In Vitro Models for Injurious Compression of Bovine and Human Articular Cartilage
by
Parth Patwari
M.D., Northwestern University Medical School, 1998
B.S. Biomedical Engineering, Northwestern University, 1994
Submitted to the Department of Electrical Engineering and Computer Science in Partial
Fulfillment of the Requirements for the Degree of
Doctor of Science in Electrical Engineering
at the
Massachusetts Institute of Technology
June 2003
© 2003 Massachusetts Institute of Technology
Signature of Author:
Department of Electrical Engineering and Computer Science
June 2003
Certified by:
Alan J. Grodzinsky
Professor of Electrical Engineering, Mechanical Engine ring, and Bioengineering
Thesis Supervisor
Accepted by:
.....................................
AArthur
r h-u--C-.
C. Sm Th
Professor of Electrical Engineering and Computer Science
Chairman, Committee on Graduate Students
MASSACHUSETTS INSTITUTE
OF TECHNOLOGY
LUL 0
03
LIBRARIES
In Vitro Models for Injurious Compression of Bovine and Human Articular Cartilage
by
Parth Patwari
Submitted to the Department of Electrical Engineering and Computer Science in Partial
Fulfillment of the Requirements for the Degree of Doctor of Science in Electrical
Engineering
ABSTRACT
Patients who have sustained a traumatic joint injury, such as a ligament rupture or
cartilage fracture, are known to have an increased risk for the development of
osteoarthritis (OA) in that joint. This has motivated the use of in vitro models of
mechanical cartilage injury in order to identify processes that could lead to cartilage
degradation. The overall aims of the work presented here has been to focus on further
identification of the effects of in vitro injurious mechanical compression of cartilage on
cell-mediated processes and to develop models for injurious compression of the
cartilage that seek to incorporate interactions with other joint tissues, such as
inflammatory mediators elaborated from the joint capsule tissue. Major results are that i)
injurious compression of newborn bovine cartilage can result in cell death predominantly
by an apoptotic mechanism; ii) incubation of mechanically injured bovine and human
cartilage with exogenous cytokines produces a synergistic increase in proteoglycan loss
from the cartilage; and iii) coincubation of cartilage with joint capsule tissue profoundly
inhibits cartilage biosynthetic activity through an IL-1-independent pathway. We also
characterize the activity and activation of ADAMTS-4 (aggrecanase-1) in the cartilage
tissue, one of the major enzymes responsible for proteoglycan degradation. These
results may help to identify some of the interactions between mechanical stimuli, joint
tissue damage, and chondrocyte behavior which lead to unbalanced cartilage
degradation and arthritis.
Thesis Supervisor: Alan J. Grodzinsky
Title: Professor of Electrical Engineering, Mechanical Engineering, and Bioengineering
Thesis Committee Members:
Dennis M. Freeman
Title: Assistant Professor, Electrical Engineering and Computer Science, MIT
Christopher H. Evans
Title: Professor of Orthopaedic Surgery, Harvard Medical School
2
Table of Contents
FO RE W ARD .................................................................................................--.......--------....-.----------------.---------.------------- 7
CH APTER
1 ...........................................................................................-------
........--
-- -----.. . . . . . . . . . . . .
10
IN VITRO MODELS FOR INVESTIGATION OF THE EFFECTS OF ACUTE MECHANICAL INJURY ON CARTILAGE ........... 10
12
In tro d u ctio n .......................................................................................................................................................
12
In Vivo Studies of TraumaticJoint Injury.......................................................................................................
15
In Vitro Studies of MechanicallyInduced CartilageInjury...........................................................................
19
Acute MechanicalInjury IncreasesMMP-3 Gene Expression .................................
21
CartilageMatrix Degradationin Response to Acute MechanicalInjury ......................................................
Effects of Acute Mechanical Injury to Cartilageusing a Cartilage and Joint Capsule CoincubationSystem ..22
-----------....................
31
QUANTITATIVE INVESTIGATION OF CELL DEATH AFTER INJURIOUS COMPRESSION OF BOVINE CALF ARTICULAR
----.............
CARTILAGE BY ELECTRON MICROSCOPY................................................................................................
31
CH APTER 2 ...........................................................................................................................
33
In trodu c tio n .......................................................................................................................................................
34
Me th o ds ..............................................................................................................................................................
.---. ....----............ 34
C artilag e Harv est.........................................................................................................................-35
Inju rio u s com p ression .....................................................................................................................................................
..--- 3 5
...
.................... ......--TU N E L stain in g...................................................................................................................
...... .. ........... 3 5
....
.....
E lectro n m icro sco p y..................................................................................................................
-------------........... 36
.....
Experimental Design ...................................................................................................................
. . ................ 3 6
..
S tatistical An alysis ..........................................................................................................................
36
R es u lts ................................................................................................................................................................
......... 36
Injurious Compression ................................................................................................................................
Cellular M orphology on Electron M icroscopy................................................................................................................37
37
Cell Death in Freeze-Thawed Cartilage ..........................................................................................................................
37
Apoptosis after Injurious Compression ...........................................................................................................................
Cell M orphology after Injurious Compression by Electron M icroscopy...................................................................
CH AP
3.TER
.........................................................................................................................................
38
38
D isc u ssio n ..........................................................................................................................................................
....--------.. . . .
46
PROTEOGLYCAN DEGRADATION AFTER INJURIOUS COMPRESSION OF BOVINE AND HUMAN ARTICULAR CARTILAGE
46
IN VITRO: INTERACTION W ITH EXOGENOUS CYTOKINES...................................................................................
In trodu c tion ..............................................................................................
.......................................................
Materialsand Methods ..............................................................................................
..... ...................-Bovine articular cartilage tissue ...................................................................................................
Postmortem adult human donor tissue.............................................................................................................................50
Inju rio u s co mpressio n .....................................................................................................................................................
B io ch em ical an alysis .......................................................................................................................................................
E xp re ssio n an aly sis .........................................................................................................................................................
Cycloheximide treatment and injury ...............................................................................................................................
Exogenous cytokine treatment and injury .......................................................................................................................
S ta tistic s ..........................................................................................................................................................................
R es u lts ................................................................................................................................................................
D isc uss io n ..........................................................................................................................................................
CH APTER 4 ..............................................................................................................
...............--------.....................
INHIBITION OF BOVINE ARTICULAR CARTILAGE BIOSYNTHESIS BY COINCUBATION WITH JOINT CAPSULE TISSUE:
..................
EVIDENCE FOR AN IL-1-INDEPENDENT PATHWAY ................................................................
48
49
49
51
51
52
52
53
53
54
57
69
69
71
Introduction...................................................................................................
72
...........................................................
.......
Experimental Procedures...........................................
....................... 7 2
... ...............
M ateria ls..............................................................................................
........ ............................. 73
Tissue Harvest............................................................................................
...... ---............ 73
.....
......................................
............................................................................
Tissue
Donor
Human
3
Injurious compression .......................................
Radiolabel Incorporation.........................................
Proteoglycan C oncentration ....................................
Statistical A nalysis ...............................................
Results........................................-.-.
-- - - - - 74
- ----- .............................................-....................................
................................................--- -------------.....--- 74
........................
74
- - - --........................
.. --- . . .-..............................................
- - - - - -........-.................... . . ..............................................-- - - ------......-................................................---
Coincubation of injured cartilage results in a synergistic increase in proteoglycan loss ............................................
.
74
75
Inhibition of proteoglycan biosynthetic activity of cartilage by conditioned medium from joint capsule...................77
The inhibitory activity of conditioned medium was heat labile.....................................................................78
78
IL-I blockade had no effect on the inhibitory factor ...................................................................................................
79
Combined blockade of IL-I and TNF had no effect on the inhibitory factor ............................................................
.80
........
.............................................................
tissue
donor
human
in
synthesis
cartilage
inhibits
Coincubation
80
. -- -- - - - - --.. . . . . . . . . . . . . . . . . . . .
......
-.
D iscussion...........................................................................
CH APTER
5 ..........................................................................................-
..
. ... ---------------------..............................
89
AGGRECAN CLEAVAGE AND ADAMTS-4 PROCESSING IN IL-I-STIMULATED BOVINE CALF ARTICULAR CARTILAGE
89
....... ......................................................................
EX PLAN TS................................................................................
90
...... . ......................................................
91
.............................................
..
.
-..
.
..
Methods..........................................
91
...................................-----.-...............................................
Cartilage explant and culture...............
92
....- -------------................
....................................
B iochem ical A nalysis of Proteoglycans .........................................
92
......
..
(aggrecanase-1)........................................
of
ADAMTS-4
Western analysis
... . . .... -------............................................ . . . 9 3
R esu lts .......................................................
93
Kinetics of aggrecan degradation in calf explants treated with IL-I ...........................................................................
95
.....
.................................................................
protein
ADAMTS-4
for
system
explant
of
Analysis
96
...... . . ...........................................................................
Discu ss io n .........................................................
Introduction...................................................
APPENDIX A ....................................................................................................-----------......................................
103
TISSUE MATURATION AND ANTIOXIDANTS ALTER THE APOPTOTIC RESPONSE OF ARTICULAR CARTILAGE AFTER
103
---..................................
--- ---..........-M ECHAN ICAL INJURY .........................................................................-......
1 05
........ . . ..............................................................................
In trodu ctio n ...................................................
106
.
..............................................................................
.........
.......................................
Materialsand Methods
106
.................... ........ ...... ---.--------------.........................................
M aterials..................................................................
Isolation and culturing of articular cartilage explants....................................................................................................107
........---------..................... 107
Injurious C om pression ...........................................................................................................
................................... 107
Histological detection of apoptosis .........................................................................................
8
of
apoptosis............................................................................................................................10
U ltrastructural d etection
10 8
B io sy nth etic activ ity......................................................................................................................................................
10 9
S tatistica l an a ly sis .........................................................................................................................................................
1
10
--.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
-.......R esu lts ............................................................................................................
...........-----------------...................... 1 10
D etection of Apoptosis................................................................................................
I11
M nTM PyP inhibits apoptosis produced by injurious com pression ...............................................................................
1
12
D isc ussio n ........................................................................................................................................................
A PPEND IX B...............................................................................................................
........
-------------...................
121
PRELIMINARY RESULTS: EFFECTS OF INJURIOUS COMPRESSION ON DEGRADATIVE ENZYME EXPRESSION AND
ACTIV ITY ......................................................................................................
A PPEN DIX C ..........................................................................................................
........................................................
12 1
. -------------------------................... 124
QUANTIFICATION OF OP-1 PROTEIN LEVELS AFTER INJURIOUS COMPRESSION OF HUMAN DONOR KNEE AND ANKLE
CA RTILA G E ............................................................................................-.......................
-....................................... 124
APPENDIX D .....................................................................................................-------------------..............................
129
PREDICTORS OF PROTEOGLYCAN LOSS IN NORMAL HUMAN POST-MORTEM KNEE AND ANKLE CARTILAGE........ 129
129
...... . ..................................................................
Introduction .........................................
130
. . .......................................................................
Meth o ds.....................................
130
................... . . ..............................................................................
Injurio u s co m p ression ...........................
130
........-----..................
..
S tatistical A n aly sis .........................................................................................................--.----
4
R esu lts an d D iscuss io n .................................................................................................................................... 1 3 1
A PPENDIX E ........................................................................................................................................................... 135
EFFECT OF G LUCOSAM INE ON CARTILAGE BIOSYNTHETic ACTIVITY ................................................................... 135
APPENDIX F ........................................................................................................................................................... 141
M ISCELLANEOUS EXPERIMENTAL RESULTS ..........................................................................................................141
RE FEREN CES ........................................................................................................................................................ 150
AIR
)\,R Ro Atj+4 W. REM~
t,
'JR.
4%R2WM
r) , ET-NL.
111k,
A.
5
S0 0 E
RX.
1%4
bk
%
W 4 [m A R -w%
[AtR
W r-M.t
)T-
imai,
im t± R m yp,.
WLA,4m
, FAM.
nc, itftm Tim
v A. m o, A & ARW Ri
~
i wTh *A, A,
(Aiti
T fo i
R
imo &U,-V ,
mR T) z-%A AM
ItJ.Mi
iu
EJt
6
IWANIMzk44 -
Foreward
The overall aims of the work presented in this thesis have been to i) focus on further
identification of effects of in vitro injurious mechanical compression of cartilage on cellmediated processes and ii) develop models for injurious compression of the cartilage
which seek to incorporate interactions with other joint tissues such as elaborated
inflammatory mediators.
(1) Chapter one: We present an introductory discussion of in vitro mechanical cartilage
injury models, including an overview of insights from previous investigations and initial
results of our ongoing studies of injurious compression in newborn bovine articular
cartilage. The text of the chapter is included as published in Clinical Orthopaedics and
Related Research in 2001.
(2) Chapter two: Chondrocyte cell death after cartilage injury would clearly have
important effects on cartilage matrix repair and maintenance. To further investigate cell
death in our model, we performed a quantitative analysis of cell morphology by electron
microscopy after injurious compression, and directly compared these results to the
TUNEL assay. This analysis confirmed that injurious compression can cause a
significant increase of the rate of apoptotic cell death, but most interestingly, found that
the cell death that occurred was almost entirely attributable to apoptosis.
(3)
Chapter three: A set of studies were
performed to examine proteoglycan
degradation after injurious compression and are included as published in the journal
Arthritis and Rheumatism in 2003. It is well described that mechanically injured cartilage
can release proteoglycans (PG) into the culture medium, but we wanted to investigate
specific mechanisms that might be relevant to the long-term processes that occur in
vivo, such as cell-mediated enzymatic degradation. Chondrocyte gene expression after
mechanical injury had not yet been investigated, so Northern analysis was performed
on mRNA extracted from bovine cartilage over the first 24 hours after injury, and we
observed a ten-fold increase in mRNA for MMP-3 but no change in MMP-1 3 expression.
7
This result raises interesting possibilities in terms of chondrocyte mechanotransduction,
although a full understanding of the effects of mechanical injury alone on cell-mediated
degradation will require further work (see Appendix B for the results of ongoing studies).
In fact, proteoglycan release was significantly increased only for the first three days
after injury, constituted a small percentage of the total cartilage PG content, and was
not affected by the addition of the biosynthesis inhibitor cycloheximide. This suggests
that the proteoglycan release from this model of cartilage injury alone is mostly
attributable to the initial mechanical damage.
We then examined proteoglycan loss from injured bovine cartilage after addition of the
cytokines IL-1 or TNFa, which might be present in synovial fluid in increased levels after
joint injury. The combination of injured cartilage incubated with 1-10 ng/ml IL-1 resulted
in a strong synergistic increase in PG loss over either injury alone or IL-1 treatment
alone, resulting in the loss of 35-60% of the total tissue PG content. A similar synergistic
effect was seen with TNFi and injury. It is possible that this effect may represent a
clinically relevant link between mechanical injury and cartilage degradation. Finally, we
reported our initial results comparing the response of normal human knee and ankle
cartilage from the same donor to the combination of injury and IL-1 treatment.
Interestingly, although much less PG was lost than in bovine tissue, a synergistic PG
loss was observed in knee cartilage but not in ankle cartilage.
(4) Chapter 4:
We were also interested in developing an injury model to investigate
possible interactions with the synovial membrane or joint capsule tissue. Coincubation
of normal joint capsule tissue together with cartilage sharply inhibited cartilage
biosynthesis, and coincubation of injured cartilage produced a synergistic loss of PG.
We thus suspected that a cytokine such as IL-1 was responsible for these effects, and
proceeded to test this hypothesis. Surprisingly, blockade of the combination of IL-1 and
TNF had no effect on the inhibition of cartilage biosynthesis by coincubation with joint
capsule tissue. Our results suggest that the mediator is a protein released from the
capsule tissue, but its identity remains unknown. Placing capsule tissue in culture
results in a number of changes from its in vivo state, but what the tissue is responding
8
to is not clear. This may be relevant to aspects of clinical or surgical trauma to the
capsule as well as to our attempts to develop a model for joint injury that involves more
than one joint tissue.
(5) Chapter 5: Cartilage proteoglycan degradation induced by cytokines such as IL-1 is
known to occur by induction of aggrecanase activity. The experiments described here
investigated the processes involved in activation of ADAMTS-4 (aggrecanase-1) in
cartilage tissue on stimulation with IL-1. On Western analysis of ADAMTS protein forms,
the major changes were the loss of a p75 form from the tissue and the appearance of
increased p75 and p60 in the medium. These changes were reversed by the inhibitor of
GPI anchor synthesis mannosamine, consistent with evidence from cell lines suggesting
that inactive p75 is converted to active p60 by a GPI-anchored proteinase (such as MTMMP4). The techniques refined specifically for cartilage tissue in this study are
expected to allow further investigation of aggrecanase degradation in injured cartilage.
9
Chapter 1
In Vitro Models for Investigation of the Effects of Acute Mechanical
Injury on Cartilage
This chapter is reproduced as published in the journal "Clinical Orthopaedics and
Related Research," number 391S, pp. S61-S71, and co-authored by:
Parth Patwari, MD*; Jakob Fay, BSc*; Michael N. Cook, MS**; Alison M. Badger, PhD**;
Alex J. Kerin, PhD*; Michael W. Lark, PhD**; and Alan J. Grodzinsky, ScD*
*Continuum Electromechanics Group, Center for Biomedical Engineering and
Department of Electrical Engineering and Computer Science, Massachusetts Institute of
Technology, Cambridge MA;
**Musculoskeletal Diseases Department, GlaxoSmithKline, King of Prussia, PA.
10
Abstract
Traumatic injury to a joint is known to increase the risk for the development of
secondary osteoarthritis, but it is unclear how this process occurs. The existence of
such a discrete event that can lead to an increased risk of osteoarthritis has spurred
interest in developing in vitro models of traumatic joint injury. This manuscript reviews
some of the recent insights gained from these model systems into the pathogenesis of
osteoarthritis, including the evidence for an initial, irreversible insult to chondrocytes
during mechanical injury, the occurrence of apoptotic chondrocyte death, and attempts
to identify the effects of trauma on chondrocyte metabolic response. Results are also
presented from the authors' ongoing studies of the degradative pathways initiated by
traumatic mechanical loads, the mechanism by which chondrocytes are affected during
compression, and possible contributions of the joint capsule to posttraumatic cartilage
degradation.
Abbreviations
ELISA: enzyme-linked immunosorbent assay
MMP: matrix metalloproteinase
TIMP: tissue inhibitor of metalloproteinase
Col2CTx: collagen Type 11 C-telopeptide cross-linking domain neoepitope
PG: proteoglycan
TUNEL: terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling
ECM: extracellular matrix
GAG: glycosaminoglycan
mRNA: messenger ribonucleic acid
IL-1: interleukin-1
This research was supported in part by National Institutes of Health grant AR45779, a
grant from GlaxoSmithKline, and a fellowship from the Whitaker Foundation.
11
Introduction
Acute traumatic joint injury is known to increase the risk for subsequent development of
osteoarthritis,'
2
but the mechanisms responsible for this process still are unclear.
Posttraumatic osteoarthritis may account for substantial morbidity in a relatively young
and active population, with many patients showing radiographic signs of osteoarthritis
by 10 years after anterior cruciate ligament or meniscal injury.3 The canine anterior
cruciate ligament transection model has been the most commonly used animal model
for posttraumatic osteoarthritis, in which surgical transection of the ligament followed by
normal
activity leads to osteoarthritis-like joint degeneration. This
model
has
considerably advanced the understanding of how joint instability contributes to
osteoarthritis. In particular, joint laxity after ligament injury may result in greatly
increased repetitive loads with everyday activity.
In addition, there recently has been a resurgence of interest in models simulating in vitro
the mechanical overloading of articular cartilage that occurs during trauma to the joint.
This interest may be partly attributable to the increasing evidence that in addition to
ligament damage and the resulting joint instability, acute joint trauma may cause
immediate, irreversible damage to the cartilage matrix and/or to the chondrocytes which
may contribute to the progression of osteoarthritis. For example, although advances in
ligament repair have made substantial progress in pain relief and return of joint function
for patients after an injury, it still is uncertain whether these repairs ultimately reduce the
risk of subsequent osteoarthritis.4 This manuscript describes the results and limitations
of selected in vivo and in vitro model systems which have been used to investigate the
effects of injurious loading, in the context of data from in vitro injury models currently
being investigated by the authors.
In Vivo Studies of Traumatic Joint Injury
Ground-breaking
insights
concerning
the
molecular
mechanisms
of
cartilage
degeneration in vivo have come from a series of analyses of synovial fluid samples
taken from Swedish patients after an anterior cruciate ligament or meniscal tear.5 7 This
cross-sectional study involved samples obtained from patients at presentation for an
12
injury and at subsequent points as many as 15 years after the injury. Synovial fluid
samples also were taken from healthy athletes, patients with primary osteoarthritis, and
patients with other joint conditions. One finding was that the concentration of
proteoglycan fragments in the synovial fluid was found to be elevated two- to threefold
after injury, and that these levels were similar those found in patients with primary
osteoarthritis.6 The synovial fluid also was analyzed by ELISA for several proteins at
each of the points in time. One of the proteins measured, stromelysin-1 (MMP-3),
represents one of the major pathways for proteoglycan degradation. Synovial fluid
levels of MMP-3 were markedly increased at presentation, to 50-100 times the levels in
healthy athletes, then decrease slowly over a period of months, but remained elevated
for many years (Fig 1).5 Tissue inhibitor of metalloprotease was also elevated, and the
ratio of MMP-3 to TIMP varied considerably over time, but this ratio generally remained
higher than the reference group throughout.6 The levels of free and active protein were
not quantified by these analyses (the ELISA seemed to measure mostly the proenzyme
form of MMP-3). Nevertheless, the results suggest that MMP-3 is upregulated after an
injury and thus may be responsible for some of the proteoglycan loss which
accompanies osteoarthritis progression.
In addition to proteoglycan degradation after traumatic knee injury, the same series of
synovial fluid samples revealed that collagen degradation pathways also seem to be
upregulated.7 In this study, joint fluid was analyzed for the presence of a neoepitope in
the C-telopeptide cross-linking domain of Type 11 collagen. The presence of this
fragment
indicates
digestion
of
mature,
cross-linked
collagen
by
a
matrix
metalloprotease. Synovial fluid Co12CTx was elevated 15-fold over healthy volunteers at
presentation and for several weeks thereafter, and remained significantly higher than
healthy volunteers for a period of 15 years.
In a related study involving quantification of collagen damage after an anterior cruciate
ligament injury, Price et al.8 obtained cartilage biopsies from a low weight bearing site
on the intercondylar notch (n = 17) during arthroscopy after an anterior cruciate ligament
injury. Collagen denaturation in these samples was calculated as the proportion of
13
collagen digested by u-chemotrypsin. Although less than 5% of collagen was denatured
in control cartilage from healthy patients, approximately 10% of collagen was denatured
in normal-appearing cartilage from patients after anterior cruciate ligament injury.
Surprisingly, the amount of denatured collagen in cartilage from patients with late-stage
osteoarthritis at joint replacement did not appear much different from that in patients
after anterior cruciate ligament injury.
Together, these clinical studies suggest that proteoglycan and collagen degradation
rates are altered significantly within days of the injury, and that these changes are not
transient; they remain altered for years. If these degradative processes had been only in
response to altered mechanical loading subsequent to injury, such as joint instability
resulting from ligament injury or subchondral bone remodeling resulting from fracture,
then one might expect that degradation would begin slowly and peak with development
of osteoarthritis. In contrast, these data suggest that the injury itself, and the immediate
response of the joint tissues to the original mechanical insult, may initiate a degradative
process that leads to an increased risk for osteoarthritis even in the absence of
subsequent alterations in the mechanical environment.
This hypothesis is consistent with the early work of Radin et al., 9-1 who found that
impact trauma alone to the patellofemoral joint led to osteoarthritis in animal models.
Radin et al. focused on the relationship between fractures at the bone-cartilage
interface and subsequent endochondral bone thickening and cartilage degradation.
However, a recent animal model has shown that cartilage degradation results even from
impacts that do not seem to cause endochondral bone fractures. In the model of
Newberry et al.,1 a load was dropped on the knee joint of a mature rabbit. Cartilage
shear stiffness was observed to decrease by 3 to 12 months after injury, suggestive of
cartilage matrix degradation. Although the response of bone to the impact and the
subsequent changes in joint stresses clearly are important in osteoarthritis pathology,
this model suggests that injury to the cartilage may be sufficient to initiate osteoarthritislike processes. A potential confounding factor in this study was that since the joint
capsule was opened in order to place pressure-sensitive film, an inflammatory response
14
could have been induced in the sensitive synovium, which could initiate certain
degradative processes.
