tissue stretched to
6.7% strain showing substantial fibril disorganization (the fiber that contains this region was ruptured indicating a possible retraction of fibrils). (G–H) Tissue
stretched to 5.8% strain displaying collagen fibril rupture (the fiber containing this region appeared intact). Hence, shortly after injury some fibers are intact and
appear normal, some fibers are completely ruptured and a dgapT exists between ruptured fibers, and additionally some fibers are intact but contain ruptured
collagen fibrils that contain a dgapT between ruptured fibril ends. As such, if the tissue gap remains and is present when new tissue dfills inT without substantial
contracture, the repaired fiber or fibril would be longer than its pre-injury length and as such it will be more lax.
as compared to healing from a complete tear. In a more
typical wound model, type III expression is substantial and
precedes type I expression (Boykiw et al., 1998; Eckes et al.,
1996; Frank et al., 1983; Riches, 1996; Williams et al.,
1984). Herein type I collagen, and supporting proteoglycans,
appears to be up-regulated at an earlier stage with a more
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
substantial response. Therefore this model could lend unique
insight into wound healing, tissue remodeling, and fibroblast
behavior, and hence additional work is required to differentiate intrinsic behavior from the more extensively studied
extrinsic healing response.
In conclusion, this study demonstrates that subfailure
damage to connective tissue induces intrinsically mediated
remodeling with increased collagen, proteoglycan, and
proteinase gene expression 7 days after a significant
subfailure stretch was applied to collateral ligament. The
apparent harmonization of these events provides compelling evidence that between 3 and 7 days, or at 7 days, the
tissue is undergoing a dramatic remodeling response. In
addition, the strong increase in type I collagen expression,
which is significantly greater than type III expression,
supports the concept of remodeling without significant scar
deposition. These findings distinguish this injury from
pathologies such as tendonitis where inflammation is
known to play a role. Furthermore, this model raises the
likely possibility that fibroblasts within additional tissues
in the body respond in a similar fashion and deposit type I
collagen to repair moderately (i.e. non-clotting and nonscarring) damaged tissue instead of depositing a type III
fibrotic matrix. As such this model can provide unique
insight into acute tissue damage and offer a potentially
powerful model to examine fibroblast behavior and
collagen fibrillogenesis.
4. Experimental procedures
4.1. Animal procedures and experimental design
This study was approved by the Institutional Animal
Use and Care Committee. Thirty-two male Sprague–
Dawley rats were used as an animal model. Under
general anesthesia, one randomly selected MCL of each
animal was approached via skin incision without damage
to other joint structures. The MCL received a subfailure
(Grade II) injury in the same manner as previously
described in lateral collateral ligaments (Provenzano et al.,
2002a). This injury method pulls laterally on the ligament
with a controlled displacement while restraining the
insertion sites. The contralateral MCL received a sham
surgery. The incision was closed in a routine manner.
Animals were euthanized at time zero for mechanical
analysis (n=4 rats) to characterize the injury, after 24 h to
test for the presence of infiltrating macrophages and
neutrophils (n=4 rats), and at 1, 3, or 7 days for gene
expression analysis (n=8 rats/group).