In Vitro Studies of Mechanically Induced Cartilage Injury
Whereas animal and human studies have focused on injuries to the intact joint, several
in vitro models have focused on understanding the effect of the injurious mechanical
compression primarily on the articular cartilage itself. Although in vitro models cannot
address all the events that occur in response to injury, they allow quantification of
specific events and mechanisms involving the effects of well-defined loading regimens
on cartilage. In addition to describing the effects of injurious compression on the matrix
and the chondrocytes, evidence can be gathered to describe how and why the
chondrocytes respond. One could hypothesize that matrix damage occurs first,
subjecting chondrocytes to increased stresses and/or disrupting normal cell-matrix
interactions or, alternatively, that chondrocytes are directly affected by injurious
compression and respond accordingly.
Although there have been several interesting and important in vitro studies using highamplitude cyclical loading regimens, 1 3
15
this review concentrates on models attempting
to simulate a single-load acute joint injury without fatigue-induced effects. As far as the
authors are aware, the first in vitro experiment in this latter category was the study by
Repo and Finlay,1 6 who used a drop-tower to apply impact loads on cartilage-bone
cylinders cored from the tibial plateau of mature dogs. Using a strain rate of 1000 s-1,
they found that chondrocyte viability, measured by autoradiography, was approximately
0 after one impact to 40% final strain, compared with 100% viability after an impact to
20% final strain. In a more recent study, in which a drop-tower loading system also was
used, Jeffrey et al. 17 similarly observed a decrease in cell viability in bovine articular
cartilage after impact loading.
These two in vitro studies (Repo and Finlay
16
and Jeffrey et al.
17)
of impact loads on
cartilage-bone cylinders highlight the relative merits of such methodologies. As Jeffrey
et al. reported, the presence of bone greatly affected the result of an impact load on the
15
cartilage. With underlying bone attached, much less damage to the cartilage was
observed, and higher-energy impacts caused damage to the bone rather than to the
cartilage. This suggested that the bone acts as a "shock absorber" in vivo, consistent
with the prominence of subchondral fracture in animal impact models."
For this reason,
impact models that isolate cartilage from the underlying bone must not be interpreted as
an exact simulation of the in vivo stress magnitude and distribution. However, such
models allow for controlled mechanical loading of cartilage and thereby enable
mechanistic studies regarding the response of cells, matrix, and cell matrix interactions
to injurious loads. The effects of given loading parameters may be more easily
quantified and correlated with cellular and matrix changes.
For example, Torzilli et al. 18 compressed mature bovine occipital cartilage in a loadcontrolled apparatus at a constant rate of 35 MPa/s to a range of final stress values
from 0.5 to 65 MPa. In this manner, it was found that an increase in matrix damage, as
measured by swelling, and a decrease in proteoglycan synthesis occurred in the range
from 15 to 20 MPa. In addition, by staining the cartilage in situ with fluorochrome dye to
visualize live cells microscopically, cell death was qualitatively found to occur at a
threshold around 17.5 MPa. This work led to the hypothesis that damage to the
cartilage occurs when a threshold in peak stress is reached during an injury. Identifying
such a specific criterion would be important for understanding the etiology of cartilage
damage. However, other researchers have found similar trends for matrix damage, cell
biosynthesis, and cell viability associated with threshold levels of the velocity of injury
(strain rate, rather than the peak stress).1 6'1 9 In general, one must know the peak
stress, final strain, and strain rate to fully define the injurious loading scheme; only two
of these three parameters are independent, and therefore two parameters must be
specified (controlled). For example, the velocity (strain rate) and final strain (or final
peak stress) can be specified, and further studies which quantify the differential effects
of both parameters would be necessary to fully characterize the response of the
cartilage to mechanical injury.
16
In addition to cell viability, radiolabel incorporation has been measured to characterize
the chondrocyte biosynthetic response to loading. Several studies have shown that
radiolabel incorporation in cartilage explants decreased after an injury.18 2 0 However,
because the injury also can cause cell death, an important unanswered question is
whether biosynthetic activity can increase in the remaining viable cells in an attempt to
repair the cartilage matrix. As mentioned previously, Jeffrey et al.17 had measured
chondrocyte viability after impact. They had isolated chondrocytes by collagenase
digestion and then counted viable cells in a hemocytometer by staining with trypan blue.
Using this approach, they then concluded that biosynthetic activity per viable cell had
recovered to control levels by 3 days after an injury. 20 However, it is possible that this
result was attributable to overestimation of cell death, since chondrocytes may remain
viable after injury but be too fragile to survive enzymatic digestion. In contrast, Kurz et
al., 1 9 using in situ tissue explant staining with the fluorescent viability markers ethidium
bromide and fluorescein diacetate, found that biosynthesis per viable cell had not
recovered to control levels by 3 days after injury. A limitation of this method is that it is
difficult to quantify viability after injuries large enough to cause matrix disruption,
because this results in inhomogeneous cell death (near such gross fractures in the
matrix, cell death appears to be 100%). However, the trend for decreased biosynthesis
per viable cell also was apparent in cartilage loaded at rates lower than those required
to cause gross matrix disruption.
Previous studies have shown that low-amplitude dynamic compression can upregulate
chondrocyte biosynthesis of proteoglycans and proteins in normal mature and immature
cartilage explants.
3
This raises the possibility that such stimulatory dynamic
compression could be used to initiate cellular repair processes after cartilage injury.
Kurz et al. 19 investigated this hypothesis by measuring the response to dynamic
compression after an injury to bovine calf explants (Fig 2). Cartilage disks were loaded
at 3 different strain rates to 50% final strain and held there for 5 minutes. Three days
after the injury, the explants were radiolabeled with 3 H-proline and
35
S-sulfate for 12
hours while being either compressed dynamically (3% dynamic strain amplitude at 0.1
Hz, superimposed on a 10% static offset compression), or compressed to the static
17
offset alone. The data shown are normalized to the incorporation of disks which
received only the static offset compression. Control disks received no injury before
being dynamically loaded.
Sulfate incorporation (as a measure of proteoglycan synthesis) in uninjured cartilage
was stimulated by dynamic compression, as seen previously. However, in cartilage that
was mechanically injured by compression at 1 s1 to 50% strain, incorporation was
decreased, compared with cartilage injured at the same rate but not subjected to
dynamic loading afterwards. The cells in mechanically injured cartilage were not able to
respond to reparative dynamic compression stimuli. A major advantage of measuring
biosynthetic response to dynamic loading is that comparisons are made to cartilage
which also was injured but not dynamically loaded. Therefore, cell death due to the
injury was the same in both groups and not a confounding variable. The loss of the
normal stimulation of biosynthesis by dynamic loading may represent another important
contribution to the cycle of degradation in osteoarthritis cartilage.
As the experiments above make clear, cell death after injurious compression may
decrease the repair potential of cartilage. However, the recent identification of apoptotic
cell death in human osteoarthritic cartilage2 4
25
has led to an additional hypothesis.
Some studies have reported that osteoarthritic chondrocytes may undergo inappropriate
terminal differentiation; that is, they express markers for the pathway leading to
endochondral bone formation, die by apoptosis, and promote mineralization. 26 ,2 7 It also
has been
apoptosis.
reported that cell death after injurious compression can
28,29
occur by
the8
In the study by Loening et al., 2 8 cartilage explants from newborn bovine
calves were subjected to a loading regimen of compression to a final strain of 30% to
50% at a strain rate of 1 mm/second, and held for 30 minutes, with this loading repeated
six times. Apoptosis was assessed 4 days after injury by evaluation of nuclear
morphology and by TUNEL staining. Apoptotic cell death was found to increase in a
dose-dependent fashion with the peak stress attained during the injury (Fig 3). In
addition, significantly more sections with apoptotic nuclei were found in cartilage
compressed to 30% final strain (approximately 4.5 MPa peak stress) than in control
18
cartilage. This result is interesting because at this level of compression, no macroscopic
damage to the cartilage ECM was detected, as assessed by measuring tissue swelling
and GAG loss, suggesting that apoptosis can be directly triggered by loading injury to
the chondrocytes. This observation is significant not only for understanding of how
cartilage responds to injury, but also for possible development of pharmacological
intervention. Although it is not yet clear whether apoptosis promotes inflammation and
degradation, this is an interesting hypothesis for additional investigation.
Acute Mechanical Injury Increases MMP-3 Gene Expression
Based in part on the experiments and results shown in Figures 1 and 3, it was
hypothesized that injurious mechanical compression might alter the expression of
proteinases such as MMP-3 that are related to matrix degradation and cellular
apoptosis. An experimental protocol therefore was established to test this possibility.
Tissue Harvest:
Knees were obtained from 2-to-3-week-old calves on the day of
slaughter from a local abbatoir. Nine-mm diameter cylindrical cores of cartilage on bone
then were obtained from the femoropatellar groove with a drill press, placed in a holder,
and cut with a microtome into 1-mm thick slices. Finally, the slices were punched to
obtain 3-mm diameter by 1-mm thick cartilage disks. Culture: Cartilage disks were then
incubated in 0.25 mL media per disk at 370 C and 5% CO 2 . Medium was low-glucose
Dulbecco's modified Eagle medium (Gibco, Grand Island, NY), with 10% fetal bovine
serum (HyClone, Logan, UT), supplemented with amino acids, ascorbate, and
antibiotics. Injurious Compression: Injurious compression took place 2 to 3 days after
harvest.
An
incubator-housed
loading
apparatus that operates
displacement control was used to compress the cartilage disks.3 0
under load
or
The 1-mm thick
cartilage was compressed 500 pm (a final strain of 50%) at a velocity of 1 mm/second
(a strain rate of 1/second), and held for 5 seconds. Control plugs were kept in freeswelling culture conditions and received no compression. mRNA isolation:
cartilage disks were frozen in liquid nitrogen at 1,
Fifteen
3, 6, 12, and 24 hours after
compression (three at each time). Fifteen control disks, cultured in free-swelling
conditions without any compression, were frozen at the same times. The disks were
pooled, and the mRNA was extracted using methods previously described. 1
19
Specific
probes to bovine MMP-3 and MMP-13 were used to quantify mRNA expression by
Northern blotting. Hybridization was normalized to elongation factor-1a expression.
In these specimens, injurious compression of the cartilage disks to 50% strain at a
strain rate of 1 s- produced peak stresses of 11.5 ± 0.7 MPa (mean ± standard error of
the mean). In the injured cartilage, MMP-3 (stromelysin-1) mRNA expression was
increased approximately 10-fold over control cartilage (Fig 4). By comparison, no
change in the amount of MMP-13 (collagenase-3) mRNA was detected. The finding of
increased MMP-3 expression after injury of cartilage from bovine calves is perhaps
related to the report by Martin et al. 32 that levels of MMP-3 proenzyme in normal human
cartilage were increased after a mechanical loading regimen (cyclical compression at
0.5 Hz to 20% strain for 2 hours). This result also correlates well with the increased
levels of MMP-3 seen in synovial fluid of patients soon after a traumatic knee injury, as
described previously (Fig 1). In addition, Chubinskaya et al., 3 3 using in situ hybridization,
found a similar pattern of upregulation of MMP-3 and MMP-8 message, but not MMP13, in damaged human ankle and knee cartilage compared with normal controls.
However, Lark et al. 34 found the presence of MMP-3 activity (by in situ hybridization for
the VDIPEN 34 1 neoepitope of MMP-degraded aggrecan) in normal human cartilage and
in cartilage from patients who had osteoarthritis and rheumatoid arthritis. Together,
these data suggest that the presence and activity of these proteinases in acute
posttraumatic joint injuries may be different from that in long-term osteoarthritis disease,
and may highlight the role of the initial mechanical injury.
These investigations are part of efforts by many researchers to understand how various
degradative processes are involved in the PG loss seen in osteoarthritic cartilage.
Attempts to additionally characterize the pathways involved in GAG degradation in this
system are ongoing. In particular, quantifying the relative amount of MMP-generated
and aggrecanase-generated GAG cleavages would be important for identifying the
degradative pathway which is activated by mechanical cartilage trauma.
20
Cartilage Matrix Degradation in Response to Acute Mechanical Injury
Previous studies have documented the effects of injurious impact loads on GAG loss
and matrix damage.1 3 ,1 4 , 17 ,18, 2 8,35
In order to correlate such aspects of matrix
degradation with injurious strain, strain rate and/or peak stress, we have begun to look
systematically at these mechanical variables in short and long-term injury studies. An
example can be seen in the immediate effects of injurious loading on collagen damage,
as manifested in tissue swelling. Injurious Compression:
Cartilage plugs were loaded
as described above but the strain rates and final strains were varied within selected
ranges. Control cartilage remained in free-swelling culture and was not compressed.
Swelling Measurements: The wet weight of each cartilage disk was measured before
injurious compression and again 24 hours after injury. The cartilage disks then were
placed in a hypotonic solution of 0.01 mol/L NaCl for 2 hours and weighed again.36 The
total swelling after injury was calculated as the difference between the wet weight
measured after hypotonic swelling and the wet weight measured before compression.
The water content of cartilage tissue is in a state of constant balance between two
opposing forces:
the swelling pressure resulting from the electrostatic repulsion
between the charged proteoglycans is opposed by the collagen network.3 7 As a result,
measuring swelling of the tissue after injury allows one to quantify the extent of damage
to the collagen network, assuming the loss of PG is not too extensive. Incubation in
hypotonic saline additionally amplifies the swelling that would result from damage to the
tissue's collagen network.36' 37
After culture for 2 hours in 0.01 mol/L NaCl, control cartilage swelled approximately 5%
because of the increased electrostatic repulsion forces in solutions of lower ionic
strength (Fig 5). After injurious compression at a given strain rate, total swelling
increased with final strain in a dose-dependent fashion. In addition, with the exception of
the explants compressed at 0.1 s-1 to 30%, total swelling after injury at a given final
strain increases with strain rate.
21
This experiment shows an approach for studying the effects of mechanical injury on
cellular response and matrix degradation in vitro. One possible mechanism involves
acute damage to matrix, which then is sensed by the cells as an alteration in their
biomechanical or physicochemical environment. However, it also is possible that cells
may be affected directly by mechanical injury in the absence of matrix damage, as
would be suggested by the occurrence of apoptotic cells below the threshold for matrix
damage.
Such cellular injury may lead to subsequent matrix degradation that might
occur, for example, after upregulation of proteinase expression by the cells. The
techniques now are in place to confirm such hypotheses, and to correlate these
injurious responses to the parameters of strain, strain rate, or peak stress.
Effects of Acute Mechanical Injury to Cartilage using a Cartilage and Joint
Capsule Coincubation System
The authors recently have been exploring a novel in vitro injury model that combines
aspects of damage to cartilage in the presence of synovial tissue that might normally be
part of the response to injury in an intact joint. Tissue Harvest: In addition to harvest of
cartilage explant disks as described above, portions of joint capsule were cut medial
and lateral to the femoropatellar groove during dissection of the joint. The joint capsule
then was punched to form a 6-mm diameter disk, approximately 2-mm thick, and placed
in culture medium. Design: Cartilage disks received one of four treatments in a factorial
design to examine the effect of injury and coculture with joint capsule tissue: (1) Control
cartilage was cultured without joint capsule in free-swelling conditions; (2) Three days
after harvest, cartilage disks were injured by compression to 50% strain at a strain rate
of 1 S-,
as described above;
(3) A third group of uncompressed cartilage disks was
placed in coculture with the joint capsule tissue 3 days after harvest; (4) The final group
of cartilage explants was injured and then placed in coculture with joint capsule.
Radiolabel Incorporation: Six days after the intervention (injury or coculture), cartilage
disks were incubated for 6 hours with
35
S-sulfate
and 3H-proline (capsule tissue was
removed from coculture groups before radiolabeling). The disks then were washed of
unincorporated label, digested, and analyzed by scintillation counter for radiolabel
incorporation.
22
Six days after injurious compression,
sulfate incorporation was decreased
by
approximately 25% compared with uninjured controls (Fig 6). However, after incubation
with joint capsule tissue alone for 6 days, sulfate incorporation was decreased in
uncompressed explants by approximately 65% compared with controls that were not
cocultured. The combination of injurious compression and incubation with joint capsule
seemed to have an additive effect causing a further decrease in incorporation. Similar
trends were seen for proline incorporation.
These results, showing that a marked decrease in biosynthetic activity occurs after
incubation of cartilage with joint capsule alone, are consistent with the results of prior
cartilage-joint capsule coculture systems. 3 8' 3 9 Cytokines such as IL-1 are known to
cause such a marked decrease in PG synthesis. It may be that excision of the joint
capsule promoted release of cytokines by the synovium into the culture medium. The
observation that excised joint capsule markedly reduces radiolabel incorporation in vitro
suggests that the co-culture system might mimic certain aspects of intact joint injury.
That is, the joint capsule may be even more sensitive to mechanical trauma than the
cartilage, and that an inflammatory response involving the synovium may result,
releasing IL-1 and other catabolic cytokines into the synovial fluid. However, it also may
be true that excised joint capsule may behave differently from in vivo capsule. The level
of cytokines released from the excised synovium probably is markedly higher than in
vivo, and removal of the synovium from its normal mechanical environment and blood
supply may additionally modulate its response.
Although work to characterize the response of this in vitro system is ongoing, the initial
results are consistent with the hypothesis that the joint capsule, and synovium in
particular, may play an important role in cartilage degradation after acute mechanical
injury, and in the progression to osteoarthritis. It is well-established that an inflamed
synovium is a prominent feature of rheumatoid arthritis. 40 However, it also has been
shown by in situ hybridization that osteoarthritic synovium also produces cytokines and
degradative enzymes.1
It seems likely that the synovium also contributes to the
23
degradative pathways, such as MMP-3, seen to be activated after joint trauma. This
could be by production of MMPs and aggrecanases in capsule tissue itself,43 or by
release of cytokines that stimulate chondrocyte-mediated degradation.
The response of cartilage to mechanical injury and the pathogenesis of osteoarthritis
are complex and multifactorial processes. It has become increasingly appreciated that
osteoarthritis is a disease of the entire joint which it affects; the subchondral bone,
articular cartilage, joint capsule, ligaments, and nervous system all may interact to play
a role in the development of osteoarthritis. However, the events surrounding the
initiation of osteoarthritis remain unclear. The clinical data and results from impact
loading models now suggest that the initial insult to the articular cartilage may be
important and irreversible, but it still seems likely that an interplay between the cartilage
and the other joint tissues contributes to the initial response. It may be that multiple
events, chondrocyte apoptosis, subchondral microcracks, or joint instability, may each
be sufficient for initiating the interplay among joint tissues that leads to osteoarthritis. In
any case, in vivo and in vitro models of acute mechanical injury such as anterior
cruciate ligament transection and impact loading probably will continue to play an
important role in dissecting the various events that initiate cartilage damage caused by
injurious compression and the relation between this damage and ultimate progression to
osteoarthritis.
24
100 -N
=347
0
E
C
75
0
C
4--
c5
0
C
C?
25-
0
' ' ' '' "
' ' ' ' ' ' ''
1
10
'
' ' 'I'
100
1000
Time after injury (weeks)
Figure 1. Temporal pattern of MMP-3 concentration in knee synovial fluid after injury.
Synovial fluid was collected in a cross-sectional study of patients at arthroscopy or at
presentation to the emergency room. These patients had a tear of the anterior
cruciate ligament or the meniscus. The data points are shown as mean ± standard
error and the shaded region represents the standard error of the MMP-3
concentration in the reference group of healthy athletes. Reproduced by permission
of the Journal of Rheumatology from Lohmander et al. 5
25
1.5
.
ElSulfate
IProline
E
(N =7-8)
cW 1.25
0
0
........
... ....
... .............
C,)
0.75
Control
0.01
0.1
1
Strain rate (1/second)
Figure 2. Response to dynamic compression after injurious compression. Three days
after loading, cartilage disks were radiolabeled for 12 hours while being subjected to
moderate-amplitude dynamic compression. The data are normalized to injured disks
receiving the same static offset as the dynamical ly-com pressed disks (dotted line)
and are shown as mean ± standard error. Results of ANOVA for proline: p < 0.05;
ANOVA for sulfate: not significant; * represents p < 0.05 versus dynamically
stimulated, uninjured controls by Dunnett. Reproduced by permission of Elsevier
Science from Kurz et al., 19 © 2001 Orthopaedic Research Society.
26
100
n=6
80-
0
S60-
D.
0C)
0
*
20
0-
Control
6 MPa
10 MPa
20 MPa
Figure 3. Apoptosis induced by injurious compression. Bovine cartilage explant disks
were subjected to an injurious compression protocol and analyzed 4 days after injury
for apoptotic nuclei by TUNEL staining. The groups are identified by the peak
compressive stresses produced during loading. Data are shown as mean ± standard
error; stars represent p < 0.05 by paired t-test. Reproduced from Loening et al., 28@
2000 Academic Press.
27
0
(n
C/.
10
xw
z
Z
QY
E
5-
N
E
0
Z
0
MMP-3
MMP-13
Figure 4. Messenger RNA levels of MMP-3 after injurious compression. Bovine
cartilage explant disks were compressed at 1 mm/second (strain rate of 1 s1) to 50%
strain and frozen in liquid nitrogen at 1, 3, 6, 12, and 24 hours after the injury. Control
disks received no compression and were frozen at the same times. The disks in each
group were pooled for mRNA extraction. The MMP-3 expression level was anlayzed
by Northern blotting and normalized to expression of elongation factor-1
28
N = 4 -5
30
(Control: N = 10)
S25)
9
20 15
10T
0
ce*
e,
4
1
QV
Figure 5. Swelling of cartilage disks after injurious compression. Each bar is labeled
with the strain rate and final strain of compression. Twenty-four hours after the injury,
disks were placed in hypotonic (0.01 mol/L) NaCl solution for 2 hours. The total
swelling was calculated as the wet weight measured after removal from the hypotonic
solution minus the wet weight measured before injury. The data are shown as mean
± standard error.
29
1.0
I
N Proline
E Sulfate
0
N = 4 -5
0.8 -
0.
L-
I
0
0.5 -
I
N
E
L-
0.3 -
0
z
0.0 Injury
Joint Capsule
Injury plus
Joint Capsule
Figure 6. Radiolabel incorporation after injurious compression in a cartilage-joint
capsule coculture experiment. Bovine cartilage explant disks were assigned to
receive injurious compression (Injury), be cultured in the presence of joint capsule
(Joint Capsule), or both (Injury plus Joint Capsule). Control disks were incubated
without intervention. Six days after the treatment was applied, radiolabel
incorporation was measured. The data are normalized to the incorporation of control
disks and shown as mean ± standard error.
30
Chapter 2
Quantitative investigation of cell death after injurious compression of
bovine calf articular cartilage by electron microscopy
This chapter is as prepared for journal submission, to be co-authored by:
P. Patwari*, V. Gaschent, I. E. James*, E. Bergert, S. M. Blake*, M. W. Lark, A. J.
Grodzinsky* and E. B. Hunzikert
*Continuum Electromechanics Group, Center for Biomedical Engineering,
Massachusetts Institute of Technology, Cambridge, MA;
tITI Research Institute for Dental and Skeletal Biology, University of Bern, Switzerland;
tDepartment of Musculoskeletal Biology, GlaxoSmithKline, King of Prussia, PA
31
Abstract
It has been suggested that chondrocyte death by apoptosis may play a role in the
pathogenesis of cartilage destruction in osteoarthritis, but the results of in vivo and in
vitro investigations have been conflicting. To investigate further the cell death in our in
vitro model for traumatic joint injury, we performed a quantitative analysis by electron
microscopy of cell morphology after injurious compression. For comparison, the TUNEL
assay was also performed. Articular cartilage explant disks were harvested from
newborn calf femoropatellar groove. The disks were subjected to injurious compression
(50% strain at a strain rate of 100%/s), incubated for 3 days, and the fixed for histology.
By TUNEL, the cell apoptosis rate increased from 7 ± 2 % in unloaded controls to 33 ± 6
% after injury (N = 8, p = 0.01). By EM, the apoptosis rate from 5 ± 1 % in unloaded
controls to 62 ± 10 % in injured cartilage (N = 5, p = 0.02). Surprisingly, analysis by EM
also identified that of the dead cells in injured disks, 97% were apoptotic by morphology.
These results confirm a significant increase in cell death after injurious compression and
suggest that most cell death observed here was by an apoptotic process.
32
Introduction
In osteoarthritis, or degenerative joint disease, the factors leading to the degeneration of
the cartilage are not well understood. The chondrocytes are the cells responsible for the
maintenance of the cartilage extracellular matrix, which provides the mechanical
integrity and function of the cartilage tissue. Loss of chondrocytes from the tissue, seen
as the presence of empty lacunae on histology, has long been observed in late-stage
osteoarthritic cartilage. The cause is not known, but as mature chondrocytes have
limited capacity to repopulate the cartilage,44 cell death may have permanent effects on
the ability of the tissue to repair and maintain its matrix.
Recently, it has been proposed that cell death by apoptosis may be an important event
in osteoarthritic cartilage. Several investigators have reported increased rates of
apoptosis in OA cartilage, using the TUNEL staining method.2 4 26 It has further been
proposed that apoptotic cell death is part of a process which promotes mineralization of
the cartilage, and thus may also have direct contributions to the pathogenesis of
OA.26 ,2 7 However, the relative importance of this process has remained controversial, as
other investigators have failed to confirm the finding of large numbers of apoptotic cells
in OA cartilage.4 5
Researchers investigating cell death with in vitro cartilage experiments have since
emphasized that false-positive staining by the TUNEL assay can be a major limitation
(45-48,
and recently reviewed49). In particular, Chen et al. 48 found that cells in cartilage
subjected to freeze-thaw cycles were 90% TUNEL positive after three days of culture,
suggesting that TUNEL is not reliable for distinguishing apoptotic from necrotic cell
death. In addition, it is increasingly clear that modes of cell death with features of both
necrosis and apoptosis exist, and that the DNA fragments, which the TUNEL stain
measures, may not give a complete picture of the processes occurring.50'51
In vitro models of mechanical injury to the cartilage have also been used in order to
investigate chondrocyte cell death. Such models are motivated by the observation that
33
traumatic joint injury is a risk factor for the subsequent development of OA, and have
been used to investigate the effects of mechanical injury on the cartilage tissue and
cells in order to shed light on the events which may lead to the development of OA in
vivo. In these in vitro models, we and others have shown that one effect of mechanical
injury is an increase in apoptotic cell death, as assessed by TUNEL staining and by
nuclear morphology on light microscopy.2 8 5 2
In light of the questions regarding the interpretation of the TUNEL assay, we undertook
a study to examine and quantify cell death by electron microscopy (EM) in our model for
mechanical cartilage injury. Electron microscopy is the standard for morphological
assessment of cell apoptosis, and has been used for confirmation of apoptosis in
previous studies of chondrocyte cell death, but not for quantitative assessment. The
aims of this study were therefore to i) quantify the relative contributions of apoptotic and
necrotic cell death after injurious compression in newborn bovine cartilage by EM and ii)
directly compare the results of the EM analysis with the TUNEL assay.
Methods
Cartilage Harvest
Newborn bovine articular cartilage explant disks were obtained from the femoropatellar
grooves of 1 to 2-week-old calves, obtained from a local abattoir (Research '87,
Hopkinton, MA) on the day of slaughter, as previously described.2 3 In brief, cartilagebone cylinders (9 mm in diameter) were drilled perpendicular to the cartilage surface
and placed in a microtome holder. After creating a level surface by removal of the most
superficial -100 pm, the next 2 mm of cartilage were sliced by a microtome, producing
two 1-mm-thick slices.
Finally, four explant disks were punched out of each slice,
resulting in cartilage disks that were 1 mm thick and 3 mm in diameter. Cartilage was
then left in culture to equilibrate for 3 days in medium (low-glucose Dulbecco's Modified
Eagle's Medium [DMEM], supplemented with 10% fetal bovine serum, 10 mM HEPES
buffer, 0.1 mM nonessential amino acids, 0.4 mM proline, 20 pg/ml ascorbic acid, 100
34
U/ml penicillin G, 100 pg/ml streptomycin, and 0.25 pg/mI amphotericin B), in a 37
0C,
5% CO2 environment.
Injurious compression
Injurious compression was performed in a custom-designed incubator-housed loading
apparatus. 3 0 Cartilage disks were placed, one at a time, in a well in the center of a
polysulfone chamber that allowed for unconfined compression of the disk. The lid of the
chamber is attached to a load cell, and the displacement of the lid (equal to the
displacement of the cartilage surface) is measured using a linear-voltage differential
transducer. The thickness at the time of injury was measured and recorded so that the
zero-strain position coincided precisely where contact was made between the chamber
lid and the top surface of the cartilage. The loading protocol applied during injurious
compression was identical to that used in a prior report53 and consisted of a ramp
compression at a velocity of 1 mm/s (a strain rate of 1 s-) to a final strain of 50%,
followed by release.
TUNEL staining
Cartilage was flash-frozen in liquid nitrogen and sectioned. TUNEL staining was
performed as reported previously.2 8 At least 200 cells were assessed on each section
for apoptotic cell death by TUNEL staining.
Electron microscopy
Cartilage was fixed in 5% glutaraldehyde (v/v) in 0.1 M sodium cacodylate buffer. The
tissue was post-fixed for EM analysis in a solution of 1%
(w/w) osmium tetroxide and
then dehydrated in ethanol and embedded in Epon 812. Thin sections (65 nm) were
then cut and stained with uranyl acetate and lead citrate. The entire section was
systematically sampled by scanning the section for centrally-cut cells (i.e., cells
sectioned through the nucleus). From each section, approximately 100 cells from the
central and edge portions of the section were classified by morphology as either normal,
necrotic, apoptotic, or unknown. The classification criteria were developed from the
35
literature and from the results of preliminary studies in newborn bovine tissue5 4 . Details
of these criteria are summarized in Table 1.
Experimental Design
From each animal joint used, two cartilage explants (punched from the same cartilage
slice) were used in this experiment. One explant was used as a free-swelling control,
and the other was subjected to injurious compression. In addition, as a control for
necrotic cell death, three cartilage disks were subjected to freeze-thaw cycles (flashfreeze in liquid nitrogen followed by thawing in a 60 C water bath, for three cycles).
After injury or freeze-thaw, the cartilage explants were replaced in culture for an
additional three days.
Each explant was then removed from culture, and cut in half
(across a diameter, forming a semicircle). From each explant, one half was processed
for TUNEL staining, and the other half was processed for electron microscopic analysis.
Statistical Analysis
All descriptive statistics for results are given in the text as mean ± standard error and
shown in figures as box plots (brackets show range, bars show interquartiles, and the
white line shows the median). Differences between two population means were tested
with the non-parametric Wilcoxon signed-rank test (SPlus, MathSoft Inc.; now Insightful
Corp., Seattle WA).
Results
Injurious Compression
Cartilage was harvested from eight different bovine femoropatellar grooves. After three
days of equilibration in culture, cartilage explants were subjected to injurious
compression. The actual thickness of the disks immediately before injury was 1.03 ±
0.02 mm (mean ± SEM), reflecting a small amount of swelling during culture.
Compression of the cartilage to 50% strain at 1 mm/s produced peak stresses of 20 ± 1
MPa. On visual inspection after injury, cartilage appeared permanently deformed to an
elliptical shape in 3/8 disks and disrupted (fissured) in 1/8 disks.
36
Cellular Morphology on Electron Microscopy
After fixation for electron microscopy, it was determined that three unloaded and three
loaded samples were inadequately fixed and were excluded from the analysis.
Examples of cellular morphology for cells classed as normal, necrotic, and apoptotic are
shown (Fig.
1).
In freeze-thawed cartilage, necrotic cells (Fig. 1,
D-F) were
characterized by amorphous granular debris throughout the lacunae, including at the
matrix border, and general lack of membrane integrity. In contrast, apoptotic cells (Fig.
1, G-1) were shrunken and clearly retracted from the surrounding matrix. In general, the
most frequently observed ultrastructural features of apoptotic cells observed here were
the presence of nuclear blebbing, apoptotic bodies, and cell shrinkage. Infrequently
observed features were an intensified staining of the cytoplasm, plasmalemmal
blebbing, and condensation of the chromatin. To show these ultrastructural features
more clearly,
additional
micrographs of apoptotic cells are
shown
at higher
magnification in Fig. 2.
Cell Death in Freeze-Thawed Cartilage
In cartilage subjected to repeated freeze-thaw cycles and fixed three days later, a mean
of 65% of cells stained positive for TUNEL (Table 2).
In contrast, a mean of 97%
(128/132) of the cells were classified as necrotic on electron microscopic analysis of
cellular morphology.
Apoptosis after Injurious Compression
Unloaded and injured cartilage explants were fixed three days later and analyzed by
both TUNEL and EM (Fig. 3). The TUNEL assay showed a significant increase in the
percentage of apoptotic cells, from 7 ± 2 % in unloaded controls to 33 ± 6 % in injured
cartilage (N = 8, p = 0.01). The EM analysis of the second half of the same cartilage
explants also showed a significant increase in apoptotic cells, from 5 ± 1 % in unloaded
controls to 62 ± 10 % in injured cartilage (N = 5, p = 0.02). The TUNEL and EM
analyses of the same cartilage disks were also compared directly (Fig. 4). In all injured
cartilage samples, apoptosis rates were higher by EM than by TUNEL (N = 5, p =
37
0.045). The correlation coefficient between the two measures in injured cartilage was
good (p = 0.67) but the linear relationship between TUNEL and EM measurements did
not reach statistical significance (p = 0.06) due to the small sample size.
Cell Morphology after Injurious Compression by Electron Microscopy
In the five injured cartilage samples, approximately 100 cells in each sample were
examined by EM. On average, 35% were normal in morphology. Of the remaining cells,
94% (range: 84-100%, N = 5) were classified as apoptotic, 1%
(range: 0-5%) were
classified as necrotic, and 5% (range: 0-11 %) were left unclassified because the correct
classification was not clear.
Discussion
We demonstrate here the results of the first quantitative investigation of chondrocyte
cell death after injurious compression by cell morphology on electron microscopic
analysis.
In addition to supporting previous evidence that injurious compression can
result in a dramatic increase in apoptotic chondrocyte cell death, we show that
unexpectedly, apoptotic cells represented almost all of the total cell death. Unlike the
TUNEL staining assay, analysis of cellular morphology did not produce false positives
for apoptosis in freeze-thawed cartilage.
We have previously reported a significant increase in apoptotic cell death after injurious
compression using TUNEL staining and nuclear morphology on light microscopy 28.
Other researchers 4 85, 2,55 56
, have also shown apoptotic cell death in models which cause
several different kinds of mechanical cartilage damage or injury, and confirmed the
presence of apoptosis by electron microscopy. However, quantification of apoptosis has
generally relied on the TUNEL assay. Multiple concerns regarding the use of TUNEL
staining have been reported. First, TUNEL staining can be nonspecific, staining freezethawed cartilage 48 -and hypertrophic chondrocytes. 45 47 Second, false-positives have
been reported as an artifact of histological sectioning.5 7 Third, the absolute numbers
produced by the assay may be unreliable due to the sensitivity of the assay to details of
38
the staining protocol. 45 Analysis of freeze-thawed cells in this experiment confirmed that
necrotic cells were frequently stained by TUNEL, whereas morphological analysis of the
same sample confirmed that nearly all cell death was by necrosis.
Surprisingly, the results of quantitative electron microscopic analysis here not only
confirmed the increase in cell death after injury but also found more cell death than the
TUNEL assay. The high level of cell death seen here is probably related to the young
age of the newborn calves from which the tissue was taken. Lemke et al.58 have shown
that with increasing age of the animal, there is a significant decrease in the level of
apoptosis induced by injury. They injured cartilage at the same strain and strain rate
used here (50% strain at 1 mm/s velocity) and found cell apoptosis rates decreased
from 22% in cartilage from two-week-old calves to -5% cartilage from animals over 6
months old. Nevertheless, the increase in apoptotic cell death after injury remained
statistically significant in cartilage from the skeletally mature animals. Similarly, Tew et
al.56 also report a decrease in apoptotic response with age at the margins of cut bovine
articular cartilage. The explanation is unclear but newborn calf cartilage chondrocytes
are in general more metabolically active and more responsive to a variety of stimuli than
chondrocytes in adult cartilage. In addition to the age of the tissue, our in vitro injury
model does not attempt to simulate the precise forces and deformations that cartilage
would experience in a clinical joint injury. For these reasons, the absolute rate of cell
death in vitro should not be extrapolated to other situations.
The finding that cell death after mechanical injury can occur predominantly by apoptosis
is interesting from mechanistic and therapeutic perspectives. In addition to cartilage, it
has also been reported in diverse settings that mechanical forces can cause cell
apoptosis (e.g., elevated hydrostatic pressure in glaucoma, 59 decreased lumen flow in
vascular endothelium, 60 and traumatic brain50 and spinal cord 61 injury). However,
despite the progress in detailed understanding of the signaling pathways and the
molecular machinery responsible for apoptosis, little is known of the mechanism for
mechanical induction of apoptosis. Trauma-induced alterations in cell-ECM interactions
is likely to play a role.6 2 In the case of traumatic cartilage injury, the mechanism for
39
programmed cell death after this discrete event could lead to a potential target for
therapeutic intervention.
Some aspects of the chondrocyte morphology we observed in apoptotic chondrocytes
differed from earlier descriptions. In particular, chromatin condensation was not
prominent in our samples, and the lack of this classic feature of apoptosis morphology
appears
to
be
consistent
with
previous
reported
images
of
apoptotic
, 6, 3 The most specific morphological changes seen here seemed
chondrocytes.6 ,4 85, 2,55 56
to be cell shrinkage and nuclear budding. Some differences in the processes involved
in cell death and the resulting morphology may be explained by unique characteristics
of cartilage tissue. Non-programmed cell death in most tissues is classically caused by
ischemia and results in swelling of the cells and infiltration of inflammatory cells to the
tissue. But adult cartilage does not contain a blood supply, does not recruit neutrophils,
and dead cells are partially restrained from swelling by their surrounding ECM. Thus of
coagulation necrosis, ischemic necrosis, and oncosis, none seem to be a perfect term
for description of non-programmed cell death in this tissue. An investigation of
apoptotic/necrotic cell morphology in death caused by other stimuli might be interesting
as well to help characterize this issue in chondrocytes.
40
Necrosis
Apoptosis
*
*
*
*
*
*
Nuclear blebbing
Presence of apoptotic bodies
Cell shrinkage and retraction
Intense staining of the cytoplasm
Budding of the cell membrane
Condensation of the chromatin
0
0
0
0
Cell swelling
Lack of an intact cell membrane
Disintegration or ruptures of the
intracellular organelles
Dispersed nuclear chromatin
Table 1. Morphological criteria for cell death classification
TUNEL Assay
Electron Microscopy
Apoptosis
65%
4/132 (3%)
Necrosis
n/a
128/132 (97%
Table 2. Cell death classification in freeze-thawed cartilage (% of total cells)
41
B
C
D
E
F
G
H
Figure 1. Typical examples of cell morphology on electron microscopy are shown for
normal cells (A-C, cartilage was left unloaded), necrotic cells (D-F, from cartilage
subjected to freeze-thaw cycles), and apoptotic cells (G-1, cartilage was subjected to
injurious compression to 50% strain at 1 mm/s). Tissue was fixed for EM analysis
three days after the intervention. Images are shown at 1700x magnification.
42
Figure 2. Figure 2.
Higher-resolution electron micrographs of apoptotic
chondrocytes (A-D) undergoing nuclear blebbing (G) and containing apoptotic bodies
(arrows). These two features were regularly encountered in apoptotic cells. N =
nucleus; scale bar = 1 pm.
43
100
1001
80
80
0cc
-17
0-
60
S60
40
0 40 -
20-
<~
*0
0u
U,
0
0
0.
Ctrl.
20
0
Injury
Ctrl.
A. TUNEL Assay
Injury
B. Electron Microscopy
Figure 3. Quantification of cell apoptosis after injurious compression by TUNEL and
by electron microscopy. Three days after injurious compression was applied, control
and injured cartilage disks were sliced in half. From each disk, one half was flashfrozen in preparation for analysis by TUNEL staining (A), and the other half fixed for
EM analysis (B). Box plots display range (brackets), interquartiles (solid bars), and
median (white stripes). Apoptosis rates were significantly higher in injured cartilage
as assessed by either TUNEL staining (p <0.05, N = 8) or by EM (p<0.05, N = 5).
44
10080
60 -
*
40-
0
0.
* Injury
0
4i
> No Load
20-
0
0
20
40
60
80
Apoptosis Rate by TUNEL (%)
Figure 4. Apoptosis rates measured by TUNEL compared to measurement by
electron microscopy on the same cartilage explants after three days of free-swelling
culture (No Load) or after injurious compression (Injury). Correlation between the two
measurements was good (rho = 0.67) but did not reach statistical significance in
injured cartilage (p = 0.06; N = 5 cartilage disks).
45
Chapter 3
Proteoglycan Degradation after Injurious Compression of Bovine and
Human Articular Cartilage in Vitro: Interaction with Exogenous
Cytokines
This chapter is reproduced from a manuscript that was published in the journal "Arthritis
and Rheumatism" in May 2003 in volume 48, issue 5, pages 1292-1301, and coauthored by:
Parth Patwari, 1 Michael N. Cook,2 Michael A. DiMicco,l Simon M. Blake,2 Ian E.
James,2 Sanjay Kumar,2 Ada A. Cole,3 Michael W. Lark,2 and Alan J. Grodzinskyl
1Parth
Patwari, MD, Michael A. DiMicco, PhD, Alan J. Grodzinsky, ScD: Center for
Biomedical Engineering, Massachusetts Institute of Technology, Cambridge,
Massachusetts;
2
Michael N. Cook, MS, Simon M. Blake, PhD, Ian E. James, PhD, Sanjay Kumar, PhD,
Michael W. Lark, PhD: GlaxoSmithKline, King of Prussia, Pennsylvania;
3Ada
A. Cole, PhD: Rush University at Rush-Presbyterian-St. Luke's Medical Center,
Chicago, Illinois.
46
Abstract
Objective Traumatic joint injury leads to an increased risk of osteoarthritis (OA), but
the progression to OA is not well understood. We undertook this study to measure
aspects of proteoglycan (PG) degradation after in vitro injurious mechanical
compression, including up-regulation of enzymatic degradative expression, and
cytokine-stimulated degradation.
Methods Articular cartilage tissue explants were obtained from newborn bovine
femoropatellar groove and from adult normal human donor knee and ankle tissue.
Following injurious compression of the cartilage, MMP-3 and MMP-13 mRNA
expression levels were measured by Northern analysis, and proteoglycan loss to the
medium after cartilage injury was measured in the presence and absence of added
exogenous cytokine (IL-1a or TNFa).
Results During the first 24 hours after injury in bovine cartilage, MMP-3 mRNA levels
increased 10-fold over the levels in control cartilage (N = 3 experiments), whereas
MMP-13 mRNA levels were unchanged. PG loss was significantly increased after injury,
but only by 2% of the total PG content, and only for the first 3 days following injury.
However, compared with injury alone or cytokine treatment alone, treatment of injured
cartilage with either 1 ng/ml IL-1l or 100 ng/ml TNFu caused marked increases in PG
loss (35% and 54%, respectively, of the total cartilage PG content). This interaction
between cytokine treatment and injury was statistically significant. In human knee
cartilage, the interaction was also significant for both IL-1a and TNFu, although the
magnitude of increase in PG loss was lower than that in bovine cartilage. In contrast, in
human ankle cartilage, there was no significant interaction between injury and IL-1a.
The cytokines IL-lu and TNFu can cause a synergistic loss of
Conclusion
proteoglycan from mechanically injured bovine and human cartilage. By attempting to
incorporate interactions with other joint tissues that may be a source of cytokines, in
vitro models of mechanical cartilage injury may explain aspects of the interactions
between mechanical forces and degradative pathways which lead to osteoarthritis
progression.
Supported in part by grants AR-45779 and AR-2P50-39239 from the NIH, a grant from
GlaxoSmithKline, and a Whitaker Foundation fellowship
47
Introduction
Patients who have sustained a traumatic joint injury are known to have an increased
risk of osteoarthritis (OA) in that joint.1 3 6, 4 Unlike other risk factors for arthritis, such as
age and obesity, traumatic joint injury is a discrete event, and investigating its effects on
the joint could potentially help to understand the pathogenesis of osteoarthritis in
general. Important insights have come from cruciate ligament transection in animal
models of knee OA, adding to the understanding of how loss of joint stability following
ligament transection can cause abnormal loading and proprioception. Clinically,
however, despite the development of surgical interventions that can restore mechanical
stability and function to a patient's knee joint after ligament damage, these procedures
do not appear to greatly reduce the risk for development of OA.3 This suggests that in
addition to the effects of subsequent functional impairment, the initial traumatic event
may have irreversible effects on the joint tissues and resident cells.
In the past ten years, in vitro models of acute compressive trauma to the articular
cartilage have begun to document certain events that can occur immediately following
cartilage injury. In particular, several studies have characterized the resulting damage to
the
cartilage
matrix,
medium 18,20,28,65,66,67,6,
including
loss
of
proteoglycan
to
the
conditioned
increased swelling 18 ,28, and increased levels of denatured
collagen neoepitopes.13,6 8 In vitro models have also started to focus on identifying the
effects of injurious compression on the chondrocytes themselves. Under certain
conditions, injury can cause cell death by apoptosis, 2 8 ,52 and can abolish the normal upregulation of chondrocyte biosynthesis by dynamic compression.19
The mechanism for the progressive loss of proteoglycan from the cartilage matrix, one
of the earliest features of osteoarthritic cartilage, has long been an important focus of
research. In patients with an injury to the anterior cruciate ligament (ACL), Lohmander
et a15 have reported that synovial fluid levels of proteoglycan fragments and of
degradative enzymes such as pro-matrix metalloproteinase 3 (proMMP-3) are increased
within days and sustained at elevated levels over a period of years. However, the
molecular-level effects of mechanical injury to the cartilage in vitro have not yet been
48
delineated. It is also not yet clear whether the loss of proteoglycan from the cartilage in
models of mechanical injury involves up-regulation of enzymatic degradation, or is
caused by mechanical disruption of the tissue alone. For example, Quinn et a16 7 found
that their model of repeated mechanical injury resulted in increased loss of highmolecular-weight, apparently uncleaved aggrecan, and no evidence for an increase in
enzymatic cleavage.
Two aspects of the current understanding of OA have motivated us to modify our in vitro
model of mechanical injury to investigate additional factors that may be present in vivo.
First, it is held by consensus that OA is a disease of the whole joint,64 69 with possible
contributions from the bone, cartilage, synovium, muscles and nervous system. Second,
inflammatory processes such as cytokine-induced activity are increasingly suspected to
play a role in the pathogenesis of OA (for review see refs.
research using
in vitro injury models
69
and
has recently focused
70).
In this light, our
on understanding
chondrocyte mechanotransduction of injurious compression and on broadening these
models to account for interactions with other tissues, such as factors secreted by the
joint capsule. 71 To further characterize the chondrocytes' catabolic response to
mechanical injury, the present study has 3 aims: (1) to investigate the up-regulation of
matrix-degrading proteases after injury, focusing here on MMP-3 (stromelysin 1) and
MMP-13 (collagenase 3), (2) to examine proteoglycan loss after addition of either
exogenous interleukin-la (IL-1a) or tumor necrosis factor a (TNFu) to injured bovine
knee cartilage, and (3) to make an initial comparison of the response of human knee
and ankle cartilage from the same donor to the combination of injury and exogenous
IL-1.
Materials and Methods
Bovine articular cartilage tissue
Articular cartilage explant discs were obtained from femoropatellar grooves of 1 to 2week-old calves, obtained from a local abbatoir (Research '87, Hopkinton, MA) on the
day of slaughter, as previously described.2 3 Briefly, cartilage-bone cylinders (9 mm in
49
diameter) were drilled perpendicular to the cartilage surface and placed in a microtome
holder. After creating a level surface by removal of the most superficial -100
tm, the
next 2 mm of cartilage were sliced by a microtome, producing two 1-mm-thick slices.
Finally, four explant discs were punched out of each slice, resulting in cartilage discs
that were 1 mm thick and 3 mm in diameter. Cartilage from this middle-zone region in
newborn calf has previously been shown to have a reasonably homogeneous
population of cells and matrix.
Nevertheless, in all subsequent experiments, treatment
groups were matched for location and depth of the cartilage on the joint surface by
distributing one disc from a single slice to each of the different treatment groups.
Cartilage was then left in culture to equilibrate for 3 days in medium (low-glucose
Dulbecco's Modified Eagle's Medium [DMEM], supplemented with 10% fetal bovine
serum, 10 mM HEPES buffer, 0.1 mM nonessential amino acids, 0.4 mM proline, 20
pg/ml ascorbic acid, 100 units/ml penicillin G, 100 pg/ml streptomycin, and 0.25 pg/ml
amphotericin B) in a 37 OC, 5% CO2 environment.
Postmortem adult human donor tissue
Adult human donor knee and ankle joints were obtained from the Gift of Hope Organ
and Tissue Donor Network (Elmhurst, IL). All research was approved by the Office of
Research Affairs at Rush-Presbyterian-St. Luke's Medical Center and by the Committee
on the Use of Humans as Experimental Subjects at the Massachusetts Institute of
Technology. Donor tissue was excluded for diagnosis or treatment of joint disease,
based on reports from the families of the donors, and for cause of death involving blunt
trauma. In addition, all joint surfaces were scored as grade
2 on a modified Collins
scale,72 and only cartilage from an area that was smoothly reflective and unfibrillated to
visual inspection was harvested for these experiments. Human donor cartilage tissue
was harvested and cultured as described above for bovine cartilage with two
exceptions. First, knee cartilage explants were taken from the femoral condyles, the
tibial plateau, and the femoropatellar groove; in addition, ankle cartilage explants were
taken from the talar dome cartilage of the same donor limb. Second, since the layer of
adult human articular cartilage is much thinner than that of bovine calf, the explant discs
were sliced to a thickness of 0.5 mm instead of 1 mm. Again, explants were matched for
50
location and depth of the cartilage by distributing one disc from a single slice to each of
the different treatment groups.