4.2. Mechanical testing
For mechanical testing, four pairs of contralateral
ligaments were harvested and allowed to recover for 15
min while remaining hydrated in order to reduce the
551
effects of tissue handling during harvest and to ensure that
potential mechanical differences between control and
stretched tissues were due to matrix damage and not a
viscoelastic effect (Provenzano et al., 2001, 2002b). For
mechanical testing, the tissue was pulled to failure as
previously described (Provenzano et al., 2002a). Briefly,
MCL area was computed from optical measurements and
then the femur–MCL–tibia complex, with optical markers
near the tissue insertion sites for measuring displacement,
was placed into a custom designed tissue bath system
inserted into our testing machine. A small preload of 0.1 N
was applied to the tissue in order to obtain a uniform zero
(start) point. In displacement control, the ligaments were
preconditioned (~1% strain for 10 cycles), allowed to
recover, and pulled to failure at ~10%/s. Strains were
calculated from videotape (described below) after displacement-controlled tests. Tissue displacement was obtained
using video dimensional analysis to measure the change
in distance between markers. The change in distance
between optical markers was calculated by analyzing
stored digital frames with N.I.H. Image using a custom
macro to calculate the change in x–y coordinate center
(centroid coordinates) of each marker. Force was displayed
on the video screen and synchronized with displacement. From displacement data, strain was calculated as
the change between stretched length and gage length
divided by gage length: DL/L 0. Stress was calculated as
the force divided by the initial area. Stress–strain data were
fit with the probabilistic microstructural model of Hurschler et al. (1997) to determine elastic modulus and strain
at the transition between the nonlinear strain-stiffening region and the approximately linear portion of the stress–
strain curve (Hurschler et al., 1997; Hurschler and
Provenzano, 2003), which has functional implications in
physiologic loading range (see Appendix A). Parameters
for comparison are tissue length, tissue laxity, strain at the
transition between the strain-stiffening and linear regions
of the stress–strain curve, ultimate stress, strain at failure,
and elastic modulus. Tissue laxity was taken to be the
difference between control ligament length (L C) and
stretched ligament length (L S), normalized by the control
length: [i.e. %Laxity=100(L SL C)/L C] as previously
described (Provenzano et al., 2002a).
4.3. Immunohistochemistry
To test for the possible occurrence of infiltrating
macrophages and neutrophils into the ligaments following
injury, immunohistochemistry was employed. Tissues were
frozen in OCT embedding media and cut into 5-Am frozen
sections. Following tissue sectioning, the tissues were
thawed to room temperature, fixed for 30 min in formalin
and then washed in PBST (1% Tween-20 in PBS). To
quench endogenous peroxidase activity, tissue sections
were blocked in 3% H2O2 and washed again in PBST.
Tissue sections were then subjected to nonspecific protein
552
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
blocking in 5% goat serum in PBS, washed in PBST, and
incubated with the primary antibody. To detect infiltrating
macrophages the mouse anti-rat ED1 (rat CD68) antibody
(Serotec, Oxford, UK) was utilized at a 1:50 dilution for 2
h at room temperature. To detect neurtrophils, a mouse
anti-rat antibody (Serotec) to the CD43 antigen was
utilized at a 1:100 dilution for 2 h at room temperature.
After extensive washing in PBST, immunochemical labels
were applied using the STAT-Q-DAB kit (Innovex
Biosciences, San Ramon, CA) with mouse-linked (ratadsorbed) secondary antibody. Tissue sections were incubated with the secondary linking antibody, labeled with
peroxidase (HRP)-labeled strepavidin, followed by application of 3,3-diaminobenzidine substrate. The tissue