Injurious compression
Injurious compression was performed in a custom-designed incubator-housed loading
apparatus,
30
(Figure 1A). Cartilage discs were placed, one at a time, in a well in the
center of a polysulfone chamber (Figure 1A [part 1] and 1 B) that allowed for unconfined
compression of the disc. The lid of the chamber is attached to a load cell (Figure 1A,
part 3), and the displacement of the lid (equal to the displacement of the cartilage
surface) was measured using a linear-voltage differential transducer (Fig 1A, part 2).
Since cartilage swelled during the 3 days in culture prior to injury, the actual thickness
was measured and recorded so that the zero-strain position coincided precisely where
contact was made between the chamber lid and the top surface of the cartilage. The
system can be operated in closed-loop feedback control of the displacement or load. In
these studies, displacements were applied and the resulting loads measured. Injurious
compression was applied by a ramp compression to a set value of final strain, followed
by release. For bovine cartilage, compression was applied to 50% final strain at a
velocity of 1 mm/second (a strain rate of 1 s1). For adult human cartilage, with thinner
discs and with biomechanical properties different than those of calf cartilage, preliminary
experiments showed that higher strains and faster strain rates were needed to produce
levels of peak stress and visible damage qualitatively similar to bovine cartilage injury.
Thus, compression was increased to 65% final strain at 2 mm/second (a strain rate of 4
s1). A schematic of the resulting stress during compression and stress-relaxation after
release of compression is shown in Figure 1C. After injury, the macroscopic appearance
of the cartilage explant was noted by visual inspection as being either unchanged,
deformed such that the explant appeared elliptical in shape after injury, or grossly
disrupted by tissue fracture.
Biochemical analysis
The sulfated glycosaminoglycan (GAG) content of the conditioned medium was
quantified using the dimethylmethylene blue dye-binding assay.73 For each analysis, a
51
standard curve was generated with known concentrations of shark cartilage chondroitin
sulfate C (Sigma, St. Louis, MO). In separate experiments, the total sulfated GAG
content of a cartilage disc was assayed after proteolytic digestion (overnight in
proteinase K at 600C). For ease of comparison, GAG loss per explant disc was
expressed as a percentage of the total GAG content of a disc, using as typical values of
total GAG content 275 ptg/disc (39 mg/ml) for 1 mm-thick newborn bovine discs, and
125 jig/disc (35 mg/ml) for 0.5 mm-thick adult human discs.
Expression analysis
For messenger RNA (mRNA) expression analysis, bovine cartilage explants were
switched at least 12 hours before injury into a serum-free, defined medium, with 1%
ITS
(containing bovine insulin, human transferrin, and sodium selenite; Sigma) plus all
supplements as described above. In each experiment, injurious compression was
applied to 15 cartilage discs, which were returned to culture for an additional 1, 3, 6, 12,
or 24 hours (split evenly at 3 discs per timepoint), after which the discs were flashfrozen. Another 15 cartilage discs were not loaded but frozen at the same times. The
cartilage from all timepoints was then pooled together prior to analysis in order to
ensure a sufficient amount of RNA for analysis. Messenger RNA was extracted and
Northern blot analyses were performed. Specific probes were generated to bovine
MMP-3 (the probe sequence is as reported previously; 74 GenBank accession no.
AF069642) and to bovine MMP-13 (bases 865-1515 from GenBank accession no.
AF072685). Hybridization was normalized by expression of elongation factor lcc.
Cycloheximide treatment and injury
In a factorial design, bovine cartilage explants were incubated in the presence or
absence of 100 pg/ml of the protein synthesis inhibitor cycloheximide and either
subjected to injurious compression or left uncompressed. Six hours prior to injury, the
medium for all groups was replaced with a serum-free, defined medium containing 1%
ITS plus all other supplements as described above. At this time, the cycloheximide was
added to the appropriate groups in order to preincubate the explants in cycloheximide
before injury. Medium was replaced again immediately following injury for all groups.
52
Twenty-four hours after injury, conditioned medium was collected and analyzed for
sulfated GAG content.
Exogenous cytokine treatment and injury
A factorial design was also used to investigate the effect of exogenous cytokines on
injured tissue. Cartilage explants were incubated in the presence or absence of a
cytokine and
either subjected to injurious compression or left uncompressed.
Recombinant human IL-1l was used at 1-10 ng/ml and recombinant human TNFU was
used at 100 ng/ml (both from R&D Systems, Minneapolis, MN). Both interventionsinjury, and medium change with or without cytokine-were performed at the same time.
For human tissue, although explants were initially equilibrated in medium with 10%
serum as described above, preliminary experiments suggested the need to remove any
anabolic factors in serum that could mask degradation after injury. Therefore, human
explants were placed into a serum-free medium (DMEM, HEPES buffer, and antibiotics
only) 2 hours prior to intervention. After injury, cartilage was again replaced in serumfree medium (with or without exogenous cytokine). For bovine tissue, no change was
made in medium formulation, and explants were simply placed in fresh medium at the
time of the intervention. After incubation for a further 3 days following the intervention,
the conditioned medium was collected for GAG analysis. In one experiment, the
medium was replaced at this point, and the medium was also collected 7 days following
the intervention.
Statistics
All data are shown as mean ± SEM. Comparisons of normalized expression data were
tested with a Student's t-test. For comparisons of GAG loss data, paired differences
were tested with the non-parametric sign test (exact method). Several experiments used
a 2 x 2 factorial design, the two factors being injury and cytokine or cycloheximide
treatment. In these experiments, we wished to test whether there was an interaction
between the two factors (i.e., whether there is a more-than-additive effect of incubating
injured tissue with cytokines). The data were therefore analyzed to test the statistical
53
significance of an interaction between the two factors in a two-way analysis-of-variance
(ANOVA) (S-Plus, MathSoft Inc., Cambridge, MA).
Results
Compression of bovine calf femoropatellar groove cartilage to 50% final strain at 1
mm/second over two typical experiments produced mean ± SEM peak stresses of 23.1
± 0.6 MPa (N = 34 cartilage discs). Macroscopic tissue changes were limited to
irreversible deformation to an elliptical shape, and those occured in 13 of the explants
(38%), consistent with the results of previous studies of injurious compression 1 9,2 8. In
adult human cartilage, compression to 65% final strain at 2 mm/second of same-donorlimb knee and ankle tissue produced peak stresses of 11 ± 1 MPa (N = 16 cartilage
discs) in knee cartilage (pooled from all femoral and tibial cartilage sites), and 16 ± 1
MPa (N = 12 cartilage discs) in ankle talocrural cartilage. This compression resulted in
macroscopic tissue changes (elliptical appearance) to 13 of 16 knee cartilage explants
(81%) and to 2 of 12 ankle cartilage explants (17%) (P < 0.001 for difference by exact
calculation from the binomial distribution).
Conditioned medium from injured and uninjured cartilage discs was analyzed for
sulfated GAG content at 3 time points after injury. The resulting GAG loss (N = 8
cartilage discs from 3 different animals) is shown in Figure 2. The amount of GAG loss
to the medium was significantly higher after injury for the first day (P < 0.001) and for
the time between days 1 and 3 (P < 0.05), but not for the time between days 3 and 7 (P
= 0.58).
MMP-3 (stromelysin 1) and MMP-13 (collagenase 3) message levels after injury of
bovine cartilage explant discs were quantified by Northern analysis (Figure 3). MMP-3
mRNA levels in injured cartilage were increased 10 ± 2 times over the levels in control
cartilage during the first 24 hours (P < 0.05 by t-test; n = 3 experiments). In contrast,
little change was observed in the levels of MMP-13 message (0.9 ± 0.1 times control
levels, P = 0.33).
54
To inhibit protein translation by the chondrocytes, we incubated bovine calf cartilage
with 100 tg/ml cycloheximide prior to and after injury (Figure 4). We verified that this
reduced radiolabel incorporation by 90% for 3H-proline and by >99% for
35
S-sulfate
(data not shown). As expected, an increase was observed in GAG content of the
conditioned medium 24 hours after injury (P < 0.001). However, cycloheximide
treatment did not reduce GAG loss from injured cartilage any more than it reduced GAG
loss from unloaded cartilage (no statistically significant interaction between injury and
cycloheximide treatment) (P = 0.92; n = 12 cartilage discs per treatment group).
Incubation of uninjured bovine calf cartilage with 0 ng/ml, 1 ng/ml, and 10 ng/ml IL-1a
for 3 days produced, as expected, an increasing loss of sulfated GAG to the conditioned
medium in a dose-dependent manner (Figure 5A), from a mean of 5% of the total GAG
with no IL-1, to a mean of 16% with 10 ng/ml of IL-1. In contrast, incubation of IL-1 with
injured cartilage for 3 days produced a dramatic, dose-dependent release of a large
portion of the tissue's sulfated GAG, from a mean of 7% of the total GAG with injured
cartilage but no IL-1, to a mean of 60% with 10 ng/ml IL-1. Analysis of the GAG loss
data showed a statistically significant interaction between injury and IL-1 treatment (P <
0.001 for interaction by two-way ANOVA; n = 5 cartilage discs per treatment group).
Similarly, an experiment was performed to test for an interaction between injury and
treatment with 100 ng/ml TNFa in the resulting loss of GAG (Figure 5B). Observed GAG
loss from injured cartilage after 3 days was slightly more than from uninjured cartilage
(7.2% vs. 6.6% of the total cartilage GAG content, P not significant [NS]), and
incubation of uninjured cartilage with TNFa was observed to increase GAG loss to 14%
of the total (P = 0.03). The combination (incubation of injured cartilage with TNFu)
resulted in a mean loss of 54% of the cartilage GAG content to the medium after 3 days.
Analysis of the data showed a statistically significant interaction between injury and
TNFa treatment (P < 0.001; n = 6 cartilage discs per treatment group).
In adult human donor knee cartilage (Figure 6A), as with the immature bovine cartilage,
there was a statistically significant interaction between injury and 10 ng/ml IL-1a
55
treatment in terms of GAG loss to the medium after 3 days (P < 0.05; n = 8 cartilage
discs per treatment group). The cartilage was from a deceased 72-year-old male donor
with Collins grade 2 knee and ankle joint surfaces. All 3 joint surfaces (tibial plateau,
femoral condyles, and femoropatellar groove) were harvested for the experiment and
pooled in the analysis. Uninjured control discs released 4.6 ± 0.7% of their total GAG
content to the medium after 3 days. Observed GAG loss was slightly higher from injured
cartilage (6.9 ± 0.8%, P < 0.01), and incubation of uninjured cartilage with IL-lx had
little effect on mean ± SEM GAG loss (5.2 ± 0.6%; P NS). In contrast, incubation of
injured cartilage with IL-la resulted in a mean ± SEM GAG loss of 10.4 ± 1.5% of the
total cartilage GAG content.
Unlike the results from the human knee cartilage, in ankle (talocrural) cartilage from the
same donor (Figure 6B), there was no statistically significant interaction between injury
and 10 ng/ml IL-1a treatment (P = 0.50; n = 6 cartilage discs per treatment group). After
3 days, uninjured control discs of ankle cartilage released 3.4 ± 0.3% of their total GAG
content to the medium. There was no observed effect of injury (mean ± SEM GAG loss
3.4 ± 0.5%), and there were small increases in mean ± SEM GAG loss after IL-1
treatment of uninjured (4.1 ± 0.4%) and injured (4.7 ± 0.5%) cartilage.
Finally, we investigated the effects of cartilage injury and treatment with 100 ng/ml
TNFu in human knee cartilage over a 7-day period. The knee cartilage was from a
deceased 66-year-old male donor with Collins grade 1 femoral and tibial joint surfaces.
Three days after the experimental intervention (Figure 7A), analysis of GAG loss to the
medium showed a statistically significant interaction between cartilage injury and TNFa
treatment (P < 0.01; n = 8 cartilage discs per treatment group). Compared to uninjured
control discs (9.3 ± 0.2%), the mean ± SEM GAG losses after injury and TNFu
treatment showed little change (9.5 ± 0.5% and 9.1 ± 0.4%, respectively; both P NS),
while GAG loss from injured cartilage treated with TNFu was increased to 12.5 ± 0.8%
of the total GAG. The GAG loss to the medium on day 7 after intervention (i.e., the
medium conditioned from day 3 to day 7) showed a similar trend (Figure 7B), but the
interaction did not reach statistical significance (P = 0.07).
56
Discussion
We have used an in vitro model of mechanical injury to articular cartilage in order to
investigate the effects of injury on subsequent proteoglycan release and to demonstrate
that a more profound release was induced by the interaction of the cytokines IL-1
and
TNFu with injured tissue. Using our model of injurious compression, we first observed
that injury to newborn bovine cartilage produced a significant up-regulation of mRNA
levels for MMP-3 but not MMP-13. The cartilage injury alone, however, did not result in
an increase in cell-mediated enzymatic degradation of proteoglycan, as evidenced by
the lack of effect of cycloheximide on GAG loss. A series of experiments using bovine
calf and normal adult human knee cartilage then showed that addition of IL-1l
or TNFu
to injured cartilage caused a synergistic increase in the loss of GAG from the tissue.
The finding of an -10-fold increase in MMP-3 expression within 24 hours of an in vitro
injury (Figure 3) echoes the finding of increased MMP-3 proenzyme in synovial fluid
from patients within days after traumatic joint injury.5 These results raise the possibility
of increased MMP-3 activity as a link between injurious mechanical forces and cartilage
degradation. However, MMP-3 expression is known to be up-regulated without any
increase in MMP-mediated aggrecan degradation in certain situations, such as
stimulation of bovine nasal cartilage with IL-1.
In this respect it is more interesting
that the pattern of MMP up-regulation by mechanical injury may be different from that of
cytokine stimulation. Whereas we observed an increase in expression of MMP-3 but not
MMP-13 after injury, stimulation with IL-1 or TNFu has been reported to increase
expression of both enzymes in bovine and human chondrocytes.74 '76'7 The pattern of
MMP expression seen here thus suggests that mechanical forces act to stimulate MMP
expression through a different pathway than do cytokines, although direct comparison of
the effects of cytokine and mechanical stimulation in tissue would be required to support
this hypothesis.
One pathway for up-regulation of MMP-3 by mechanical disruption of tissue was
recently identified by Vincent et al., who showed that cutting porcine explants results in
57
the release of basic fibroblast growth factor from the matrix, which causes extracellular
signal-related kinase activation and up-regulation of MMP-3.
Preliminary results in our
laboratory have also shown the same pattern of MMP-3 and MMP-13 expression in
injured human knee cartilage, although with MMP-3 not as highly increased as in the
immature bovine cartilage (data not shown). Ongoing studies are expanding on these
results with quantification of other relevant protease expression levels in this model
system.
Despite the up-regulation of MMP-3 seen here, evidence from our studies and those of
others suggests that in vitro models of injury to cartilage alone do not demonstrate the
steady progression of proteoglycan release that one might expect from up-regulated
enzymatic matrix degradation. The increase in GAG loss caused by injury alone in our
model, though statistically significant, was not a large portion of the total GAG content of
the cartilage explant (1-2% in Figures 2, 4, and 5). In addition, the GAG loss to the
medium was increased only during the first 3 days after injury. While there are varying
results reported among in vitro cartilage injury models that use different animal species
and mechanical loading protocols, this is consistent with observations from a number of
investigators that:
1) GAG loss is increased predominantly during the first few days
after injury1 9,2 0 ,28 ,65
66,68
tissue GAG content1 9,28
, and 2) the increase in GAG loss is a small fraction of the total
52
,65
68
. Finally, analysis of aggrecan fragments by Quinn et
al. 67 suggests that it is predominantly release of uncleaved aggrecan that is responsible
for the increase in GAG loss after injury.
These observations have led us to hypothesize that the loss of matrix proteoglycans
seen in these models of mechanical injury to cartilage may be primarily due to
mechanical disruption or dislodgment of matrix rather than due to cell-mediated
enzymatic degradation. This hypothesis is supported by the results of the cycloheximide
study reported here (Figure 4), which demonstrate that in our model, incubation of
cartilage with the protein synthesis inhibitor cycloheximide before and after injury had no
significant effect on sulfated GAG release to the medium after injury. This does not
preclude the possibility that the GAG release is mediated by increased activity of
58
enzymes translated before the addition of cycloheximide. For example, injury could
induce the activation of degradative proenzymes already present in the tissue matrix.
This possibility could be explored further with specific enzyme inhibitors and with
investigation of released aggrecan fragment neoepitopes.
Since injury to cartilage alone does not appear to involve sustained matrix degradation,
the evidence that the pathogenesis of osteoarthritis involves an interaction among
different tissues and cell types within a joint led us to investigate possible interactions
between injured cartilage and exogenous cytokines. We report here that incubation of
injured cartilage with either IL-1 or TNFu can cause a synergistic increase in sulfated
GAG loss from both newborn bovine and normal adult human knee cartilage. These
results are significant in that they provide the first evidence identifying an interaction
between mechanical forces, factors in the joint external to the cartilage, and degradation
in a cartilage injury model. In interpreting these results, it should be noted that the levels
of cytokine used in our short-term studies were probably much higher than the levels
seen in synovial fluid in vivo. Reported concentrations of IL-1p range from 1 pg/ml in
patients with chondromalacia of the patella,79 to 20 pg/ml in patients with rheumatoid
arthritis
;80
TNFa can range from 40 pg/ml to 250 pg/ml in rheumatoid arthritis. 80
Therefore, the absolute amounts of matrix loss are not applicable to in vivo joint injuries.
Nevertheless, the interaction identified here does suggest one aspect of how low levels
of cytokine in the synovial fluid might lead to a small but sustained increase in matrix
degradation of injured cartilage tissue in vivo.
The precise mechanism responsible for the dramatic enhancement of GAG loss from
injured cartilage in the presence of IL-1a or TNFu remains to be delineated. It is
possible that purely mechanical damage to the matrix is responsible for this synergistic
effect. Since injurious compression of cartilage can produce macroscopic tissue
damage, it is also likely to increase cartilage permeability. Therefore, increased GAG
loss could result from enhanced cytokine transport into the tissue. However, it seems
unlikely that enhanced transport alone could fully explain these results. For example, we
note that in uninjured bovine tissue, the effect on GAG loss of increasing IL-1
59
concentration by a factor of 10, to 10 ng/ml, was still far less than the effect of
incubating injured tissue with 1 ng/ml IL-1 (Figure 5A). An alternative hypothesis is that
molecular-level disruption of matrix caused by mechanical injury would allow for greater
access of degradative enzymes to aggrecan cleavage sites, a mechanism that is
consistent with certain previous hypotheses of OA pathogenesis. 81 Finally, a cellmediated mechanism, such as synergistic action of degradative pathways activated by
injurious compression with pathways activated by cytokines, could be responsible for
this effect. Further studies, such as examination of the structure of the released
aggrecan fragments, are needed to test such possibilities.
The synergistic effects of cytokines and cartilage injury observed with immature bovine
tissue were also found to occur in adult human knee tissue. However, this interaction
was not statistically significant in identical experiments using ankle tissue from the same
donor. Even application of the 65% compression used in these experiments did not
increase GAG loss from the ankle cartilage (Figure 6). The higher peak stress and lower
extent of damage produced by injurious compression of ankle tissue, here, is consistent
with the earlier finding of increased GAG content and higher compressive stiffness of
human ankle tissue compared to knee tissue.82 These observed differences in the
response of injured knee and ankle cartilage explants to exogenous cytokines are
particularly relevant in light of our group's ongoing effort to identify biochemical and
biomechanical differences that may contribute to the lower incidence of OA in the ankle
compared with the knee.7 2 ,83 The results of the present study lead to the hypothesis that
the interaction between joint injury and cytokines may be an underlying difference in the
progression to OA of knee and ankle cartilage after injury.
Another important effect of injurious compression in our model is cell death by both
necrosis and apoptosis. Prior investigations with identical loading conditions have
estimated cell death after injury to be 15%-25% in newborn bovine cartilage. 19' 58 The
presence of significant cell death after injury suggests that on a per-cell basis, the
degradative effect of IL-1 on injured cartilage could be even more profound. Cell death
after injury in human tissue was not measured in these experiments, but in bovine
60
tissue, apoptotic cell death after injury decreased sharply with increasing age in bovine
cartilage, to 6% cell death in 2-year-old steers.
This suggests that the percentage of
cell death after injury would be relatively low in human tissue, a hypothesis that would
be important to test in future studies.
The results presented here extend the current literature by suggesting the importance
that factors external to the cartilage, such as cytokines in the synovial fluid, could play in
the development of cartilage degradation after acute joint trauma. Models of cartilage
injury may thus benefit from incorporating interactions with other joint tissues. In
particular, aspects of cell response to injurious compression such as the increase in
MMP-3 expression reported here may take on added importance once these
interactions with exogenous factors are accounted for.
Acknowledgments
The authors would like to express their appreciation to the Gift of Hope Organ and
Tissue Donor Network and to the donors' families for access to the human donor tissue.
61
U
B
I
70-
- Strain
60-
t
Stre ss
50-
-25
20Z
CL
40
30-
'10
W 20
5 C)
10'
0
0
A
C
0.1
0.2
0.3
-4 0
0.4
Time (s)
Figure 1. Schematic of loading device and applied compression waveforms.
A, An incubator-housed loading apparatus was used to apply injurious
compression to a single cartilage disc held inside a polysulfone chamber (1).
Transducers recorded both the displacement (2) and the load (3). B, A crosssectional view of the chamber shows the cartilage disc, at the bottom of a well in
the center of the chamber, loaded by a platen attached to the lid of the chamber.
C, A schematic of data recorded during a ramp and release compression of the
cartilage to 65% of its original thickness at a velocity of 2 mm/second shows the
generation of a peak stress of 20.1 MPa followed by stress relaxation.
62
6-
50I-
*
0 Injured
**
T
-4-
0 4-P
01
30
.,
El Unloaded
-i-
2-
0
1 0-
Day 0-1
Days 1-3
Days 3-7
Figure 2. Sulfated glycosaminoglycan (GAG) loss to the medium after injury of
newborn bovine cartilage. GAG loss is expressed as a fraction of the total GAG
content of a single disc. Statistical comparisons of GAG loss were made between
unloaded and injured discs. Values are the mean ± SEM (n = 8 cartilage discs
from 3 different animals). * = P < 0.05 and ** = P < 0.01 versus unloaded discs.
63
*
e 0
" 10 -
-Unloaded
E
Control
0
T
0
MMP-13
MMP-3
Figure 3. Matrix metalloprotease (MMP) mRNA expression within 24 hours after
injury. After injurious compression, cartilage explants were returned to culture,
then flash-frozen at time points 1, 3, 6, 12, and 24 hours after the injury (15
cartilage discs, 3 per time point). Fifteen control discs were unloaded but frozen
at the same times. Northern analysis showed a statistically significant increase of
MMP-3 mRNA levels in injured cartilage. Values are the mean ± SEM (n = 3
experiments). * = P < 0.05 versus control cartilage.
64
65-
El No CH
* Cycloheximide
0
4-
4-
0
0
0
-H I
M-
.J
0
2-
0
1 0-
--- I
Unloaded
Injured
Figure 4. Effect of cycloheximide (CH) on glycosaminoglycan (GAG) loss after
injury. Protein translation was blocked by incubation of cartilage discs with 100
pig/ml cycloheximide. GAG content of the conditioned medium was measured 24
hours after either injury or addition of cycloheximide. Cycloheximide treatment
did not reduce GAG loss from injured cartilage any more than it reduced GAG
loss from unloaded cartilage. Values are the mean ± SEM (n = 12 cartilage discs
per treatment group). P = 0.97 for interaction between injury and cycloheximide
treatment, by two-way analysis-of-variance.
65
60~
40
0
50-
Cl)
14-
40-
-J0
30-
Cl,
20-
0
El Unloaded
N Injured
I
1000
A
Z0
11
-,-
10
IL-1 Concentration (ng/ml)
60-
0 lo50-.
C Unloaded
I
N Injured
40
30 0
-i
2010-
0-
B
no TNF
100 ng/ml TNF
Figure 5. Glycosaminoglycan (GAG) loss 3 days after cytokine treatment of
injured bovine cartilage. Unloaded cartilage and injured cartilage was cultured
with either interleukin-1a (IL-1; 0 ng/ml, 1 ng/ml, or 10 ng/ml) (A), or tumor
necrosis factor a (TNF; 0 ng/ml or 100 ng/ml) (B). GAG content of the medium
was measured after 3 days of culture. Values are the mean ± SEM. P < 0.001 for
interactions between injury and treatment with either IL-1 (n = 5 cartilage discs
per treatment group) or TNF (n = 6 cartilage discs per treatment group), by twoway analysis-of-variance.
66
12-
E Unloaded
U Injured
010
0
U)
U)
I
T
06
0
4-
E
0
0
A
10 ng/ml IL-1
No IL-1
12
M 10-
-
L Unloaded
MInjured
0
0U)
0
0)
06-
E
0 2'4-
B
0
10 ng/ml IL-1
No IL-1
Figure 6. GAG loss 3 days after IL-1 treatment of human knee and ankle
cartilage. Unloaded or injured adult human donor cartilage was incubated with 0
ng/ml or 10 ng/ml IL-1l, and GAG content of the medium was measured after 3
days of culture. Values are the mean ± SEM. A, Incubation of injured knee tissue
with IL-1 caused a synergistic increase in loss of GAG (P < 0.05 for interaction
between injury and IL-1 treatment by two-way analysis of variance [ANOVA]; n =
8 cartilage discs per treatment group). B, In ankle tissue from the talar dome, the
interaction between injury and IL-1 treatment was not significant (P = 0.50 by
two-way ANOVA; n = 6 cartilage discs per treatment group). See Figure 5 for
definitions.
67
14'0
cY,
12-
0
4-
o
t.0
T
10-
0
0- 40
0
CD
* Unloaded
U Injured
-i-
8-
CD 64-
CDE
0
S.
to-
2-
0
A
I_
'I_
'
no TNF
14-
Ot12-
_I
100 ng/ml TNF
l Unloaded
0 Injured
0 10
0 0
S8-
-J
E4 -
4-
0
C 2-
B
100 ng/ml TNF
no TNF
Figure 7. Time course of glycosaminoglycan (GAG) loss after TNFu treatment of
adult human knee cartilage. Unloaded and injured cartilage was incubated with 0
ng/ml or 100 ng/ml TNFu. After 3 days, the medium was replaced with fresh
solutions of TNF, and incubated for 4 more days. Values are the mean ± SEM.
A, In the medium collected at day 3, there was a synergistic increase in GAG
loss from injured tissue treated with TNF (P < 0.01 for interaction between injury
and TNF treatment by two-way ANOVA; n = 8 cartilage discs per treatment
group). B, In the medium collected at day 7, the interaction between injury and
TNF treatment did not reach statistical significance (P = 0.07 by two-way
ANOVA; n = 8 cartilage discs per treatment group).
68
Chapter 4
Inhibition of bovine articular cartilage biosynthesis by coincubation
with joint capsule tissue: Evidence for an IL-1 -independent pathway
This chapter is prepared for journal submission, to be co-authored by:
Parth Patwaril, Stephanie N. Lin , Ada A. Cole 3 , Sanjay Kumar 2, Alan J. Grodzinskyl
'Continuum Electromechanics Group, Center for Biomedical Engineering,
Massachusetts Institute of Technology, Cambridge, MA;
2
Musculoskeletal Diseases Department, GlaxoSmithKline, King of Prussia, PA;
3Rush
University at Rush-Presbyterian-St. Luke's Medical Center, Chicago, IL.
This research
was supported in
part by NIH grant AR45779, a grant from
GlaxoSmithKline, NIH SCOR grant AR-2P50-39239, and a fellowship from the Whitaker
Foundation.
69
Abstract
The reason for the increased risk for development of osteoarthritis (OA) after acute joint
trauma is not well understood. In particular, substances secreted from the joint capsule
such as interleukin-1 (IL-1) and tumor necrosis factor-alpha (TNF) may be important in
causing cartilage degradation. We coincubated cartilage tissue explants with joint
capsule tissue from newborn calf joints and from normal adult human post-mortem knee
joints. In selected experiments, cartilage was also subjected to injurious compression
(50% strain at a velocity of 1 mm/s). Under certain conditions, the combination of injury
and coincubation resulted in a synergistic increase in proteoglycan release from the
cartilage. However, the most prominent effect of coincubation was a -70% reduction in
cartilage proteoglycan biosynthesis in bovine calf tissue and a -40% reduction in adult
human tissue. In bovine tissue, we showed that conditioned medium from joint capsule
was equally effective at inhibiting cartilage biosynthesis. The inhibitory effect of
conditioned medium was heat-labile, and cartilage biosynthesis was recoverable.
However, blockade of IL-1 with soluble IL-1 receptor or IL-1 receptor antagonist
(IL-1 RA) had no effect on the inhibition of cartilage biosynthesis. Combined blockade of
IL-1 and TNF with IL-1RA and etanercept also had no effect on inhibition of
biosynthesis. These results suggest that excision and culture of joint capsule tissue
from a normal, uninflamed bovine calf joint causes the release of a nontoxic protein that
inhibits cartilage biosynthetic activity by an IL-1 and TNF-independent pathway.
Keywords: cartilage, synovium, injury, interleukin-1
70
Introduction
Osteoarthritis (OA) is the most common joint disease and leads to a substantial burden
of disability in the elderly population. OA is considered to be a disease of the entire
synovial joint, characterized pathologically by destruction of the articular cartilage and
changes to the underlying bone, synovium, ligaments, and muscle.64 The cause of OA
is unclear, but it is broadly understood that the degradation of the cartilage is due to an
interaction between the mechanical loading of the joint and its biological response that
leads to an imbalance between anabolic and catabolic activity of the chondrocytes, the
cells responsible for maintenance of the cartilage matrix.81 84 8 5
A traumatic joint injury, such as a sports injury leading to a ligament rupture or a
cartilage fracture, leads to a substantial increase in risk for subsequent development of
OA 1 2,8 6 . To investigate the events after joint injury which may eventually lead to the
development of OA, in vitro models that apply injurious mechanical loading to the
cartilage have been developed. In particular, several studies have characterized the
resulting damage to the major structural constituents of the cartilage matrix, including
loss of proteoglycan to the conditioned medium 18,20,65,67 and increased levels of
denatured collagen neoepitopes.13,6 8 In vitro models have also identified several effects
of injurious compression on the chondrocytes, such as apoptotic cell death,2 85 2 and loss
of the normal ability of dynamic compression to stimulate chondrocyte biosynthesis. 19
While these models have focused on the cartilage (or cartilage attached to subchondral
bone), it may be important to investigate possible interactions with other joint tissues,
since in the long term it is clear that the whole joint is a participant in the progression of
OA. For example, multiple joint tissues often are acutely damaged by a joint injury, 7
and cytokine levels have recently been shown to be elevated in the synovial fluid of
patients after an anterior cruciate ligament injury.8 8 Thus, even those injuries which
cause only damage to the articular cartilage may be affected in the short term by
inflammatory or degradative mediators associated with other joint tissues such as the
synovium.
71
The role of the synovium and inflammation in the pathogenesis of OA remains
controversial. Synovial inflammation is present clinically and histologically in some OA
patients, 89-91 but this may be a secondary response. Animal models have suggested
that OA can develop with only moderate levels of ultrastructural and hypertrophic
changes in synoviocytes and without infiltration.92 94 However, at the molecular level,
evidence increasingly points to a pivotal role for inflammatory
mediators and
inflammation-related signaling pathways in the cartilage degradation observed in
OA.69,70,85,95-97
With these considerations in mind, we have begun to explore in vitro several aspects of
how joint capsule tissue and inflammatory mediators could interact with the cartilage
after a traumatic joint injury. We have previously reported that injured cartilage which is
then cultured in the presence of the cytokines IL-1 and TNFac can result in a synergistic
loss of tissue proteoglycans.
In addition, we have investigated a model in which
cartilage is coincubated with excised joint capsule tissue from a normal joint. In this
model, the excision and culture of capsule tissue could be considered as a traumatic
event for the joint capsule and may thus allow exploration of some interactions between
joint capsule and cartilage tissues that occur in a traumatic injury to multiple joint tissues
in vivo. Our initial observation from this model was that coincubation of joint capsule
tissue resulted in a profound inhibition of cartilage biosynthetic activity even in uninjured
cartilage. 71 The specific objectives of this study were therefore to further characterize
the effects of coincubating joint capsule tissue with cartilage on cartilage biosynthetic
activity, and to study the role of IL-1 and TNFL in this model using cytokine blockade.
Experimental Procedures
Materials
Etanercept was from Immunex Corp. (Seattle, WA); the recombinant human cytokines
interleukin-1a (IL-1), soluble IL-1 type I
receptor (sIL-1r), interleukin-1
receptor
antagonist (IL-1RA), and tumor necrosis factor a (rh TNF) were from R&D Systems
(Minneapolis, MN). Recombinant bovine TNFx (rb TNF) was a generous gift of Dr.
72
Theodore H. Elsasser (United States Department of Agriculture, Beltsville, MD).
Radiolabeled
35
S-sulfate and
3H-proline
was from New England
Nuclear (now
PerkinElmer Life Sciences, Boston, MA). Additional supplies were from Sigma Chemical
Co. (St. Louis, MO) where not otherwise noted.
Tissue Harvest
Articular cartilage explant disks were obtained from femoropatellar grooves of 1 to 2week-old calves, obtained from a local abbatoir (Research '87, Hopkinton, MA) on the
day of slaughter, as previously described 2 3. In brief, cartilage-bone cylinders (9 mm in
diameter) were drilled perpendicular to the cartilage surface and placed in a microtome
holder. After creating a level surface by removal of the most superficial -100 pm, the
next 2 mm of cartilage were sliced by a microtome, producing two 1-mm-thick slices.
Finally, four explant disks were punched out of each slice, resulting in cartilage disks
that were 1 mm thick and 3 mm in diameter. Treatment groups were matched for
location and depth of the cartilage on the joint surface by distributing one disk from a
single slice to each of the different treatment groups. Cartilage was then left in culture
to equilibrate for 3 days in low-glucose Dulbecco's Modified Eagle's Medium (DMEM)
supplemented with 10% fetal bovine serum, 10 mM HEPES buffer, 0.1 mM nonessential
amino acids, 0.4 mM proline, 20 pg/ml ascorbic acid, 100 U/ml penicillin G, 100 pg/mI
streptomycin, and 0.25 pg/ml amphotericin B (standard medium), in a 37'C, 5% CO 2
environment. In addition to harvest of cartilage explant disks, portions of joint capsule
were cut medial and lateral to the patella and suprapatellar fat pad during dissection of
the joint. No attempt was made to isolate the synovium. The joint capsule tissue, which
varied from 0.5- to 3-mm-thick, was punched and cultured in standard medium. To block
for the varying thickness and location of the capsule, adjacent capsule punches were
distributed evenly among treatment groups.
Human Donor Tissue
Normal adult human donor knee joints were obtained from the Gift of Hope Organ and
Tissue Donor Network (Elmhurst, IL). All research was approved by the Committee on
the Use of Humans as Experimental Subjects at the Massachusetts Institute of
73
Technology. Post-mortem knee joints were washed, placed in DMEM, and harvested
within 48 hours as described previously.5 3 0.5-mm-thick, 3-mm-diameter cartilage disks
were obtained from femoropatellar, femoral condylar, and tibial plateau articular
cartilages and placed into standard medium (see above). In addition, portions of joint
capsule tissue adjacent to the attachment site of the capsule to the tibia were dissected
and punched to form 3-mm-diameter disks.
Injurious compression
In selected experiments, compression was performed in a custom-designed incubatorhoused loading apparatus.3 0 Cartilage disks were placed, one at a time, in a well in the
center of a polysulfone chamber that allowed for unconfined compression of the disk.
Injurious compression was applied by a ramp compression to 50% final strain at a
velocity of 1 mm/s (a strain rate of 100%/s), and compression was released within five
seconds (the same protocol described in Patwari et al. 5 3 ).
Radiolabel Incorporation
Three days after the intervention (injury or coincubation), cartilage disks were
radiolabeled by incubation with fresh medium containing 10 pCi/ml each of
35
S-sulfate
and 3H-proline (all capsule tissue was removed before the label). The disks then were
washed of unincorporated label, digested in protease K, and analyzed by scintillation
counter for incorporated radiolabel.
Proteoglycan Concentration
The sulfated glycosaminoglycan (GAG) content of the conditioned medium was
quantified using the dimethylmethylene blue dye-binding assay, 98 with standard curves
generated from known concentrations of shark cartilage chondroitin sulfate C.
Statistical Analysis
Most experiments used a 2x2 factorial design, where one treatment was coincubation,
and the purpose of the experiment was to test whether a second treatment (such as IL1 blockade) modified the effect of coincubation. Therefore, in each experiment our
74
primary hypothesis of interest was whether there was a significant interaction between
the effects of the two treatments. This was tested in a two-way analysis-of-variance
(ANOVA) model. Where inferences over a population of experiments (or animals) were
made by analysis of more than 3 repeated experiments, a linear mixed-effects mode
99
was used to include experiment-to-experiment variation as a random effect, and model
the experimental treatments as fixed effects (S-Plus, MathSoft Inc.; now Insightful
Corp., Seattle, WA). Otherwise, comparisons were made by two-tailed t-tests with
unequal variance estimates.
Results
With a long-term goal of exploring the response of joint capsule tissue to joint injury and
possible interactions with the cartilage, we investigated the effect of injuring cartilage
and then coincubating it with joint capsule tissue. In these experiments, 6-mm-diameter
capsule disks were coincubated in the same culture well as three to five 3-mm-diameter
cartilage disks in 1.0 to 1.5 ml of medium. We have previously reported that after 6 days
of culture under these conditions, radiolabel incorporation into the coincubated cartilage
is decreased by -70% for
incubated alone.
35
S-Sulfate
and by -50% for 3H-proline, compared to cartilage
We did not observe adherence of capsule tissue to culture wells and
cartilage until after the first week in culture. Significant attachment between cartilage
and capsule tissue from normal joints has been reported in previous coincubation
systems.100 However, this may have been due to the use of centrifuged minced
synovium. Consistent with our results, another group has reported no difference in GAG
loss between cartilage in contact with and cartilage separated from synovium. 3 9
Coincubation of injured cartilage results in a synergistic increase in proteoglycan loss
In a factorial design, cartilage disks were subjected to no treatment, coincubation with
capsule tissue, injury, or injury plus coincubation.
Both treatments (injury and
coincubation) were begun 3 days after harvest, and GAG loss to the medium was
measured 3 days after treatment. Since the GAG content of the capsule tissue was
negligible compared to the cartilage (data not shown), all GAG was attributed to release
75
from the cartilage. Data from four separate experiments (each using cartilage from a
different animal) is combined and summarized in Figure 1. A mixed effects model was
used to account for experiment-to-experiment variation and test for the interaction
between the effects of coincubation and injury on GAG release. The synergistic
increase in GAG release from injured cartilage coincubated with capsule tissue
compared to either injury or coincubation individually was statistically significant (p =
0.003).
This data was similar to our previous observation of a strong synergistic increase in
GAG loss from injured cartilage treated with the cytokines IL-1 and TNFu. These
cytokines
are
also
well-described
inhibitors
of cartilage
biosynthetic
activity.
Furthermore, IL-1 is well-documented as a molecule which can be released by excised
normal synovial membrane in culture,10 1 and appears to be responsible for much of the
proteoglycan-degrading activity of medium conditioned by inflamed synovium. 10 2 We
therefore proceeded to test whether the effect of joint capsule coincubation on cartilage
was consistent with a protein factor released from the joint capsule tissue and whether it
was abolished by blockade of IL-1 and TNFx.
Characterization of the inhibition of the biosynthetic activity of cartilage by coincubation
with joint capsule
For these further investigations, we focused on the inhibition of radiolabeled sulfate
incorporation, which is a measure of proteoglycan synthesis in newborn bovine
cartilage.
This inhibition was the most prominent effect of coincubation in our model,
and has also been less well characterized by previous investigators. To allow for more
efficient sampling in these experiments, coincubation conditions were slightly modified:
each cartilage disk was incubated together with a 3-mm-diameter sample of capsule
tissue in a separate well of a 96-well plate.
Under these conditions, data from five separate experiments showed that inhibition of
biosynthesis was as profound as in our initial results.
71
Three days after incubation with
joint capsule tissue, sulfate incorporation into cartilage disks was decreased by 72 ± 5%
76
compared with control cartilage that was incubated alone (p < 0.001 for effect of
coincubation by mixed-effects analysis, 27 total samples per treatment group). Similarly,
coincubation reduced proline incorporation by 51±4 % (p
< 0.001).
However,
coincubation of injured cartilage here showed no evidence of a synergistic loss of GAG
(p = 0.78, data not shown).
Inhibition of proteoglycan biosynthetic activity of cartilage by conditioned medium from
joint capsule
To test whether the inhibitory activity is due to a protein factor released from joint
capsule tissue, we next incubated cartilage disks with conditioned medium from joint
capsule tissue. Inhibition of sulfate incorporation by incubation of cartilage with capsuleconditioned medium was not significantly difference from the inhibitory effect of direct
coincubation (data not specifically shown; inhibition of sulfate incorporation by
conditioned medium was greater than 50% in Fig 2A and 2B). This demonstrated that
the inhibitory effect of coincubation was due to a factor released from the capsule tissue
in culture, consistent with the observations of Jubb and Fell (1980).
To determine whether the inhibitory effect of the capsule-conditioned medium was
reversible, we performed a recovery experiment. Radiolabel incorporation into cartilage
was measured in four groups of cartilage, which were cultured in either standard
medium or capsule-conditioned medium and then either radiolabeled immediately, or
washed with fresh standard medium and allowed to recover for 5 days in standard
medium before radiolabeling (Fig. 2A). Incorporation was normalized to the mean value
of cartilage labeled immediately after incubation in standard medium. In cartilage
labeled
immediately
after
in
incubation
capsule-conditioned
medium,
sulfate
incorporation was decreased by 63%. This inhibitory effect was significantly reversed by
recovery (p < 0.001 for interaction by two-way ANOVA). The recovery of cartilage
biosynthetic activity suggests that the inhibitory effect of capsule-conditioned medium
was not attributable to chondrocyte toxicity.
77
The inhibitory activity of conditioned medium was heat labile
Conditioned medium was collected daily from joint capsule tissue cultures and pooled.
Standard medium and conditioned medium were boiled for 10 minutes, cooled, and
sterile filtered. Cartilage disks were then incubated for 3 days in standard medium,
conditioned medium, boiled standard medium, or boiled conditioned medium. After the
3-day incubation, radiolabel incorporation was measured and normalized to the mean
incorporation of the standard medium group (Fig. 2B). Conditioned medium decreased
sulfate incorporation by 59%. This effect was significantly reversed by boiling (p < 0.001
for the interaction term by two-way ANOVA). Inhibition of proline incorporation by
conditioned medium was also significantly reversed by boiling (p < 0.001 for interaction
by ANOVA, data not shown). The loss of the inhibitory activity of the boiled conditioned
medium suggests that the factor released from capsule tissue is a heat-labile and likely
a protein that is denatured on boiling.
IL-1 blockade had no effect on the inhibitory factor
We proceeded to test whether the inhibition of cartilage synthesis by coincubation was
mediated by IL-1 by blockade with recombinant human IL-1 soluble type II receptor
(slL-1 r) (Fig. 3A). Coincubation inhibited sulfate incorporation into the cartilage by 80%.
Addition of 5 pg/ml sIL-1r to the medium had no more effect on coincubated cartilage
than on control cartilage incubated alone (p = 0.32 for interaction by ANOVA). As a
positive control for the activity of sIL-1r blockade, the effect of 1 ng/ml rhlL-1
was
significantly reversed by sIL-1r (p < 0.001 for interaction by ANOVA). The results were
similar for proline incorporation (p = 0.87 for interaction between coincubation and slL-1 r
treatment by ANOVA, data not shown).
Since recombinant bovine IL-1 was not available, we also tested the effect of blockade
with recombinant human IL-1 receptor antagonist (IL-1RA). IL-1RA binds directly to the
IL-1 receptor, so cross-reactivity with the bovine receptor would be required in order to
block the effect of rhIL-1. The results for blockade with IL-1 RA (Fig. 3B) were similar to
those observed for blockade with IL-1 sr. Addition of 250 ng/ml IL-1ra had no significant
78
effect on the inhibition of sulfate incorporation by coincubation (p = 0.34 for interaction
by ANOVA) but significantly reversed the effect of 1 ng/ml hrlL-1c
(p < 0.001 for
interaction by ANOVA). Similarly, IL-1 RA had no significant effect on inhibition of proline
incorporation by coincubation (p = 0.46 for interaction by ANOVA, data not shown).
These data suggest that the factor responsible for the inhibitory activity of capsuleconditioned medium is not IL-1. However, it is possible that the human slL-1 r or IL-1 RA
did not cross-react with bovine IL-1a or IL-1 P even though the sIL-1r and the IL-1RA
blocked the effect of rhlL-1a.
Combined blockade of IL-I and TNF had no effect on the inhibitory factor
To test the effect of blockade of TNF in our system, we used etanercept, a soluble FcTNF receptor Fc fusion protein. Since etanercept is based on the human p75 TNF
receptor, we tested whether etanercept also blocks the effect of bovine TNFa (Fig. 4A).
After incubating cartilage for three days with 100 ng/ml bovine recombinant TNFu,
sulfate incorporation was reduced by 21%
compared to controls. This effect was
significantly reversed by addition of 25 pg/ml of etanercept to the culture medium (p <
0.001 for interaction by ANOVA).
Having shown that etanercept blocks the effect of bovine TNFa in bovine tissue, we
tested the effect of combined blockade with 25 pg/ml of etanercept and 250 ng/ml IL1 RA on cartilage radiolabel incorporation, cultured alone or coincubated with joint
capsule (Fig. 4B). After three days of culture, coincubation reduced sulfate incorporation
by 72%. Addition of IL-1 RA and etanercept had no significant effect on coincubation (p
= 0.36 for interaction by ANOVA). As a positive control for the blockade, inhibition of
sulfate incorporation by 1 ng/ml hrlL-1a and 25 ng/ml hrTNF was significantly affected
by addition of IL-1 ra and etanercept to the medium (p<0.001 for interaction by ANOVA).
Similarly, inhibition of proline incorporation by coincubation was not significantly affected
by combined blockade with IL-1ra and etanercept (p = 0.75 for interaction by ANOVA,
data not shown).
79
Coincubation inhibits cartilage synthesis in human donor tissue
The effect of coincubation was tested in tissue from a human donor knee joint. For
consistency with prior human tissue experiments, after equilibration for three days in
standard medium, capsule and cartilage tissue was transferred into a serum-free
medium (standard medium as above with no fetal bovine serum). Cartilage was then
either incubated alone or in the same well as joint capsule tissue. After 3 days, cartilage
was radiolabeled for 24 hours with 20 pCi/ml each of
35
S-sulfate and
3H-proline.
Radiolabel incorporation was normalized to the mean values of cartilage incubated
alone. As shown in Fig. 5, coincubation resulted in a -40%
decrease in sulfate
incorporation and a -20% decrease in proline incorporation.
Discussion
With the long-term goal of developing an in vitro injurious compression model that could
reveal interactions between the joint capsule tissues and the cartilage, we coincubated
cartilage tissue with joint capsule tissue excised from normal femorotibial joints. We
report here that in tissue from newborn calf joints, coincubation of joint capsule with
injured cartilage can cause a synergistic release of cartilage proteoglycan, and that
coincubation can strongly reduce chondrocyte biosynthesis even in the absence of
cartilage injury. This decrease in biosynthesis was observed for coincubation of
cartilage and joint capsule tissue from normal adult human post-mortem knee joints as
well. Based on these results, we then hypothesized that an inflammatory cytokine such
as IL-1 or TNFa might be responsible for this effect. Surprisingly, blockade of the
combination of IL-1 and TNFa had no effect on the inhibition of cartilage biosynthesis by
coincubation with joint capsule. Although it is unlikely, due to the use of human reagents
in the bovine system, we cannot completely rule out a role for IL-1 or TNFu in this
system. Initial characterization of the effect of coincubation showed that an inhibitory
factor was released from capsule tissue, that the resulting inhibition of cartilage
biosynthesis was reversible, and that the activity of the factor was lost on boiling. This
suggests that the inhibitory factor was nontoxic, heat-labile, and likely a protein.
80
The observation that coincubation of injured cartilage with joint capsule tissue can result
in synergistic increase in proteoglycan release, taken together with our previous report
that incubation of injured cartilage with IL-1 or TNFu results in a marked synergistic
increase in proteoglycan release, further suggests an aspect of how cartilage
degradation following joint injury may involve interactions with factors such as
inflammatory cytokines that are released by other joint tissues. The mechanism for the
synergistic proteoglycan loss is not yet known. It could reflect synergistic activation of
degradative pathways at the cellular level. Alternatively, it may be caused directly by
mechanical damage to the matrix. This could lead, for example, to more rapid transport
of signaling molecules, or to increased exposure of matrix molecules to degradative
enzymes.
However, while the
synergistic loss of cartilage
proteoglycan
was
reproducibly observed in a set of four experiments, this effect was sensitive to specifics
of the culture conditions and not observed in follow-up experiments that focused on the
inhibition of cartilage synthesis.