sections were then washed, dehydrated, cover-slipped,
and imaged with light microscopy. Spleen was stained as
a positive control.
4.4. RNA isolation and reverse transcription
For gene expression analysis, eight contralateral pairs of
tissues were harvested per group. Under anesthesia, tissues
were harvested without surrounding scar tissue, rinsed with
DNase-, RNase-, and protease-free phosphate-buffered
saline to remove blood or other contaminants from the
tissues, placed in RNAlater solution (Ambion, Austin,
TX), and stored at 20 8C. Utilizing methods similar to
Reno et al. (1997), which combines the TRIzol method
with the column fractionation steps of the RNeasy Total
RNA kit (Qiagen, Valencia, CA), total RNA was later
isolated. Tissues were pooled in groups of two to ensure
quantifiable RNA. Tissues were placed in a liquid nitrogen
cooled Braun Mikro-Dismembrator Vessel (B. Braun
Biotech International, Allentown, PA) and reduced to
powder. One milliliter of Trizol Reagent (Life Technologies, Grand Island, NY) was then added to the homogenate
and incubated at room temperature for 5 min. The
homogenate was then centrifuged to remove insoluble
tissue debris, followed by the addition of 0.2 mL of
chloroform to the supernate. After the addition of chloroform, the samples were mixed vigorously, incubated for 3
min at room temperature, and transferred to a pre-spun 2.0
mL heavy phase-lock gel tube (Brinkman-Eppendorf,
Westbury, NY). The organic and aqueous phases were
then separated by centrifugation at 14,000g for 5 min,
after which an equal volume of 70% EtOH was added to
Table 1
PCR primer sequences
Target gene
Primer sequence 5V–3V
Amplicon length
Collagen I (col1a1)
FP: CGT GAC CAA AAA CCA AAA GTG C
RP: GGG GTG GAG AAA GGA ACA GAA A
FP: CGA GGT AAC AGA GGT GAA AGA
RP: AAC CCA GTA TTC TCC GCT CTT
FP: CTG CTG TAT GTC CGG CTG TCT T
RP: GCT GCG CTT GAT CTC GTT CC
FP: GAG GGC TCC GGT GGC AAA TA
RP: TGG CAT CGG TAG GGG CAC ATA
FP: GTC CGC TCC CAA AGT CCC TAC A
RP: AAG CCC GGT GAA GCG ATT GA
FP: TTC CAA GGG CAA TGC TAC AAG T
RP: TCA TGG CCC ACA CGA TTC ACA A
FP: CAG GGG AAG GGG AAC AAA GAA G
RP: ATG CCC GGA TGG AAC AGT AAC
FP: AAA GAA CAT GGT GAC TTC TAC C
RP: ACT GGA TTC CTT GAA CGT C
FP: GGT CGC AGT GAT GGC TTC CTC T
RP: CAC ACC ACA CCT TGC CAT CGT T
FP: TGC GAC CGT GAT AAT GTG AAC C
RP: ATG GGC TGG CTG GCT TGA ATC
FP: GAC GGT GGG GCC TAT TTC TG
RP: TTC CGG TCT TTG GCT ATT TTG A
FP: CGC CAG AAC CGT GCA GAT TAT G
RP: AAG ATG GCC ACG GTG ATG TTC G
FP: ACA GCT TTC TGC AAC TCG
RP: CTA TAG GTC TTT ACG AAG GCC
FP: TCA GCC GAT TTG CCA TTT CAT A
RP: AAC ACG CCA GTC GCT TCA CA
FP: GAC TGT GGA TGG CCC CTC TG
RP: CGC CTG CTT CAC CAC CTT CT
186
Collagen III (col3a1)
Fibromodulin
Decorin
Lumican
Versican
Biglycan
MMP-13
MMP-2
Cathepsin K
Cathepsin L
TRAP
TIMP-1
TNF-a
GAPDH
349
267
218
296
140
264
283
376
205
237
297
335
251
239
GenBank accession numbers for col1a1, col3a1, decorin, lumican, fibromodulin, versican, biglycan, MMP-13, MMP-2, cathepsin K, cathepsin L, TRAP,
TIMP-1, TNF-a, and GAPDH are Z78279, XM_343563, Z12298, X84039, X82152, AF072892, U17834, M60616, NM_031054, AF010306, NM_013156,
M76110, L31883, AF329982, and AF106860, respectively.
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
the RNA-containing aqueous phase. Total RNA was then
isolated from the ethanol mixture using the column
fractionation steps of the RNeasy Total RNA kit. Yield
and purity of RNA were quantified by spectrophotometric
measurement at 260, 280, and 320 nm (UltraSpec 3000
Spectrophotometer, Pharmacia Biotech, Cambridge, UK).
After RNA quantification, reverse transcription into cDNA
was performed at 42 8C in a 20 AL total volume
containing 5 AL total RNA, 1 AL oligo(dT)15 (500 ng/
AL; Promega, Madison, WI), RNase-free water, and 9 AL
of First Strand Synthesis buffer containing 40 units of
RNase inhibitor (Ambion), 500 AM dNTP mix, 10 mM
DTT, and 200 units of Superscript II reverse transcriptase
(Invitrogen, Life Technologies).