Our data suggests that the inhibition of chondrocyte biosynthesis by excised capsule
tissue is due to an IL-1-independent mechanism. Despite the isolation of IL-1 (catabolin)
from minced porcine synovial tissue from a normal MCP joint,101 and blockade of
proteoglycan loss from cartilage coincubated with inflamed synovium by anti-IL-1
antibodies, 01 2 1 0 3 to our knowledge the inhibition of cartilage biosynthesis by excised
capsule tissue from a normal joint has not been specifically characterized in previous
reports. In our studies, this inhibitory effect was not only strong but also quite
reproducible despite variability in joint capsule tissue composition, as no attempt was
made either to isolate the synovial membrane nor to ensure its presence in each
capsule tissue sample. In addition, our initial data (not shown) suggested that
coincubation of cartilage with other synovial joint connective tissues, including ligament
and tendon, also had a strong inhibitory effect on cartilage biosynthetic activity. In this
respect it is interesting that it has been reported that a cartilage-degrading factor was
released even by cultured heart valve and aorta.10 4 Finally, we show that coincubation
of normal adult human knee cartilage with joint capsule tissue from the same joint
81
inhibits chondrocyte biosynthesis, providing additional evidence for the general nature
of this inhibition.
The observation of an inhibitory effect in human tissue also suggests opportunities for
further studies since blockade of selected mediators could be more readily tested in
human tissue. In addition to IL-1 and TNFu, several members of the gp130-binding
cytokine family sharply inhibit cartilage biosynthesis (e.g., IL-6, oncostatin-M, and
leukemia inhibitory factor).10 5 Interestingly, chondrocytes show little response to IL-6
without the addition of exogenous IL-6 soluble receptor, which could be provided by the
1 05 107
cells of the capsule tissue. The reason that the joint capsule tissue from a normal joint releases of a mediator that
inhibits chondrocyte biosynthesis is not clear. One would suspect that the release of this
factor is related to the excision of the tissue, the change in its mechanical and
biochemical environment, and loss of blood supply when placed in culture. It is possible
that trauma induced by excision of the capsule tissue might mimic certain aspects of
intact joint injury. That is, like the chondrocytes, the capsule fibroblasts and/or
synoviocytes may be sensitive to mechanical disruption of their matrix caused by a joint
injury, even without a gross tear or puncture of the capsule.
However, it is possible that other aspects of capsule excision and culture lead to the
release of the inhibitory protein. For example, just the act of slicing cartilage tissue has
been shown to release matrix stores of bFGF. 78 Similar mechanisms which do not
depend on protein signaling and synthesis may be involved here. Necrotic cell death
has recently been shown to result in release of high mobility group box-1 protein
(HMGB-1 ),10 8 a recently identified cytokine that can activate synovial macrophages. 109
Since excision of the joint capsule leads to loss of its normal blood supply, it is likely that
some capsular fibroblast and synoviocyte cell death occurs, and it is therefore possible
that the release of HMGB-1 from dead cells is responsible for the inhibition of cartilage
biosynthesis. Although not tested here, and the effect of HMGB-1
on cartilage
biosynthetic activity is not known, this hypothesis is interesting as it would be consistent
82
with cytokine release after excision and culture of a variety of vascularized tissues,
while predicting no release in vivo or in coculture experiments with isolated cells.
83
D No Load
U Injury
50
40
-
30
-
20
-
0
-I
0
100
Cartilage Alone
Coincubation
Figure 1. Cartilage was subjected to injurious compression to 50% strain at 1 mm/s
(injury), coincubated with joint capsule tissue (coincubation), or both, and GAG loss
from cartilage was measured after three days of culture. Cartilage that was injured and
then coincubated released significantly more GAG than the additive effects of
coincubation or injury alone (p=0.003 for interaction by ANOVA, 13 samples/group over
4 experiments). The box plots display range (brackets), interquartiles (boxes), and
median (stripes) for the unadjusted data.
84
C_ 1.4
0
1.2
0~
1----r-------
1.0
0
C
0.8
N Recovery
0.6
(D
0
0.4
0.2
1
0.0
Standard
Medium
a)
CO
Capsule Cond.
Medium
U)0
1.4-
0
1.2-
0
1.0
I
CL
0
z
--
0.8
-
0.6
-
0.4
-
0.2
-
EBoiled
0.0 -i
Standard
Medium
Capsule Cond.
Medium
Figure 2. A) Cartilage disks were incubated for 3 days in either standard medium or
capsule-conditioned medium and then either radiolabeled immediately (white bars), or
allowed to recover for 5 days in standard medium before radiolabeling (black bars).
Sulfate incoporation was normalized to the mean value of disks incubated in standard
medium and labeled immediately. The inhibition of sulfate incorporation by capsuleconditioned medium was significantly reversed by recovery (p<0.001 for interaction by
ANOVA, N=5 disks per treatment group). B) Medium was boiled for 10 min. and sterilefiltered. Cartilage disks were incubated for 3 days in either standard medium or capsuleconditioned medium (white bars), or in medium that was boiled (black bars). The
inhibition of sulfate incorporation by capsule-conditioned medium was significantly
reversed by boiling (p<0.001 for interaction by ANOVA, N=6 disks per treatment group).
85
01.2
soluble
IL-1 receptor
& 1.0
0
-. 6 0.4
0
0.2
0
A
Control
Coincubation
IL-1
IL-1 receptor
antagonist
0
0 1.0 Z 0.0o
0.60.40.2 -
B
Control
Coincubation
IL-1
Figure 3. IL-1 Blockade. Sulfate incorporation was measured after cartilage was
incubated for 3 days alone (control) or in the presence of joint capsule tissue
(coincubation). IL-1 was blocked by addition of 5 pg/ml human recombinant soluble IL-1
receptor (3A, IL-1sr) or by addition of 250 ng/ml of human recombinant IL-1 receptor
antagonist (3B, IL-1ra). In both experiments, IL-1 blockade had no significant effect on
the inhibition of sulfate incorporation by coincubation (both p=NS for interaction of
coincubation and IL-1 blockade by ANOVA). As a positive control for IL-1 blockade, in
both experiments IL-1 blockade had a significant effect on the inhibition of sulfate
incorporation by 1 ng/ml human recombinant IL-1 (both p < 0.001 for interaction by
ANOVA). N = 5 (3A) and 7 (3B3) cartilage disks per treatment group.
86
o
-
------
0
--
0.8.
0.
0
*25 pglml
r-.6-
Etanercept
1D.4
C0.6
0.6
0.2
0
0.0
A
bovine TNF
no TNF
r- 1.2-
u
o
0-1
-u
IL- ra + etanercept
--
-
--
-
--
-
--
-
-
0
"0.8
~08
~0.6750.40.20
0.0-
B
Co-culture
IL-1+TNF
Control
Figure 4. Sulfate incorporation was measured after cartilage was incubated for 3 days
and normalized to the mean values of untreated cartilage. A) Activity of etanercept
against bovine TNF. Cartilage disks were incubated in 100 ng/ml bovine recombinant
TNFa (brTNF) with or without 25 pg/ml etanercept. The inhibition of sulfate
incorporation by brTNF is significantly reversed by addition of etanercept (p < 0.001 for
interaction of etanercept and brTNF by ANOVA; N = 8 cartilage disks per group). B)
Combined blockade of IL-1 and TNF. Cartilage disks were incubated alone or
coincubated with joint capsule tissue for three days. IL-1 and TNF were blocked by
addition of 250 ng/ml IL-1 ra and 25 pg/ml etanercept. However, the combined blockade
had no significant effect on the inhibition of sulfate incorporation by coincubation (p =
0.36 for interaction by ANOVA, N = 6 cartilage disks per group). As a positive control,
inhibition of sulfate incorporation by 100 ng/ml hrTNFa and 1 ng/ml hrlL-1i is
significantly reversed by incubation with etanercept and IL-1ra (p<0.001 for interaction
by ANOVA; N = 4-6 cartilage disks per group).
87
1.2C:
0
0
L
1.0
1l
-
+
M Proline
T
*
0.8-
I-
0
CO
Sulfate
0.6-
a)
0.40
z
0.2I
0.0
Coincubation
Cartilage Alone
Figure 5. Normal adult human donor knee cartilage and capsule tissue. Radiolabel
incorporation was measured after cartilage was coincubated for 3 days with joint
capsule tissue and normalized to the mean of control cartilage that was incubated
alone. Coincubation with joint capsule tissue significantly decreased incorporation of
both sulfate and proline compared to cartilage incubated alone (*p < 0.01 by t-test, N =
10 per group).
88
Chapter 5
Aggrecan cleavage and ADAMTS-4 processing in IL-1 -stimulated
bovine calf articular cartilage explants
This chapter is prepared for journal submission, to be co-authored by:
Parth Patwaril, Gui Gao 2,Vivian Thompson 2, Alan J. Grodzinsky', and John D. Sandy 2
89
Introduction
Enhanced release of proteoglycans from the articular cartilage to the synovial fluid is a
prominent feature of early and late osteoarthritis. The cause is not yet clear, but it
appears to involve an interplay between joint mechanics and chondrocyte response
which leads to a pathological excess of enzymatic degradative activity. Characterization
of enzymatic activity stimulated by cytokines such as interleukin-1 (IL-1) has led to the
identification of the aggrecanase family of enzymes, which cleave aggrecan, the largest
and most abundant proteoglycan in cartilage, at specific sites on the aggrecan core
protein.' 1 0 1 1 3 Aggrecanases have also been demonstrated to be active in cleavage of
proteoglycans in cartilage tissue and synovial fluid from patients with osteoarthritis.11 4
As a highly-glycosylated proteoglycan, electrostatic repulsion among the negativelycharged GAG chains of aggrecan contributes a large portion of the compressive
strength of cartilage, and depletion of aggrecan from the cartilage leads to loss of the
mechanical function of the tissue.
The steps involved in biosynthesis and activation of the proteolytic activity
of
ADAMTS-4 (aggrecanase-1) have recently been described in a human chondrosarcoma
cell line stably transfected with full-length ADAMTS-4. It appears that furin-mediated
cleavage of proenzyme followed by additional MMP-dependent cleavage(s) within the
C-terminal region are required for full aggrecanase activity.'1 5'1 1 6
ADAMTS
described, 1
However, the
protein forms present in native cartilage tissue have not been fully
7
and processing of these forms in tissue stimulated with IL-1, which
induces aggrecanase activity, could be important for understanding the activation of the
enzyme in vivo.
In addition, the mechanism of aggrecanase inhibition by mannosamine (ManN) in IL-1treated cartilage explants is not yet clear. We have previously hypothesized1
8
that
ManN may act by blocking the synthesis of glycosylphosphatidylinositol (GPI)-linked
proteins, such as MT4-MMP, which may be required for C-terminal truncation of
ADAMTS. Since it has been shown that mannosamine blocks the appearance of the
90
fully active 60 kDa form of ADAMTS4 in the chondrosarcoma cell system,'
we decided
to examine whether its modulatory effects on IL-1-induced aggrecanase activity and
protection of biomechanical properties in native cartilage tissue is also accompanied by
alterations in the extent of conversion of the pro-ADAMTS-4 (68/75 kDa) to the fully
active 60 kDa form.
We examined aggrecan degradation in calf explants treated with IL-1 over a six day
time course, and further examined these samples for evidence of changes in the
abundance and molecular form of ADAMTS4. We show here that the most prominent
change in tissue abundance of ADAMTS4 with IL-1 stimulation was the loss of the p75
species from the tissue and the finding of increased abundance of both the p75 and p60
species in the conditioned medium. These effects of IL-1 on ADAMTS4 protein forms
were largely reversed by addition of 1.35 mM mannosamine.
Methods
Cartilage explant and culture
Articular cartilage disks were obtained from the femoropatellar groove of one-to-twoweek-old calves using methods similar to those described in detail previously.2 3 In brief,
9 mm diameter cylinders of full-thickness cartilage and bone were cored from the
articular cartilage and inserted into a sample holder of a sledge microtome. After
removing sufficient superficial cartilage to create a flat surface (usually less than 500
pm), the next two sequential 0.5 mm thick slices were cut. From each of these slices, a
dermal punch was used to produce 4 - 5 cartilage disks (3 mm in diameter and 0.5 mm
in thickness). Groups of cartilage disks were incubated at 37
C02,
0C
in an atmosphere of 5%
in wells containing 0.25 ml/disk of a serum-free culture medium (low-glucose
DMEM [Gibco, Grand Island, NY] with 100 U/mI penicillin G, 100 pg/ml streptomycin
and 0.25 pg/ml amphotericin B). Medium was changed every third day for timecourse
experiments. Medium was not replaced for 6-day experiments.
91
Treatment groups consisted of four cartilage disks which were incubated in either
medium alone (controls), medium plus 10 ng/ml human recombinant IL-1l
(R&D
Systems, Inc., Minneapolis, MN), medium plus 1.35 mM mannosamine
(Sigma
Chemical Co., St. Louis, MO), or medium plus both 10 ng/ml IL-1 and 1.35 mM ManN.
Each experiment had four to seven such treatment groups, allowing disks among
different groups to be matched for their original location along the surface of the joint.
Biochemical Analysis of Proteoglycans
Tissue and medium were analyzed for composition of aggrecan fragments by Western
blotting as previously described.11 9 Aggrecan remaining in the cartilage tissue was
extracted in 4M guanidine (1.2 ml per 75 mg wet weight) for 48 hr at 4 IC in the
presence of proteinase inhibitors. Tissue extracts and conditioned medium were
isolated by ethanol precipitation and deglycosylation. Portions of medium and tissue
extracts (corresponding to 10 pg GAG in each case) were loaded on 4-12%
polyacrylamide gels (Novex) for Western analysis and probed with polyclonal antisera to
the bovine G1 domain (G1-2), the G3 domain (TYKHRL), and the C-terminal
neoepitopes TFKEEE 1 687, TAGELE1 5 0 1 , and NITEGE 373 . Tissue (digested overnight with
proteinase
K)
and
conditioned
medium
that
was
analyzed
for
sulfated
glycosaminoglycan (GAG) content was assayed by reaction with dimethylmethylene
blue (DMMB) dye, with known quantities of shark chondroitin sulfate (Sigma) used as
standards.
Western analysis of ADAMTS-4 (aggrecanase-1)
Pilot studies were done to optimize the extraction conditions for ADAMTS-4 in cartilage.
Cartilage (15 mg wet weight) was freeze-milled in a Biopulverizer (Biospec Products,
Bartlesville, OK) and the powder was extracted for 20 h at 4 'C in 3 volumes (pl/mg wet
weight) of a) 50 mM Tris HCI, pH 7.0; b) 50 mM Tris HCI, 100 mM NaCl, pH 7.0; c)
50 mM Tris, 100 mM NaCl, 0.5% Nonidet P-40, pH 7.0; or d) 4M guanidine, 50 mM Tris
HCI, pH 7.0. Macromolecules in the 4M guanidine extract were purified by Sephadex
G50 and DE 52 as previously described for aggrecan fragments and the unbound
(protein) and bound (proteoglycan) fractions were analyzed. Other cartilage samples
92
were treated at 100 degrees for 10 min in 50mM Tris HCI, 2.5% SDS. The results (not
shown) indicated that the highest yield of immunoreactive product was consistently
obtained with extractant c). Poor recoveries were seen with 4M guanidine extracts and
boiling in SDS. Therefore, all data shown in the present paper were obtained by
extraction in 50 mM Tris, 100 mM NaCl, and 0.5% Nonidet P-40 at pH 7.0, and dilution
of 5-10 pl of this extractant in gel-loading buffer before SDS-PAGE.
Western analysis was done with two affinity-purified antibodies which exhibit very
different reactivities for the different molecular forms of ADAMTS-4 (see Gao et al."').
Anti-VMAH was raised to a peptide (Va1394 -Pro 4 03 ) in the catalytic domain. It detects
most forms which contain the catalytic domain with equal sensitivity. Anti-YNHR was
raised to a peptide (Tyr5 90-Pro6 O3 ) within the Cys-rich region. It reacts with the 75 kDa
form of ADAMTS-4 much more strongly than with other forms (compare rTS4 on Figs. 3
and 4). All soluble samples were loaded on an equivalent tissue weight basis and
separate blots were probed with anti-YNHR or anti-VMAH.
Results
Kinetics of aggrecan degradation in calf explants treated with IL-1
The kinetics of aggrecan degradation in this explant system was established by
treatment of cultures with IL-1 cc at 1, 10 and 100 ng/ml for 6 days. The percent of total
GAG remaining in the tissue at 2, 4 and 6 days is shown (Fig. 1), and since we were
interested
in
identification of intermediates of both
aggrecan and ADAMTS-4
processing, we chose the 10 ng/ml condition for a more detailed analysis.
Explant cultures treated with 10 ng/ml IL-1 were terminated at various times (0, 0.5, 1,
1.5, 2, 2.5, 3, 4.5, and 6 days) during the 6-day experiment. The GAG content of the
collected conditioned medium was measured (Fig. 2A), and Western analysis of the
aggrecan remaining in the tissue was performed with an antibody to aggrecan G1
domain (Fig. 2B). The composition of the control tissue (maintained in culture without IL1 for 6 days) and fresh tissue (maintained for zero days) was primarily core species 1
93
(full-length aggrecan) and low-abundance species a and b (see Sandy et al.
0
for
peptide notations). Early in the treatment period (days 1 and 1.5), species a and b were
essentially eliminated and from days 2-6 there was an accumulation of species 4 and 5.
The identities of species 4 and 5 were confirmed by Western analysis with anti-GELE
and anti-KEEE (data not shown). Most interesting was the finding that the N-terminal
product G1-NITEGE
(species 6 doublet) was not observed until day 4.5, which
corresponded to the time at which the GAG content of the tissue decreased sharply.
This finding is consistent with earlier work showing that the cleavage at Glu 373-Ala 374 is
the last of the cleavages to occur in the aggrecanase-mediated degradation of aggrecan
1
in vitro12 1 and in cartilage.
17 1 2 2
,
This conclusion was also supported by the analysis of medium fragments with antiaggrecan G3 domain (LEC-7), which showed that C-terminal processing occurs very
early in this system. Medium collections were made from explants terminated at days
0.5, 1, 1.5 and 2. After a medium change at day 2, further collections were made at
days 2.5, 3, and 4.5, and after a second change at day 4.5, a final collection was made
at day 6. The GAG content of the medium samples was measured, and Western
analysis was performed on an equal-GAG basis (Fig. 2C). Starting with the conditioned
medium from day 0-0.5, the analysis shows the four expected G3-positive species,
termed 7, 8, 9 and 10, that are generated by single aggrecanase-mediated cleavages at
Glu 1 890 , Glu1 790, Glu1 685 and Glu 1 499 respectively (numbering from accession number
P136081 2 3). Interestingly, the largest fragments, 9 and 10, were present in mediums
collected from days 2-2.5 and 2-3 but not in mediums from days 2-4 and 2-4.5. This
suggests that these larger species were further degraded in the medium to species 8
and 7 during the period from day 3-4.5, a finding consistent with their absence in the
medium collected over the day 4.5-6 period. Together this indicates that the medium
contains an active aggrecanase at these later stages of tissue degradation. The G3positive fragment at about 45 kDa presumably represents the product of further
proteolysis of species 7 by an unknown proteinase; however, its precise structure has
not been further investigated.
94
Analysis of explant system for ADAMTS-4 protein
Since it was clear that an active aggrecanase was present in this system and published
studies indicate that ADAMTS-4 (rather than TS-1 or TS-5) is mostly responsible for this
activity,
17,124-126
we analyzed both tissue and medium for ADAMTS-4 in control cultures,
IL-1-treated cultures, and IL-1-treated cultures in which the aggrecanase-mediated
degradation had been blocked by 1.35 mM mannosamine.
Extracts of untreated control tissue at days 0, 2, 4 and 6, probed with anti-VMAH (Fig.
3A), exhibited at least six TS-4 species which have been termed 1) p125, 2) p75, 3)
p60, 4) p42, 5) p40 and 6) p38. They appear to represent: 1) a proenzyme form; 2) the
product of furin-mediated removal of the prodomain; and 3, 4, 5 and 6) all products of
progressive MMP-mediated
15
or autocatalytic 1 6 removal of the C-terminal region of
the protein. Most forms appeared to be present in the fresh tissue at day 0, and their
abundance was not markedly altered during the 6 days of culture. The exceptions were
species 5 and 6, which were lost from the tissue and appeared in the medium. The
equivalent probe of these samples illustrates the relatively constant tissue content of
p75 over the 6 days.
In the IL-1 treated explants, there were marked changes in the tissue abundance and
distribution of some of these species. The most obvious change was the complete
disappearance of p75 from the tissue extract, which was especially noticeable on gels
probed with anti-YNHR (Fig. 3E), which is more reactive for the p75 form. The inclusion
of 1.35 mM ManN in IL-1-treated explant cultures at least partially blocked this loss of
the p75 species. However, the fate of the p75 in the IL-1-treated cartilage was not clear
from these results, since the abundance of ADAMTS4 forms detected in the medium did
not appear to be appreciably different between the IL-1 and IL-1 plus ManN groups with
anti-VMAH (3B), and neither the p75 nor the p60 species was detected in the medium
with anti-YNHR (3E).
To investigate this result further, an additional experiment was performed in order to
search for the presence of the p75 or lower-MW-forms in the unextracted tissue and in
95
the medium. The experimental groups were freshly harvested tissue (Day 0) and the
three treatments as above (control, IL-1, and IL-1 plus ManN), terminated after 6 days.
Western analysis performed with anti-YNHR again saw a clear loss of p75 from the
tissue on treatment with IL-1, but also detected the appearance of some p75 in the
medium. Since it was possible that the p75 in IL-1-treated cartilage had shifted to an
intracellular location or was present in further-processed forms which were not as easily
extracted, the residual unextracted tissue from this experiment was boiled in loading
buffer and Western analysis of the supernatant was performed. However, probing with
anti-VMAH revealed no changes in abundance of ADAMTS4 protein forms (results not
shown). We then further investigated the ADAMTS4 protein present in the conditioned
medium by attempting to remove the aggrecan from the medium and concentrating it
prior to electrophoresis. Western analysis with anti-YNHR (Fig. 4, medium 2) revealed
that the abundance of p75 in the medium was greatly increased, and the abundance of
the p60 form was increased as well, whereas no changes in lower-molecular-weight
forms were noted.
Discussion
The results we present here showed that under IL-1 treatment of bovine calf articular
cartilage, there was a loss of pre-existing p75 ADAMTS4 protein from the tissue, and
the appearance in the conditioned medium of increased levels of both the p75 and the
active p60 ADAMTS4 protein forms. The presence of 1.35 mM mannosamine appeared
to markedly, if not completely, block these changes in ADAMTS4 protein forms,
consistent with a role for ADAMTS4 in the aggrecan degradation observed. An
accompanying time course of aggrecanase activity under IL-1 treatment detailed a shift
from early aggrecanase cleavage in the CS-2 domain of aggrecan to a sharp increase
in interglobular domain cleavage between days 3 and 4.5.
In our study, Western analysis of cartilage tissue detected constitutive expression of
multiple ADAMTS4 protein forms, and though there was a shift of p75 from the tissue,
the overall abundance of the ADAMTS4 forms did not appear to be markedly
96
upregulated with IL-1 treatment. This is consistent with the results of Pratta et al.
127,
who show constitutive cellular expression of ADAMTS4 protein forms in articular
cartilage by immunostaining with anti-VMAH, and no marked changes under IL-1
treatment, although the protein forms detected by immunostaining are not clear. In
contrast, Tortorella et al., 1 1 7 also using anti-VMAH, detected ADAMTS4 protein forms at
approximately 60, 45, and 30 kDa in cytokine-stimulated bovine nasal cartilage, but not
in control tissue. Based on the pilot studies performed for the present study (see
methods), this difference in detection of constitutive expression of ADAMTS forms may
be explained by the different tissue extraction protocol used here, as guanidinium
chloride may not produce optimal extraction of ADAMTS4 protein from the tissue.
The identities and activities of the p75, p60, and lower-molecular-weight forms we found
to be constitutively expressed in the tissue have not been further investigated here.
However, the p75 has been previously identified as the result of furin-mediated
1 51 1 6
cleavage of the prodomain, and appears to have little aggrecanase activity. '
Furthermore, as aggrecanase activity, purified from the conditioned medium of IL-1stimulated bovine nasal cartilage, was originally reported to be mediated by a protein of
-62 kDa,113 the p60 form identified here is likely an active form of the protein.
The major findings in this study, the loss of p75 from IL-1-treated tissue and the
appearance of p75 and p60 in the medium, with reversal by the inhibitor of GPI anchor
formation ManN, are therefore quite consistent with the previously hypothesized model
of ADAMTS4 activation which requires C-terminal processing of the p75, by either
autocatalysis or by a GPI-anchored proteinase, to produce an active p60 aggrecanase.
The significance of the prominent loss of p75 to the medium is not clear, but as the IL-1 treated tissue degrades about 80% of its matrix GAG over 6 days, this may represent a
passive loss of GAG-associated protein. The recent demonstration of the affinity of
furin-cleaved ADAMTS4 for GAG and the presence of GAG-binding consensus
sequences at the C-terminal end of the molecule 1
6
helps to support this possibility.
Therefore, taken together with the lack of any other clear changes in ADAMTS4 protein
forms, we hypothesize that our finding of increased p60 in the medium of IL-1-
97
stimulated tissue represents the major active aggrecanase in our system. These results
do not, however, rule out the possibility that the p60 is also passively lost by association
with GAGs, or that transformation of already present and activated p60 to a more
soluble form is the major change induced by IL-1.
98
1. 0
0.8
cc
0.6
0(
0 0.4
*Controls
.1 ng/mL
10 ng/mL
0.2
100 ng/mL
0.0
0
4
2
6
Time (days)
Figure 1 Kinetics of aggrecan degradation. 0.5-mm-thick cartilage tissue explants were
treated with IL-1 a at 1, 10 and 100 ng/ml for 6 days in serum-free medium. The percent
of the total GAG content remaining in the tissue at 2, 4 and 6 days is shown as mean
SEM with N = 6.
99
1.0
0.8-
E
4)
W
0.6-
*J
r
0
0.4'
O
0.20.0
A
0
1
2
3
A
4
5
6
Time (days)
IL-1 Treatment (time in days)
A
0
1
1.5
2
3
2.5
4.5
6
a
4,5
b-
-4--148
6 *
M-60
B
4-42
Medium Collection Time Period (Days)
0-0.5 0-1 0-1.5 0-2 2-2.5 2-3
2-4.5
4.5-6
-250
10+*
-148
8-+-
+-60
G3frag
C
C
m
-
o
42
Figure 2 Cartilage explant cultures treated with 10 ng/ml IL-1 were terminated at
various times (0, 0.5, 1, 1.5, 2, 2.5, 3, 4.5, and 6 days) during a 6-day
experiment. The GAG content of the collected conditioned medium was
measured (A), and Western analysis of the aggrecan remaining in the tissue was
performed with an antibody to aggrecan G1 domain (B). Western analysis of
aggrecan released to the medium was performed (C) with an antibody to
aggrecan G3 domain (LEC-7).
100
Control
Tissue
Medium
Day: 0 2 4 6 2 4 6
z
IL-1
Tissue
IL- + ManN
Medium
0 2 4 6 2 4
6
Tissue
Medium
0 2 4 6 2 4 6
130-*
9448-
636-
A
B
C
130-il
z
48-
5 --a
6 -P
36-
-
D
E
F
Figure 3 Time course of ADAMTS-4 protein forms in cartilage tissue and medium.
Cartilage tissue was incubated in serum-free medium only (Control), 10 ng/ml IL-1
(IL-1), or IL-1 plus 1.35 mM mannosamine (IL-1 + ManN) for 0, 2, 4, and 6 days.
Western analysis was performed with anti-VMAH (A-C) and with anti-YNHR (D-F).
101
Tissue
kDa
Medium (1)
Medium (2)
C
C
203119-
91
O
48-
34Day
0
C
IL
IM
IL IM
Day6
IL
IM
Figure 4 Western analysis of ADAMTS-4 in cartilage tissue and medium, probed with
anti-YNHR. Cartilage tissue was incubated in medium without additives (C), with 10
ng/ml IL-1 (IL), or with IL-1 plus 1.35 mM ManN (IM) for 6 days. Medium(2) was
concentrated and dialyzed prior to loading on the gel.
102
Appendix A
Tissue Maturation and Antioxidants Alter the Apoptotic Response of
Articular Cartilage after Mechanical Injury
Bodo Kurz PhD, Angelika Lemke, Melanie Kehn, Christian Domm, *Parth Patwari MD,
*Eliot Frank PhD, *Alan J. Grodzinsky ScD, and Michael SchOnke PhD
Anatomisches Institut der CAU zu Kiel, Olshausenstr. 40, 24098 Kiel, Germany
and
*Massachusetts Institute of Technology, Cambridge, USA
103
Abstract
Objective To study the influence of tissue maturation and antioxidants on apoptosis in
bovine articular cartilage induced by injurious compression.
Methods Bovine articular cartilage disks were obtained from the femoro-patellar
groove of animals aged 0.5- to 23-month-old, and placed in culture. Cartilage disks
were pre-incubated overnight with the cell-permeable superoxide dismutase (SOD)
mimetic Mn(Ill)porphyrin (0-12.5 pM) or alpha-tocopherol (0-50 pM) and then injured by
a single unconfined compression to a final strain of 50% at a velocity of 1 mm/s. After 4
days of additional incubation, the disks were fixed and embedded for light and electron
microscopic investigations. Apoptotic cells were quantified morphologically by nuclear
blebbing on light microscopy. The antioxidative action of the SOD mimetic was
confirmed by histological examination of cartilage tissue after incubation with nitrobluetetrazolium.
Results Injurious compression induced significantly more apoptosis in cartilage disks
from newborn calves (22% of the cells) in comparison to cartilage from more mature
cows (2-6%). In cartilage from 22-month-old animals, the SOD mimetic reduced the
percentage of apoptotic cells induced by injury in a dose-dependent manner (with
complete inhibition by 2.5 pM), while alpha-tocopherol had no effect. Neither antioxidant
altered protein biosynthesis or cellular ultrastructure.
Conclusion Our data suggest that the apoptotic response of articular cartilage to
mechanical injury is affected by maturation and is mediated in part by reactive oxygen
species. The antioxidative status of the tissue might be important for prevention of
mechanically-induced cell death in articular cartilage.
104
Introduction
The etiology of osteoarthritis (OA), a degenerative joint disease, is still not fully
understood. Along with several other factors, mechanical overload and cell death may
1 1 2 8, 1 2 9
be important in contributing to the degeneration of articular cartilage. ,
Chondrocyte cell death must have important consequences in cartilage tissue since
these cells represent 1-10% of the tissue volume and have a very low regenerative
capacity.1 3 0 As each cell is responsible for the maintenance of its surrounding
extracellular matrix, cell death could play a significant role in the imbalance in
degradative activity that leads to OA.
In addition, it has been proposed that apoptotic cell death is present in cartilage tissue
from OA patients and may play a causative role in OA pathogenesis.2 4 2 7 In particular,
Hashimoto et al.25 reported that destabilization of rabbit knee joints induces apoptosis in
articular cartilage and that the prevalence of apoptotic cells is significantly correlated
with OA grades. This theory remains controversial, as other investigators were not able
to confirm this finding, 45 which may have been attributable to false-positive results of the
TUNEL assay.4 6 48 ' 6
However, several studies have shown that in vitro mechanical injury by various
2 8 4 85 2
compressive loading protocols can cause significant apoptotic cell death. , ,
Investigators from our groups have shown that injurious compression of newborn bovine
cartilage can induce apoptosis, as assessed by both TUNEL and by nuclear
morphology (that is, the presence of nuclear disintegration or blebbing), accompanied
by cartilage swelling, release of matrix proteoglycan, and loss of the anabolic response
to low-amplitude dynamic compression in the remaining cells. 1 9,2 8 We were therefore
interested in using this model to investigate additional aspects of mechanically induced
apoptosis. Since degenerative diseases are correlated with age and it is known that
biomechanical and biochemical properties of articular cartilage vary at different stages
of age and maturity,1 3 1" 3 2 we hypothesized that the apoptotic response of the tissue to
mechanical injury might also be affected by maturation of the tissue.
105
In addition to characterizing this apoptotic response in order to develop insights into the
mechanotransduction of mechanical injury, the induction of programmed cell death is
clinically interesting since it may be possible to prevent cell death after traumatic joint
injury
3
. We were therefore interested in testing whether apoptosis could be inhibited in
our model. Several authors have shown that reactive oxygen species (ROS) and
especially superoxides are involved in some of the pathways leading to programmed
cell death 1 3 4 -1 3 7. As a result, antioxidative substances have been shown to inhibit
apoptosis in a number of cell types, 136,138-140 including chondrocytes. 141
We therefore hypothesized that antioxidative scavenger mechanisms would influence
the induction of apoptosis by mechanical injury. This hypothesis is supported by a
recent report from investigators in our groups that a diet enriched with vitamins and
selenium increased the expression of antioxidative enzymes in articular cartilage and
significantly reduced the incidence of mechanically induced OA in the STR/1 N
mouse. 1 4 2 To test this hypothesis in our model, we used two molecules which are
considered to have different antioxidative functions. Manganese(ll) tetrakis (1-methyl-4pyridyl)porphyrin pentachloride (MnTMPyP) is a molecule with superoxide dismutase
mimetic properties, 14 3 1 45 and alpha-tocopherol (vitamin E) inhibits peroxidation of
membrane molecules by hydroxyl radicals. 146
The objectives of this study were therefore to test the influence of i) tissue maturation
and ii) the antioxidative scavengers MnTMPyP and alpha-tocopherol on apoptosis
induced by injurious compression of bovine articular cartilage.
Materials and Methods
Materials
The cell permeable superoxide dismutase mimetic (SOD-M) Manganese(ll)tetrakis (1methyl-4-pyridyl) porphyrin pentachloride (MnTMPyP) was from Alexis Biochemicals
(Fig. 1). All other chemicals were from Sigma unless otherwise specified.
106
Isolation and culturing of articular cartilage explants
Articular cartilage disks were obtained from the bovine femoropatellar groove as
previously described2 3 for animals ranging in age from 2-3 weeks (0.5 months) to 23
months. In brief, cartilage-bone cylinders (9 mm in diameter) were drilled perpendicular
to the cartilage surface and placed in a microtome holder. After removing enough
cartilage to create a level surface, the top 1 mm of the cartilage was sliced by the
microtome to produce a 1 mm thick slice. Five to six explant disks (3 mm in diameter x 1
mm in thickness) were punched out of each slice and equilibrated for 24 hrs in culture
medium (low glucose DMEM supplemented with 10% FBS, 10 mM HEPES buffer, 1mM
sodium pyruvate, 0.1 mM non-essential amino acids, 0.4 mM proline, 20 pg/ml ascorbic
acid, 100 U/ml penicillin G, 100 pg/ml streptomycin, and 0.25 pg/ml amphotericin B) in a
370C, 5% C02 environment. The explants were distributed among the different
experimental groups matched by their original anatomical location.
Injurious Compression
Before
injurious
compression, some of the explants were
pre-incubated
with
antioxidative substances overnight. Injury was then applied to groups of 3-4 cartilage
explants at a time.
The explants were placed in chambers and radially unconfined
compression was applied by an incubator-housed loading device as described
previously. 1 9,3 0 Controlled displacement ramps to 50% final strain were applied at a
ramp velocity of 1 mm/s (corresponding to a strain rate of 100%/second), and held for 5
minutes at the final strain. The force produced during compression was also recorded,
which allowed calculation of the peak stress applied to the cartilage during the injury.
After compression, the explants were cultured for another 4 days, and biochemical or
microscopic measurements were made as described below.
Histological detection of apoptosis
After mechanical compression and subsequent culture the cartilage explants were fixed
overnight using 4% paraformaldehyde and embedded in paraplast. Serial sections (7
pm) were taken through the entire thickness of the explant disks sagitally, immobilized
107
on glass slides, and stained with Mayer's hematoxylin to quantify the percentage of
cells showing nuclear blebbing. In some cases, sections were stained for the presence
of TUNEL-positive cells according to the manufacturer's protocol (ApopTag peroxidase
in situ apoptosis detection kit, Oncor, Gaithersburg, MD). Three to five sections were
evaluated from each explant disk. Since cutting of the explants induces apoptosis at the
edges of the tissue, the margins of the sections (-150 pm thickness) were excluded.
Using a Zeiss Axiophot with a 40x objective, positive and negative cells were counted in
3 optical fields in each section (60-100 cells/field). One optical field was located in the
center of the explant sections and 2 were located more peripherally, close to the
corners of the sections (but not including the margins). The mean value from each field
was recorded. In a secondary analysis, cell apoptosis rates in the central and peripheral
fields were compared to each other. Encoded labels were used on all samples to
ensure blind scoring.
Ultrastructural detection of apoptosis
To further validate the determination of apoptosis by nuclear blebbing on light
microscopy, ultrastructural changes were evaluated by electron microscopy. Explants
were fixed with 2.5% glutaraldehyde in PBS overnight at 40C, washed and incubated for
1 hour with 1% Os04. Samples were embedded in araldite and 50 nm thick sections
were prepared, incubated for 15 min with saturated uranylic acetate in 70% methanol,
hydrated in descending concentrations of methanol and incubated in a lead-citrate
solution under a C0 2-poor atmosphere for 5 min. The sections were visualized using a
ZEISS EM900.
Biosynthetic activity
The wet weight of each cartilage explant was measured before injury. After the
experimental treatment, the explants were placed in fresh culture medium containing 10
pCi/ml 3H-proline for 18 hours under free-swelling conditions. Unincorporated radiolabel
was removed by washing in PBS containing 0.5 mM proline (the solution was replaced
three times after 20 min each) and digested overnight in 1 ml of papain solution per
explant at 650C (papain 2.125 U/ml in 0.1 M Na 2 HPO 4 , 0.01 M Na-EDTA, 0.01 M L-
108
cysteine, pH 6.5). 100 pl/sample were added to 2 ml scintillation fluid and measured in a
Beckmann scintillation counter. Counts were expressed in cpm/mg wet weight and
normalized to the radiolabel incorporation of uninjured control tissue, which was set to
100%.
Statistical analysis
To test the effect of maturation on cell apoptosis rate after injury, apoptosis was
measured in paired unloaded control and injured cartilage explants from animals of
different ages. Because the resulting data included both animal-to-animal variation as
well as age-related effects, a mixed-effects model9 9 was required to analyse the data.
Age was modelled as a fixed effect, each animal was modelled as a random effect, and
the paired difference in apoptosis rate (apoptosis after injury minus apoptosis in control
for each pair of matched explants) was used as the outcome measure. The model was
fit by restricted maximum likelihood estimation of parameters in S-Plus (MathSoft; now
Insightful Corp., Seattle, WA).
Dose-response experiments with 4 doses SOD-M and alpha-tocopherol were designed
to test the effect of ROS scavengers on apoptosis produced by injury. The effect of
dose on the paired difference in apoptosis rate between injured and control explants
was tested with a linear regression model. A nonlinear response was assumed, and
doses were applied over an exponential range. The data were linearized by rank
transformation of both dose and outcome (i.e., dose was converted from 0, 0.0025,
0.025, and 2.5 pM SOD-M to 0, 1, 2, and 3).147 As two animals were used to generate
the cartilage explants, and the anatomical site of each cartilage explant was not
completely matched among all eight experimental groups, animal and site were also
included as covariates in the model.
An experiment to test the effect of pre-incubation in SOD mimetic also included data
from two animals and was analysed by linear regression with animal included in the
model as a covariate. Control and experimental groups were otherwise compared by
two-tailed Student's t-test and differences were considered significant at alpha = 0.05.
109
Results
Detection of Apoptosis
Apoptosis was demonstrated morphologically by counting cells showing nuclear
blebbing (a specific morphological indicator for apoptosis 4 8,1 4 8,1 4 9) in serial histological
sections of articular cartilage explants (Fig. 2A and B). In some experiments, apoptotic
cells were counted in histological sections after labelling with TUNEL staining (Fig. 2C
and D). Furthermore, nuclear blebbing was demonstrated by electron microscopy (Fig.
3B), confirming the presence of apoptosis in these samples.
There was no significant difference between the percentage of cells positive for
apoptosis by TUNEL staining vs. by nuclear blebbing in injured cartilage explants from
6-month-old cows (nuclear blebbing 7.3 ± 2.0% vs. TUNEL 8.0 ± 1.9%; n=3; mean
values ± SD). However, consistent with prior reports, we observed in preliminary
experiments that the results of the TUNEL assay were highly sensitive to small changes
in the protocol, so we primarily quantified apoptosis by histomorphometric examination
of chondrocytes on light microscopy in all the results that follow.
Apoptosis produced by injurious compression is dependent on maturation of the tissue
A single axial compression to a final strain of 50% at a velocity of 1 mm/s induced mean
peak stresses of 17 MPa in newborn tissue, 25 MPa in younger (6-16-month-old) tissue
and 29 MPa in 22-23-month-old tissue.
The cell apoptosis induced by injurious compression was measured by nuclear blebbing
in central fields of cartilage from five animals of varying ages (Fig. 4). Nuclear blebbing
was present in 22% of the cells in injured cartilage from newborn (0.5-month-old)
calves, compared to 5.9% (6 months), 5.4% (16 months), 2.3% (20 months) and 5.7%
(23 months), in tissue from more mature animals. Due to the nonlinear response of
apoptosis with age, age was dichotomised for the statistical analysis as newborn vs.
more mature (6 to 23 months). The change in apoptosis after injury was significantly
110
decreased in the more mature cartilage compared to the newborn cartilage (-18 ± 3%,
mean ± SEM, p = 0.01, in a mixed-effects model).
In contrast, when cells in the more peripheral corner fields were measured for nuclear
blebbing, the change in apoptosis after injury was not significantly affected by age (1 ±
4%, mean ± SEM, p = 0.87, in a mixed-effects model). This was attributable to
differences in apoptosis rates between cells in the central field and the peripheral field
which varied with maturity. In newborn tissue the percentage of apoptotic cells in the
central fields of the explants was higher than in the peripheral fields (central minus
peripheral rates: 7.1% ± 3.2%, mean ± SEM, N = 3 animals). In contrast, in the more
mature tissue (6 to 23 months old) apoptosis was more common in the peripheral fields
(central minus peripheral rates: -7.5% ± 1.4%, mean ± SEM, N = 4 animals).
MnTMPyP inhibits apoptosis produced by injurious compression
A control experiment was performed to confirm the antioxidative action of the SOD-M on
cartilage. Explants were incubated with 1.25 mg/ml nitroblue-tetrazolium (NBT) in Hanks
buffer with or without 2.5 pM SOD-M for 90 min. NBT is transformed predominantly by
superoxide anions produced by living cells into the blue dye formazan1 5 0 . Cryosections
showed that cartilage explants incubated with NBT alone were strongly stained (Fig.
2E), while addition of the SOD-M inhibited the formation of formazan (Fig. 2F).
Over the range of doses used here, neither SOD-M or alpha-tocopherol had any
significant effect on proline incorporation of unloaded cartilage explants from 23-monthold cows after 3 days of incubation (data not shown). Furthermore, incubation of injured
cartilage in SOD-M did not alter proline incorporation [Control: 100
+
9; loaded tissue:
95 ± 8; loaded tissue + SOD-M: 98 ± 9. Mean values ± SEM, data in % normalized to
control, n=4]. Finally, by electron microscopy there was no obvious change in
ultrastructural cell morphology of unloaded cartilage incubated with 2.5 pM SOD-M (Fig.
3A), unloaded cartilage incubated with alpha-tocopherol (not shown), or injured cartilage
incubated with 2.5 pM SOD-M (Fig. 3C).
111
In cartilage from 22-month-old cows, the SOD mimetic reduced the percentage of
mechanically-induced apoptotic cells significantly in a dose-dependent manner (p<0.001
for linear effect of rank-transformed dose on rank-transformed change in apoptosis after
injury, by linear regression, adjusted for animal and site) (Fig. 5), while alpha-tocopherol
(in concentrations up to 50 pM) had no effect on the apoptotic response (data not
shown). A concentration of 2.5 pM SOD-M was needed to decrease the apoptotic
response to the level of the uninjured controls.
To test whether the presence of the SOD mimetic during the injurious compression was
required to inhibit apoptosis, cartilage from 22-month-old cows was either incubated in
2.5 pM SOD-M
starting before injury, or incubated in 2.5 pM SOD-M
starting
immediately after injury. Control tissue was not loaded but was incubated in SOD-M at
the same times. Nuclear blebbing was measured four days after injury [Increase in
apoptosis rate after injury (apoptosis rate after injury minus apoptosis rate in uninjured
cartilage with the same SOD-M treatment): no SOD-M: 11.0 ± 1.8 %;
injury: 1.4 ± 0.1 %;
SOD-M before and after injury: 0.5 ± 0.1 %;
SOD-M after
mean values +/- SEM;
n=7]. Since the equal variance assumption of linear regression did not appear valid, the
outcome measure (change in apoptosis after injury) was rank transformed. Both
treatments significantly reduced apoptosis after injury, and in addition there was a small
but statistically significant reduction in apoptosis for treatment before and after injury
compared to treatment only after injury (p < 0.001 for all 3 comparisons by linear
regression, adjusted for animal).
Discussion
It has now been demonstrated by several investigators that mechanical compressive
injury can induce apoptotic death in chondrocytes under certain circumstances in
vitro. 28,48,2 We have shown here that there is a significant effect of tissue maturation on
the apoptotic cell death seen produced by injurious compression, with much higher cell
death in newborn bovine cartilage. We found that the apoptotic cell death can be
inhibited
by
incubation
with
the
superoxide
112
dismutase
mimetic
MnTMPyP,
demonstrating that this response to mechanical injury is mediated at least in part by the
generation of reactive oxygen species.
The reason for the increased apoptotic cell death in the most immature tissue after
mechanical injury is not known but this observation is consistent with that of Tew et
al.,5 6 who reported that cell death after cutting bovine cartilage explants was similarly
much higher in newborn tissue than in mature tissue. It seems very likely that the
increased cell death we observed in newborn cartilage is related to the increased
metabolic activity and mitotic rate of this tissue,44
'
as hypothesized by Tew et al.
In addition, the differences in apoptosis with maturation may be explained in part by
associated differences in the biomechanics and biochemical composition of the
cartilage. Our observation that there was more apoptosis in the central region of
newborn cartilage disks, but not more mature tissue, suggests a hypothesis for how this
could occur. As newborn bovine cartilage matures, the most prominent changes appear
to be an increase in stiffness and a higher collagen content.13 1 Therefore, the radially
unconfined compression applied here would be expected to produce more radial
bulging of the cartilage disks in newborn cartilage. This could explain the distribution of
cell death in newborn cartilage, since this radial strain would be at a minimum at the top
and bottom edges of the cartilage, where the peripheral sections exhibited less cell
death. In contrast, it is interesting to note that the peak stresses produced by injurious
compression to 50% strain at 1 mm/s were higher in the mature cartilage (consistent
with a stiffer tissue), suggesting that the peak stress was not the mechanical parameter
responsible for the decrease in the apoptotic response with tissue maturity.
In our in vitro experiment with articular cartilage tissue from the more mature, 2-year-old
cows, we found that a Mn(Ill) porphyrin SOD mimetic, MnTMPyP, prevented apoptotic
cell death that was produced by injurious compression, with complete inhibition by 2.5
pM. Control experiments confirmed that the prevention of apoptosis seen here is not a
consequence of reduced metabolic activity or cellular toxicity, as assessed by protein
synthesis and cellular ultrastructure. To our knowledge, this is the first report that
113
chondrocyte death after mechanical injury is mediated at least in part via the generation
of ROS. A linkage between mechanical injury and ROS would be consistent with the
report of Kaiki et al.1
3
that injection of hydrogen peroxide acted synergistically with
running activity to produce OA in rat knees. Although the origin of the ROS here
remains unknown, MnTMPyP inhibited a large proportion of apoptosis even when
added to the explants after injurious compression, so it is likely that the ROS are
generated primarily at a some point after injury in the pathway leading to apoptosis.
In contrast to the effect of the SOD mimetic MnTMPyP, there was no influence of alphatocopherol on apoptosis induced by injurious compression. Alpha-tocopherol is not able
to scavenge superoxides or hydrogen peroxide but inhibits peroxidation of membrane
molecules by hydroxyl radicals, 46 suggesting that the role of superoxide scavenging
may be particularly important in the cell death seen here. However, further studies
would be needed to specifically identify any particular ROS-mediated pathway, since
although scavenging of superoxides has been shown to be a major function of the Mn
(Ill) porphyrins in living cells, 1 4 3 1 4 4 these molecules can also scavenge molecules
descended from superoxides, such as peroxynitrite15 4 or hydrogen peroxide.1 55
It is not yet clear what the effect of a traumatic joint or cartilage injury is on chondrocyte
viability in vivo. The absolute levels of apoptosis generated by the injurious compression
model used here should not be interpreted as a simulation of those occurring clinically,
due to important differences in mechanical conditions. However, several initial clinical
investigations have recently reported a substantial increase in apoptotic cell death in
fragments obtained after intra-articular fracture 133,156 and in cartilage biopsies after joint
injuries.1 57 Although these reports appear to have relied primarily on TUNEL staining,
the induction of apoptosis after a discrete event in vivo would certainly be important to
investigate as a possible target for pharmacological intervention. It has been shown by
several
investigators that cartilage
apoptosis
can
be
prevented
by caspase
inhibitors.5 2 ,15 8 , 1 59 The results of this investigation, together with the evidence that a diet
enriched in antioxidants can reduce development of mechanically-induced OA in an
animal model,142 suggest that the antioxidative status of the tissue may have
114
importance as another possible target for prevention of chondrocyte death and
osteoarthritis.