4.5. Real-time quantitative PCR
For real-time quantitative PCR, primers for collagen I
(col1a1), collagen III (col3a1), decorin, lumican, fibromodulin, biglycan, versican, rat collagenase (matrix metalloproteinase (MMP)-13), gelatinase-A (MMP-2), cathepsin
K, cathepsin L, tartrate-resistant acid phosphatase (TRAP),
tissue inhibitor of metalloproteinase 1 (TIMP-1), tumor
necrosis factor-a (TNF-a), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were utilized (Table 1).
The primer sets were either derived from previously
published reports for MMP-13 (Knittel et al., 1999) or
designed from sequences available from GenBank using
the PrimerSelect module of the Lasergene 5.01 software
(DNAStar, Madison, WI) and BLASTR (Basic Local
Alignment Search Tool; http://www.ncbi.nlm.nih.gov:80/
BLAST). Note, versican has four isoforms (Lemire et al.,
1999), so the versican primer was chosen in the conserved
region for all four isoforms for mouse, rat, bovine genes.
In order to perform quantitative analysis of each gene,
QPCR standards for each of the genes were prepared from
purified reverse transcription-PCR products of the target
sequences. From spectrophotometric quantitation of the
PCR products, the number of copies per AL of each
standard was calculated, and 10-fold serial dilutions
(ranging from 1109 to 10 copies/AL) were prepared.
For mRNA quantification, real-time quantitative-PCR was
performed using a BIO-RAD iCycler iQ Real-time PCR
system (BIO-RAD, Hercules, CA). All reactions were
carried out in a total volume of 20 AL containing 1
Platinum Quantitative PCR Supermix (Life Technologies),
10 nM fluorescein, 200 nM forward primer, 200 nM
reverse primer, 0.25 Sybr Green, template cDNA, and
RNase-, DNase-, pyrogen-free H2O at an annealing
temperature of 55–60 8C (depending upon the primers).
For each gene, an assessment of quality control was
performed by (1) visualizing RT-PCR products for
expected size and the lack of large molecular weight
genomic DNA contamination, (2) by including in each
QPCR experiment two negative controls (one with no
template and one with reverse transcriptase excluded), and
553
(3) by examining PCR melt-curves following QPCR to
ensure specificity. All reactions were specific. The cDNA
copy numbers of the target gene were analyzed after being
normalized to the copy number of GAPDH.
4.6. Scanning electron microscopy
Scanning electron microscopy was performed as previously described (Provenzano et al., 2002a) on additional
(to those listed above) ligaments stretched to measured
levels of strain.
4.7. Statistical analysis
For statistical analysis, paired t-tests were applied to
analyze differences in mechanical properties at time zero
and gene expression within a time group. Gene expression
data were log-transformed to conform to the t-test
assumptions. All groups that were significantly different
had at least a 10-fold change in gene expression, except two
samples that had 3- and 4-fold changes, respectively.
Acknowledgements
This work was funded through grants from N.A.S.A.
(NAG9-1152), the N.S.F. (CMS9907977), and a N.I.H.
Predoctoral Training Grant (#T32 GM07215) in Molecular Biosciences (AAO). The authors thank Dr. David
Hart and Carol Reno for assistance with obtaining sample
preparation equipment for RNA extraction, Ron McCabe
and Anthony Escarcega for technical assistance with
laboratory equipment, the Materials Science Center for
access to the scanning electron microscope, and Phil
Oshel and the Biological and Biomaterial Preparation,
Imaging, and Characterization Laboratory (BBPIC) for
technical assistance with electron microscopy specimen
preparation.
Appendix A. Determination of the toe-to-linear region
transition in fibrous connective tissue
The ligament stress–strain curve displays nonlinear
strain-stiffening behavior in the dtoeT region after which
a more linear region is encountered. This early portion of
the stress–strain curve is physiologically relevant. Nonrecoverable stretch changes in the mechanical behavior in
this region can result in joint laxity and abnormal motion
leading to degenerative joint disease. The model described
in this appendix quantifies the transition from the dtoeT
region to the linear region and has been described in detail
elsewhere (Hurschler et al., 2003). Stress–stretch data were
fit with the probabilistic microstructural model of Hurschler et al. (1997) which utilized a Weibull probability
distribution function (pdf) to represent collagen fiber
554
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
recruitment and hence load bearing capability as the tissue
is stretched and the characteristic crimp pattern of the
collagen fiber is removed. The model has the form:
rt ðkt Þ ¼
Zkt
Pw ðks Þr33 ðkt =ks Þdks ;kt Nc
ð1Þ
c
where P w is the Weibull pdf as a function of the
straightening stretch ratio, defined as the tissue stretch at
which an individual fiber begins to bear load (k s=l s/
l o=fiber length/tissue reference length), r 33 is the longitudinal normal stress in a fiber, and c is the location
parameter of the Weibull distribution and defines the onset
of fiber loading (Hurschler et al., 2003). Hence, tissue
stress is determined by the overall tissue stretch (stretch
ratio=k t =L/L 0=deformed tissue length/tissue reference
length) and the load bearing contribution of the fiber
population. After the data have been well described by the
model, parameters from the Weibull pdf are input into the
Weibull cumulative distribution function (cdf), F(k):
Fðkt Þ ¼
Zkt
c
Pðks Þdks ¼
exp Zkt
c
ks c
d
b
d
ks c
d
b dks:
b1
ð2Þ
After integration, Eq. (2) can be algebraically manipulated to yield the inverse cdf,
kt ¼ d½ lnð1 F Þ1=b þ c
ð3Þ
that defines the stretch ratio at which any fraction of the
fibers ( F) have been recruited. Hence, Eq. (3) defines the
stretch (and hence stress) at which the tissue exits the nonlinear strain-stiffening dtoe-regionT and enters the more
linear portion of the curve.
References
Andriachi, T., et al., 1987. Ligament: injury and repair. In: Woo, S.L.-Y.,
Buckwalter, J.A. (Eds.), Injury and Repair of the Musculoskeletal Soft
Tissues. AAOS, Park Ridge, IL.
Birk, D.E., Trelstad, R.L., 1984. Extracellular compartments in matrix
morphogenesis: collagen fibril, bundle, and lamellar formation by
corneal fibroblasts. Journal of Cell Biology 99, 2024 – 2033.
Birk, D.E., Trelstad, R.L., 1986. Extracellular compartments in tendon
morphogenesis: collagen fibril, bundle, and macroaggregate formation.
Journal of Cell Biology 103, 231 – 240.
Birk, D.E., Zycband, E.I., Winkelmann, D.A., Trelstad, R.L., 1989.
Collagen fibrillogenesis in situ: fibril segments are intermediates in
matrix assembly. Proceedings of the National Academy of Sciences of
the United States of America 86, 4549 – 4553.
Birk, D.E., Nurminskaya, M.V., Zycband, E.I., 1995. Collagen fibrillogenesis in situ: fibril segments undergo post-depositional modifications
resulting in linear and lateral growth during matrix development.
Developmental Dynamics 202, 229 – 243.
Birk, D.E., Zycband, E.I., Woodruff, S., Winkelmann, D.A., Trelstad, R.L.,
1997. Collagen fibrillogenesis in situ: fibril segments become long
fibrils as the developing tendon matures. Developmental Dynamics 208,
291 – 298.
Boykiw, R., Sciore, P., Reno, C., Marchuk, L., Frank, C.B., Hart, D.A.,
1998. Altered levels of extracellular matrix molecule mRNA in healing
rabbit ligaments. Matrix Biology 17, 371 – 378.
Bromme, D., Okamoto, K., Wang, B.B., Biroc, S., 1996. Human cathepsin
O2, a matrix protein-degrading cysteine protease expressed in
osteoclasts. Functional expression of human cathepsin O2 in Spodoptera frugiperda and characterization of the enzyme. Journal of
Biological Chemistry 271, 2126 – 2132.
Bune, A.J., Hayman, A.R., Evans, M.J., Cox, T.M., 2001. Mice lacking
tartrate-resistant acid phosphatase (Acp 5) have disordered macrophage
inflammatory responses and reduced clearance of the pathogen,
Staphylococcus aureus. Immunology 102, 103 – 113.
Chakravarti, S., Magnuson, T., Lass, J.H., Jepsen, K.J., LaMantia, C.,
Carroll, H., 1998. Lumican regulates collagen fibril assembly: skin
fragility and corneal opacity in the absence of lumican. Journal of Cell
Biology 141, 1277 – 1286.
Chapman, H.A., Riese, R.J., Shi, G.P., 1997. Emerging roles for cysteine
proteases in human biology. Annual Review of Physiology 59, 63 – 88.
Creemers, L.B., Jansen, I.D., Docherty, A.J., Reynolds, J.J., Beertsen, W.,
Everts, V., 1998. Gelatinase A (MMP-2) and cysteine proteinases are
essential for the degradation of collagen in soft connective tissue.
Matrix Biology 17, 35 – 46.
Dwyer, K.W., Provenzano, P.P., Muir, P., Valhmu, W.B., Vanderby Jr., R.,
2004. Blockade of the sympathetic nervous system degrades ligament in
a rat MCL model. Journal of Applied Physiology 96, 711 – 718.
Eckes, B., Aumailley, M., Krieg, T., 1996. Collagens and the reestablishment of dermal integrity. In: Clark, R.A.F. (Ed.), The Molecular and
Cellular Biology of Wound Repair, 2nd ed. Plenum Press, New York
p. 493.
Everts, V., van der Zee, E., Creemers, L., Beertsen, W., 1996. Phagocytosis
and intracellular digestion of collagen, its role in turnover and
remodelling. Histochemical Journal 28, 229 – 245.
Ezura, Y., Chakravarti, S., Oldberg, A., Chervoneva, I., Birk, D.E., 2000.
Differential expression of lumican and fibromodulin regulate collagen
fibrillogenesis in developing mouse tendons. Journal of Cell Biology
151, 779 – 788.
Frank, C., Woo, S.L., Amiel, D., Harwood, F., Gomez, M., Akeson, W.,
1983. Medial collateral ligament healing. A multidisciplinary assessment in rabbits. American Journal of Sports Medicine 11, 379 – 389.
Fruensgaard, S., Johannsen, H.V., 1989. Incomplete ruptures of the anterior
cruciate ligament. Journal of Bone and Joint Surgery. British Volume
71, 526 – 530.
Garnero, P., Borel, O., Byrjalsen, I., Ferreras, M., Drake, F.H., McQueney,
M.S., Foged, N.T., Delmas, P.D., Delaisse, J.M., 1998. The collagenolytic activity of cathepsin K is unique among mammalian proteinases.
Journal of Biological Chemistry 273, 32347 – 32352.
Graham, H.K., Holmes, D.F., Watson, R.B., Kadler, K.E., 2000. Identification of collagen fibril fusion during vertebrate tendon morphogenesis. The process relies on unipolar fibrils and is regulated by
collagen–proteoglycan interaction. Journal of Molecular Biology 295,
891 – 902.
Hellio Le Graverand, M.P., Eggerer, J., Sciore, P., Reno, C., Vignon, E.,
Otterness, I., Hart, D.A., 2000. Matrix metalloproteinase-13 expression
in rabbit knee joint connective tissues: influence of maturation and
response to injury. Matrix Biology 19, 431 – 441.
Henthorn, P.S., Millan, J.L., Leboy, P., 1999. Acid and alkaline
phosphatase. In: Seibel, M.J., Robins, S.P., Bilezikian, J.P. (Eds.),
Dynamics of Bone of Cartilage Metabolism. Academic Press, San
Diego, pp. 137 – 150.
Hou, W.S., Li, Z., Gordon, R.E., Chan, K., Klein, M.J., Levy, R., Keysser,
M., Keyszer, G., Bromme, D., 2001. Cathepsin k is a critical protease in
synovial fibroblast-mediated collagen degradation. American Journal of
Pathology 159, 2167 – 2177.
P.P. Provenzano et al. / Matrix Biology 23 (2005) 543–555
Hurschler, C., Loitz-Ramage, B., Vanderby Jr., R., 1997. A structurally
based stress–stretch relationship for tendon and ligament. Journal of
Biomechanical Engineering 119, 392 – 399.
Hurschler, C., Provenzano, P.P., Vanderby Jr., R., 2003. Application of a
probabilistic microstructural model to determine reference length and
toe-to-linear region transition in fibrous connective tissue. ASME
Journal of Biomechanical Engineering 125, 415 – 422.
Jack, E.A., 1950. Experimental rupture of the medial collateral ligament of
the knee. Journal of Bone and Joint Surgery 32B, 396 – 402.
Jepsen, K.J., Wu, F., Peragallo, J.H., Paul, J., Roberts, L., Ezura, Y.,
Oldberg, A., Birk, D.E., Chakravarti, S., 2002. A syndrome of joint
laxity and impaired tendon integrity in lumican- and fibromodulindeficient mice. Journal of Biological Chemistry 277, 35532 – 35540.
Kirschke, H., Barrett, A.J., Rawlings, N.D., 1998. Lysosomal Cysteine
Proteinases. Oxford Press, New York.
Knittel, T., Mehde, M., Kobold, D., Saile, B., Dinter, C., Ramadori, G.,
1999. Expression patterns of matrix metalloproteinases and their
inhibitors in parenchymal and non-parenchymal cells of rat liver:
regulation by TNF-alpha and TGF-beta1. Journal of Hepatology 30,
48 – 60.
Kresse, H., Schonherr, E., 2001. Proteoglycans of the extracellular matrix
and growth control. Journal of Cellular Physiology 189, 266 – 274.
Laws, G., Walton, M., 1988. Fibroblastic healing of grade II ligament
injuries. Histological and mechanical studies in the sheep. Journal of
Bone and Joint Surgery. British Volume 70, 390 – 396.
Lemire, J.M., Braun, K.R., Maurel, P., Kaplan, E.D., Schwartz, S.M., Wight,
T.N., 1999. Versican/PG-M isoforms in vascular smooth muscle cells.
Arteriosclerosis, Thrombosis and Vascular Biology 19, 1630 – 1639.
Li, Z., Hou, W.S., Bromme, D., 2000. Collagenolytic activity of cathepsin
K is specifically modulated by cartilage-resident chondroitin sulfates.
Biochemistry 39, 529 – 536.
Lord, D.K., Cross, N.C., Bevilacqua, M.A., Rider, S.H., Gorman, P.A.,
Groves, A.V., Moss, D.W., Sheer, D., Cox, T.M., 1990. Type 5 acid
phosphatase. Sequence, expression and chromosomal localization of a
differentiation-associated protein of the human macrophage. European
Journal of Biochemistry 189, 287 – 293.
Muir, P., Hayashi, K., Manley, P.A., Colopy, S.A., Hao, Z., 2002.
Evaluation of tartrate-resistant acid phosphatase and cathepsin K in
ruptured cranial cruciate ligaments in dogs. American Journal of
Veterinary Research 63, 1279 – 1284.
Oddie, G.W., Schenk, G., Angle, N.Z., 2000. Structure, function, and
regulation of tartrate-resistant acid phosphatase. Bone 27, 575 – 584.
Panjabi, M.M., Yoldas, E., Oxland, T.R., Crisco III, J.J., 1996. Subfailure
injury of the rabbit anterior cruciate ligament. Journal of Orthopaedic
Research 14, 216 – 222.
Parak, W.J., Dannohl, S., George, M., Schuler, M.K., Schaumburger, J.,
Gaub, H.E., Muller, O., Aicher, W.K., 2000. Metabolic activation
stimulates acid production in synovial fibroblasts. Journal of Rheumatology 27, 2312 – 2322.
Provenzano, P.P., Lakes, R.S., Keenan, T., Vanderby, R.J., 2001. Nonlinear ligament viscoelasticity. Annals of Biomedical Engineering 29,
908 – 914.
Provenzano, P.P., Hayashi, K., Kunz, D.N., Markel, M.D., Vanderby Jr., R.,
2002a. Healing of subfailure ligament injury: comparison between
immature and mature ligaments in a rat model. Journal of Orthopaedic
Research 20, 975 – 983.
555
Provenzano, P.P., Heisey, D., Hayashi, K., Lakes, R., Vanderby Jr., R.,
2002b. Subfailure damage in ligament: a structural and cellular
evaluation. Journal of Applied Physiology 92, 362 – 371.
Provenzano, P.P., Muir, P., Hao, Z., Vanderby Jr., R., 2002c. Apoptosis in
ligament after a subfailure stretch. Proc. of the World Congress on
Biomech., Calgary, CA.
Reno, C., Marchuk, L., Sciore, P., Frank, C.B., Hart, D.A., 1997. Rapid
isolation of total RNA from small samples of hypocellular, dense
connective tissues. Biotechniques 22, 1082 – 1086.
Riches, D.W.H., 1996. Macrophage involvement in wound repair,
remodeling, and fibrosis. In: Clark, R.A.F. (Ed.), The Molecular and
Cellular Biology of Wound Repair, 2nd ed. Plenum Press, New York
p. 95.
Saftig, P., Hunziker, E., Wehmeyer, O., Jones, S., Boyde, A., Rommerskirch, W., Moritz, J.D., Schu, P., von Figura, K., 1998. Impaired
osteoclastic bone resorption leads to osteopetrosis in cathepsin-Kdeficient mice. Proceedings of the National Academy of Sciences of the
United States of America 95, 13453 – 13458.
Schindelmeiser, J., Schewe, P., Zonka, T., Munstermann, D., 1989.
Histochemical and immunological demonstration of purple acid
phosphatase in human and bovine alveolar macrophages. Histochemistry 92, 81 – 85.
Scott, J.E., Orford, C.R., Hughes, E.W., 1981. Proteoglycan–collagen
arrangements in developing rat tail tendon. An electron microscopical
and biochemical investigation. Biochemical Journal 195, 573 – 581.
Scuderi, G.R., Scott, W.N., 2001. Classification of knee ligament injuries.
In: Insall, J.N., Scott, W.N. (Eds.), Surgery of the Knee, vol. 1.
Churchill Livingstone, Philadelphia, p. 590.
Svensson, L., Aszodi, A., Reinholt, F.P., Fassler, R., Heinegard, D.,
Oldberg, A., 1999. Fibromodulin-null mice have abnormal collagen
fibrils, tissue organization, and altered lumican deposition in tendon.
Journal of Biological Chemistry 274, 9636 – 9647.
Tepel, C., Bromme, D., Herzog, V., Brix, K., 2000. Cathepsin K in thyroid
epithelial cells: sequence, localization and possible function in
extracellular proteolysis of thyroglobulin. Journal of Cell Science 113
(Pt. 24), 4487 – 4498.
Viidik, A., 1972. Simultaneous mechanical and light microscopic studies of
collagen fibers. Zeitschrift fur Anatomie und Entwicklungsgeschichte
136, 204 – 212.
Vogel, K.G., Trotter, J.A., 1987. The effect of proteoglycans on the
morphology of collagen fibrils formed in vitro. Collagen and Related
Research 7, 105 – 114.
Williams, I.F., McCullagh, K.G., Silver, I.A., 1984. The distribution of
types I and III collagen and fibronectin in the healing equine tendon.
Connective Tissue Research 12, 211 – 227.
Xia, L., Kilb, J., Wex, H., Li, Z., Lipyansky, A., Breuil, V., Stein, L.,
Palmer, J.T., Dempster, D.W., Bromme, D., 1999. Localization of rat
cathepsin K in osteoclasts and resorption pits: inhibition of bone
resorption and cathepsin K-activity by peptidyl vinyl sulfones. Biological Chemistry 380, 679 – 687.
Yahia, L., Brunet, J., Labelle, S., Rivard, C.H., 1990. A scanning electron
microscopic study of rabbit ligaments under strain. Matrix 10, 58 – 64.
Zhang, Y., Cao, L., Yang, B.L., Yang, B.B., 1998. The G3 domain of
versican enhances cell proliferation via epidermial growth factor-like
motifs. Journal of Biological Chemistry 273, 21342 – 21351.