Acknowledgments
The authors would like to thank Rita Kirsch, Claudia Kremling and Frank Lichte for their
excellent technical support. The project was also supported by a research grant from
the "Forschungsschwerpunkt Muskel- und Skelettsystem, CAU Kiel, Germany", and by
NIH grant AR-45779 (MIT).
115
CH 3
N+
H3C-N
Mr +
N ## N
N-CH3
-5CI-
N+
CH3
Fig. 1: Schematic formula of the superoxide dismutase mimetic Mn(lII)tetrakis
(1-methyl-4-pyridyl) porphyrin pentachloride.
116
C
D
-
=ANN
E
F-
-
Fig. 2: Light microscopy of cartilage tissue explants from 23-month-old cows. A and
B: Mayer's hematoxylin. In contrast to normal cells in unloaded explants (A, arrow),
paraffin sections of explants 4 days after injurious compression showed cells with
prominent nuclear blebbing (B, arrow; see also higher magnification in insert). C and
D: TUNEL staining (brown dye). TUNEL staining of unloaded tissue showed no
positive nuclei (C, arrow) while TUNEL-positive nuclei were present in tissue after
injurious compression (D), which again showed nuclear blebbing as well (D, arrow;
see also higher magnification in insert). E and F: Cryosections of articular cartilage
after incubation for 1 hour with nitroblue tetrazolium (NBT). Cartilage with NBT only
(E) showed strong transformation of NBT into the blue dye formazan in the
cytoplasm of the cells (E, arrowheads; see also higher magnification in insert) which
was reversed by addition of 2.5 pM superoxide dismutase mimetic (F, arrowheads).
Bars = 25 pm.
117
B
Fig.
$-K
3: Ultrastructure of articular chondrocytes 4 days after injurious
compression. A: Nucleus of a chondrocyte from uninjured control tissue. B:
Apoptotic nucleus of an injured chondrocyte showing nuclear blebbing (arrow).
C: Incubation with the superoxide dismutase mimetic (SOD-M, 2.5 pM) inhibits
nuclear blebbing of the cells and has no toxic influence on the ultrastructure of
articular chondrocytes in comparison to control tissue. Arrows in A and C show
cell organelles such as the Golgi apparatus and the rough endoplasmic
reticulum. Bar = 1.2 pm.
118
30
25
-
20
-
I
0
4I)
~.
0
17 Unloaded
U Injured
15-
010 0
50-
T
T
Y
T
0.5
16
6
20
23
Age (months)
Fig. 4: The percentage of cells showing nuclear blebbing 4 days after injurious
compression. The increase in apoptosis after injury was higher in cartilage from
newborn cows (0.5 months) compared to cartilage from more mature cows (6 to
23 months) (p = 0.01 by mixed-effects analysis; 18 injured and uninjured
cartilage disks from 5 animals). Values shown are mean +/- SEM.
119
8-
I
7-
L Unloaded
* Injured
610-
I
5-
0- 43-
I
21-
EI
00
0.0025
0.025
2.5
Superoxide Dismutase Mimetic (pM)
Fig. 5: Effect of a superoxide dismutase mimetic on apoptosis 4 days after
injurious compression in cartilage explants from 22-month-old cows. The
superoxide dismutase mimetic had a significant and dose-dependent effect on
the increase in apoptosis produced by injury (p<0.001 for linear effect of ranktransformed dose on rank-transformed change in apoptosis after injury, adjusted
for animal and cartilage site). Values shown are mean +/- SEM.
120
Appendix B
Preliminary Results: Effects of Injurious Compression on Degradative
Enzyme Expression and Activity
kDa
U C
U C
U C
ADAMTS5
MMP3
220
9766
_
4
p75
p60
46
30
ADAMTS4
Figure B1: Protein forms of degradative enzymes in conditioned medium
after injurious compression. Cartilage was equilibrated for three days,
washed into serum-free ITS medium, and either subjected to injurious
compression at 1 mm/s to 50% final strain (C) or left uncompressed (U).
After three days of further culture, medium was collected and analysed by
Western blotting with probes to the catalytic domains of aggrecanase-1
(ADAMTS4), aggrecanase-2 (ADAMTS5), or stromelysin-1 (MMP3).
Analysis of the conditioned medium from injured cartilage by Western blotting showed
an increase in abundance of an ADAMTS4 protein form, no obvious changes in
ADAMTS5 protein, and increased abundance of MMP3 protein forms. The finding of an
increased abundance of the p60 ADAMTS4 protein form is particularly interesting, as
this appears to be the active aggrecanase in this cartilage explant system' 15 116 (see
discussion above in Chapter 5). Further refinement of the ADAMTS4 protein extraction
process by Gao et al. now allows further refinement of Western analysis of conditioned
medium. In contrast, no effect was seen on ADAMTS5 protein. This result is consistent
with the report of Tortorella et al. 11 7 that ADAMTS5 message is constitutively expressed
but not induced by IL-1 or TNF alpha, although the relation between mRNA expression
and protein activity is not clear for either ADAMTS4 or 5.
121
Increased expression of MMP3 protein forms in the medium three days after injury
supports our finding of increased MMP3 message in the tissue within 24 hours after
injury (Chapter 3). However, analysis of aggrecan neo-epitopes is necessary to confirm
an increase in either MMP or aggrecanase activity. We therefore proceeded to analyse
the aggrecan fragments released to the medium.
U C
UC
kDa
220
66G1-TEGE 373
16 kDa
22 14
IL-1
Anti-NITEGE
NITEGE
Blocked
Figure B2: Western analysis of aggrecanase-generated aggrecan fragments in
conditioned medium after injurious compression. Cartilage was subjected to
injurious compression at 1 mm/s to 50% final strain (C) or left uncompressed
(U), and returned to culture for three days. Western analysis of aggrecan
fragments was performed on the conditioned medium with a neo-epitope
antibody for the xxNITEGE aggrecanase cleavage. As a positive control (IL-1),
incubation of cartilage with IL-1 was confirmed to release the GI-NITEGE
fragment into the medium. A 16 kDa aggrecan fragment was identified in the
conditioned medium of the injured cartilage (anti-NITEGE), and the specificity of
this band for the NITEGE antibody was confirmed by its disappearance on
blockade by excess NITEGE neo-epitope (NITEGE blocked).
The finding of release to the medium of a 16 kDa aggrecan fragment after injury is
intriguing but its identity is as yet unknown. Initial attempts to sequence the fragment
were not successful due to lack of sufficient protein. It is possible that this fragment
represents the 32-amino-acid protein segment released by aggrecanase cleavage at
TEGE 37343 74 ARGS and MMP cleavage at DIPEN 34 1 43 4 2 FFGVG. Small neo-epitope-
containing fragments have not been fully characterised, but this has previously been
122
suggested as the identity of a small -35 kDa aggrecan fragment isolated from porcine
laryngeal cartilage160
123
Appendix C
Quantification of OP-1 protein levels after injurious compression of
human donor knee and ankle cartilage
Most research on the effects of injurious compression of the cartilage has focused on
catabolic processes such as GAG and collagen degradation. However, the failure of a
normal anabolic response to repair the damage to the cartilage may play an equally
important role in the eventual destruction of the cartilage in OA. In fact many
researchers have observed evidence that chondrocytes undertake a repair response in
the early stages of OA. At the molecular level, an anabolic response may be stimulated
by several growth factors. Osteogenic protein-1 (OP-1), a member of the bone
morphogenetic protein (BMP) family, is also known as BMP-7. Already in clinical use for
treatment of bone fractures, it has also been shown to have anabolic effects on
chondrocytes 6 1 1,6 2 and may inhibit damage stimulated by inflammatory mediators1 3 .
Furthermore, it has been proposed that the response of normal adult human donor
tissue differs with age 164 and between the knee and ankle joint83 . We were therefore
interested in testing whether the level of OP-1 protein is upregulated after cartilage
injury in vitro, and whether this response varies between knee and ankle joint articular
cartilages.
Adult normal human donor cartilage was obtained from the Gift of Hope Donor Bank in
Chicago and injured as described previously
5.
The ankle and knee joints were shipped
to MIT on ice and bathed in DMEM. 3-mm-diameter by 0.5-mm-thick disks were
obtained from tibial plateau, femoropatellar groove, femoral condyle, and talar dome
cartilages (Fig. CI).
In selected ankle joints, cell viability was assessed by a live-dead assay kit (Molecular
Probes) using calcein AM and ethidium homodimer solutions. A 3-mm-diameter
cartilage disk was sliced in half through a diameter with a razor blade and then
sectioned again to produce a thin 0.5-mm by 3-mm rectangular slice. The slice was
incubated in the ethidium homodimer and calcein AM for 5 min. and visualized under
124
ultraviolet light. In an early donor joint shipment, cell viability was quite low and was
therefore not used experimentally. For future joints, blood was more thoroughly washed
from the joints before shipment, and cell viability has since been nearly 100% (Fig. C2).
Figure C1 Cartilage harvest from a normal adult human talar dome. The donor was a
60-year-old male (AM1940) who died of cardiopulmonary arrest. The right ankle joint is
shown before (right) and after (left) drilling of cartilage-bone cylinders during harvesting.
Figure C2 Cell viability in adult human donor talotibial cartilage. The donor cartilage is
from the same ankle joint shown in Figure C1. A 3-mm-diameter cartilage disk was
sectioned with a razor blade and is approximately 100 um thick. Calcein AM is
converted to a fluorescent molecule by esterase activity in cytoplasm of live cells
(green). Cells lacking membrane integrity allow ethidium to intercalate into nuclear
material and fluoresce red.
125
All cartilage was cultured in DMEM with 10% FBS and supplements as usual after
harvest. Three days after harvest, the plugs were washed into serum-free medium plus
HEPES buffer and antibiotics only. In general, injury consisted of unconfined
compression to 65% strain at 2 mm/s with no hold time, and control consisted of no
loading. However, earlier experiments used 10%-serum in the medium and was loaded
to only 50%. The changes were made because these earlier conditions did not reliably
cause any cartilage damage either visually or in terms of GAG loss. As per usual,
treatment groups were blocked for location (cartilage slice). Three days after injury,
cartilage disks were flash-frozen in liquid nitrogen. OP-1 protein levels were measured
by ELISA as previously described164.
Results are summarized for ankle cartilage (Table 1; Fig. C3) and for knee cartilage
(Table 2; Fig. C4). For early experiments where donor numbers have not been matched
with the data, the covariates listed in the table are not known for all observations. The
observed mean OP-1 levels increased after injury in both joints but did not reach
statistical significance (p = 0.38 for both by Wilcoxon signed rank test for paired
differences, injury vs. unloaded control, N = 7 donor joints). Due to the high variability in
the data, a larger sample size would be required to test our hypothesis.
126
Table 1. Ankle cartilage
GAG Loss
OP-1 (nq/q dry wt)
(ug/plu)
Donor
Name
AM1766
AM1787
AM1786
AM1819
AM1940
AM1980
Age
Gender
24
69
73
83
72
60
64
M
M
M
M
M
M
F
Collins Strain
Grade
50
50
1
0
60
0
60
65
0
65
0
65
0
Mean
N
SE
p-value (Wilcoxon signed rank test)
Control
Injury
10
130
48
91
29
53
195
144
138
116
79
85
84
111
80
108
Difference
after Injury
134
8
68
-12
56
31
-84
29
7
26
0.38
Difference
after Injury
-1.14
-0.78
0.33
0.00
0.27
1.43
0.02
6
0.37
200 -
150 -
-C
c -o
C
-0
cD.c
C
E
cL
50
0
0I
Injury
Control
Figure C3 OP-1 protein levels 3 days after injury in normal adult human donor ankle
tissue. OP-1 levels were increased by 29±26 ng/g tissue dry weight (mean ± SE),
which was not a statistically significant difference (p=0.38 by Wlicoxon rank sum test for
paired differences, N = 7 donor joints).
127
Table 2. Knee cartilage
GAG Loss
OP-1 (nq/q dry wt)
(uc/plug)
Donor
Name
AM814
AM1819
AM1940
AM1980
Age Gender
71
24
72
60
64
M
M
M
M
F
Collins Strain
Grade
2
2
1
Control
6
3
31
23
6
69
104
60
60
50
65
65
65
Injury
9
8
71
34
111
60
92
Difference Difference
after Injury after Injury
4
5
40
11
2.8
105
0.3
-9
1.1
-12
1.38
3
0.75
55
7
15
0.38
Mean
N
SE
p-value (Wilcoxon signed rank test)
200 -
150 0) C)
0) 0
~0
0)
100 -
C
0
E
50 -
C
-1
0 +
Injury
Control
Figure C4 OP-1 protein levels 3 days after injury in normal adult human donor knee
tissue. OP-1 levels were increased by 55±15 ng/g tissue dry weight (mean ± SE), which
was not a statistically significant difference (p=0.38 by Wlicoxon rank sum test for paired
differences, N = 7 donor joints).
128
Appendix D
Predictors of Proteoglycan Loss in Normal Human Post-mortem Knee
and Ankle Cartilage
Introduction
In vitro cartilage injury models may be useful for identifying the mechanical parameters
loading that are most responsible for cartilage matrix damage, as well as for injury to the
chondrocytes. This information may be important for two major reasons. First, it may
lead to characterization of tolerances of the cartilage matrix in terms of loading
parameters. For example, Newberry et al. have attempted to collect data that would
"...begin to identify ranges of tissue stress that could be defined as 'safe' and 'unsafe' in
terms of developing a chronic disease, osteoarthritis" in a rabbit model of impact-injuryinduced OA. 12 Second, identification of the important loading parameters for causing
chondrocyte injury may give insights into the mechanism of mechanotransduction that is
responsible for these effects.
The simplest possible result would be that a threshold for one loading parameter exists
that can separate injurious from non-injurious stimuli. In particular, Torzilli et al. have
suggested that cartilage injury may occur when peak stress exceeds a certain
threshold.
However, the variation in multiple loading parameters could be equally
important in explaining the observed effects of injury. We therefore analyzed results
from our model of injurious compression in order to develop a statistical model to
identify a set of predictors for the loss of proteoglycan (PG) from the cartilage after
injury, with particular emphasis on testing whether peak stress during injurious
compression to a fixed strain and at a fixed velocity is correlated with the loss of PG.
Injurious compression was performed on both knee and ankle cartilage from adult
normal human post-mortem donor joints, allowing for comparison of this analysis
between knee and ankle cartilage.
129
Methods
Injurious compression
Normal adult human post-mortem ankle and knee cartilage was harvested from tissue
donors, equilibrated for three days, and subjected to injurious compression to 65%
strain at a velocity of 2 mm/s, as previously described.5 3 Injured and location-matched
uninjured control cartilage was incubated for 3 days further, and the conditioned
medium was then collected and analyzed for PG content by the DMMB assay.
Statistical Analysis
The outcome measure of interest for this study was the increase in PG loss after injury,
calculated as a paired difference of PG loss from injured disks minus PG loss from the
location-matched control disk. A pre-specified step-wise procedure was used to select
predictors of the increase in GAG loss after injury in a linear mixed-effects model.
During model selection, all models were fit by maximum likelihood estimation and
comparisons were made by likelihood ratio test. In the final model, the reported p-values
were also calculated by likelihood ratio test but the coefficients were estimated by
restricted maximum likelihood.
The mixed-effect model included person-to-person variability as a random effect and the
possible predictor variables as fixed effects. These variables were: 1) donor age; 2) the
original anatomical location of the cartilage disks; 3) peak stress during loading; 4)
approximate swelling (change in thickness, as a percentage of original thickness)
between harvest and injury; 5) peak stress during loading (in MPa) and 6) damage
score after injury. Anatomical location included an indicator variable for whether the disk
was from the medial or lateral aspect of the articular surface (0 if medial; 1 if lateral); the
number of the core that the disk was from (cores were numbered from proximal to distal
along the femur and from anterior to posterior on the tibia and talus); and the depth of
the slice that the disk was from (0 for the top slice; 1 for deeper slices). After injury,
cartilage damage was graded by visual inspection as unchanged (0), slightly flattened
or non-circular (1), clearly elliptical (2), or grossly fractured (3).
130
Results and Discussion
Ankle cartilage was obtained from three donors, who ranged in age from 60 to 72 years
old. The cartilage surfaces were all rated as Collins grade 0. Injury produced peak
stresses of 13.9 ± 4.6 MPa (mean ± SD) and no observed damage. The resulting
increase in PG loss to the medium 3 days after injury (PG loss from injured cartilage
minus PG loss from matched uninjured controls) was 0.0 ± 1.2 pg/disk (mean ± SD).
Knee cartilage was obtained from the same three donors, and the cartilage surfaces
were rated Collins grade 1 to 2. Injury produced peak stresses of 13.5 ± 4.6 MPa and
the damage score ranged from 0 to 2 (mode, 0). The increase in PG loss after injury
was 1.5 ± 2.1 pg/disk.
We first examined the unadjusted linear relationship between peak stress and increase
in PG loss after injury (Fig. Dl).
Ankle Cartilage
6 -
5
5
4
4.
O -..: 3-
-
53
-2
. C
Knee Cartilage
7
6
2-
2@
2 %
--E10 -
* *,
.
-0 -
-2 -
1%
-2
5-3
-3
0
A
*..OS
10
20
0
30
B
Peak Stress (MPa)
10
20
30
Peak Stress (MPa)
Figure D1. Increase in PG loss after injury vs. peak stress during injury for human
knee cartilage (A) and ankle cartilage (B) from multiple donors. In knee cartilage, there
was a significant decrease in PG loss with increasing peak stress (p < 0.001 by mixedeffects analysis; N = 38 observations from 3 donors), whereas there was no significant
linear relationship in ankle cartilage (p = 0.51; N = 17 observations from 3 donors). The
dashed line indicates the estimated linear effect of peak stress.
131
Results for the model selection are shown in Tables 1 and 2 (N: knee, 38 total
observations from 3 donors; ankle, 17 total observations from 3 donors). After
adjustment for other covariates, peak stress remained a significant predictor for PG loss
after injury from knee cartilage (p = 0.001).
Table 1. Human Ankle Cartilage
Coefficient SE
0.46
(Intercept) 0.61
0.52
-1.06
Lateral
p value
0.03
Table 2. Human Knee Cartilage
Coefficient
7.22
(Intercept)
-0.80
TP
-0.20
Stress
-1.75
Depth
-2.38
Damage
Depth:Damage 1.57
SE
1.37
0.31
0.06
0.64
1.00
0.58
p value
0.007
0.001
0.004
0.01
0.02
Interestingly, in injured adult human knee cartilage from normal post-mortem donors,
loss of tissue PG did not increase with increasing peak stress during injury. This
contrasts with the conclusion of Torzilli et al. that cell death and collagen damage in
their model occurred above a critical threshold stress of 15 MPa. The differences
between the results of the two studies emphasize that the observed effect of any one
loading parameter is dependent on how the other loading parameters are varied in the
experiment. Torzilli et al. compressed the cartilage at a fixed stress loading rate until a
range of peak stresses were generated (0.5 to 65 MPa), and measured the resulting
strain. Since both peak stress and peak strain were increased together, it is not entirely
surprising that cartilage damage increased with increasing peak stress in this
experiment. The authors do report that cartilage damage increased with increasing
compressive strain as well, demonstrating that it is not simple to separate the effects of
the two loading parameters. Although they note that more variability is explained by
peak stress, they report similar correlation coefficients for peak stress vs. PG
biosynthesis (r = 0.61 for a second-order model) and for strain vs. PG biosynthesis (r =
132
0.68 for a linear model with strain). Similarly, correlation with the swelling ratio is no
better for peak stress (r = 0.57 for a second-order model) than for strain (r = 0.68 for a
linear model).
In the experiment we report here, cartilage is compressed to one fixed strain (50%) at
one fixed strain rate (400%/s), and the peak stress was recorded. This experiment is
therefore valid only for this particular strain and strain rate, but allows one to isolate any
effects of variation in peak stress. Under these circumstances, the peak stress
generated during injury in undamaged cartilage would in turn be expected to depend
strongly its strength in unconfined compression. The decrease in PG loss with
increasing peak stress in injured knee cartilage observed here is likely related to the
generation of inhomogenous tissue mechanical properties due to matrix damage during
injury. Once injury generates cartilage matrix damage, one would expect that the
apparent strength of the tissue would be reduced and that the peak stress would not
increase at the same rate as undamaged cartilage (i.e., the tissue "yields" during
loading). This hypothesis is supported by the selection of both damage score and the
interaction term of damage and slice depth in the model for prediction of PG loss after
injury. In addition, this may explain one aspect of the different response in ankle
cartilage - that is, the trend for a positive correlation between peak stress and GAG loss
may have been observed in ankle cartilage because the cartilage was not damaged by
injury.
Torzilli et al. also report that tissue yield was important in their experiment. In fact, at
higher peak stress compressions, the tissue never generated the peak stress that it was
nominally loaded to. As a result, the cartilage was compressed to nearly 100% strain,
which would clearly crush the tissue to pieces. Although this and other nonlinearities in
stress-strain behavior are discussed, the authors do not clarify whether most of the
damage they saw was predicted by "yield mode", which would complicate interpretation
of their results.
133
The importance here of the original anatomical position of the ankle cartilage, and the
slice depth of knee cartilage, emphasizes that there is significant variability of adult
human tissue with depth and location. Thus, the removal of the superficial zone may be
a more significant limitation of the injury model used here, in comparison to bovine
cartilage, which is much thicker and more homogeneous. Additionally, although only
cartilage which was normal to appearance was used in these experiments, we were not
able to assess here whether the differences between knee and ankle could have been
related to the higher Collins grade of the knee joint surfaces. Finally, with the
parsimonious model selection procedure used here, lack of inclusion in the model does
not imply sufficient power to exclude a predictor.
Despite these limitations, our results are consistent with previous reports to date which
have shown that ankle cartilage is stiffer in compression8 2 and less responsive to
catabolic stimuli.8 3 As in our initial study of injury in human donor cartilage,5 3 we show
here evidence, now from multiple donor joints, that unlike knee joint cartilage, ankle
cartilage
shows
little evidence for increased
proteoglycan
loss after
injurious
compression. Analysis of the data that included the effects of multiple predictors
suggested that this difference was most related to the complete lack of observed
damage to injured ankle cartilage.
134
Appendix E
Effect of Glucosamine on Cartilage Biosynthetic Activity
Current therapies for osteoarthritis are limited to drugs which relieve pain and surgical
replacement of the joint, leading to intense interest in developing therapies which
actually reduce progression of the disease. One such candidate disease-modifying
agent is the amino sugar, glucosamine (GIcN). For over 20 years, multiple clinical trials
have reported significant therapeutic effect of glucosamine. 6 5-1 70 However, all of these
studies were relatively small, short-term studies. In addition, they all were, or appear to
have
been,
sponsored
by the
Rotta
Research
Laboratorium,
the
European
manufacturer of GIcN and patent-holder for its therapeutic use.
More recently, Rotta conducted a clinical trial with long-term (3 year) follow-up and
reported a benefit both for pain relief and for joint space narrowing on radiography.1
71
Although this study is impressive for its length, size, and finding of improvement in an
objective measure of disease progression, clinical trials performed by investigators
unaffiliated with Rotta have reported inconsistent results. One smaller study found no
effect in primary outcome (difference in pain score between start and week 8 of
treatment) but positive trends for glucosamine in 23/24 subscores. 1 72 Two other studies
1 73 1 74
reported no benefit of glucosamine treatment vs. placebo. ,
In the United States, consumers are turning in large numbers to the de facto
unregulated dietary supplement industry for treatment of medical conditions for which
there are limited pharmaceuticals options available. 75 Therefore, in addition to the
evidence generated by recent clinical trials, the widespread sale of and consumer use
of glucosamine for OA treatment has led to a pressing need to know whether this
molecule is effective. In 1998, the National Center for Complementary and Alternative
Medicine concluded that a phase Ill study of glucosamine and chondroitin sulfate for
treatment of OA was needed, and a multi-center clinical trial is now underway (the
135
National
Institutes of Health
Glucosamine/Chondroitin Arthritis
Intervention Trial
[GAIT]). 1 76
Despite the clinical interest in GIcN, possible mechanisms of action are unclear. The
earliest hypothesis for the mechanism of glucosamine effect has been that as a
potential precursor for synthesis of chondroitin sulfate (CS), GIcN may increase CS
production. This hypothesis was based on studies from the 1970s that CS synthesis
rates were increased in chick embryo cartilage cultures 177 and in rat cartilage' 78 with
addition of up to 1 mM GIcN. Bassleer et al. 179 have also reported that aggregated OA
chondrocytes increased GAG accumulation with addition of 0.05 to 0.5 mM GIcN. There
is also good evidence that glucosamine is preferentially taken up by chondrocytes (vs.
glucose) and incorporated into newly synthesized GAGs in bovine steer articular
cartilage.18 0
Sandy et al. have recently reported that GIcN inhibits IL-1-induced aggrecanase activity
in chondrocyte cultures. 11 9 It is possible that this effect is due to depletion of intracellular
ATP levels, as the concentrations required to inhibit aggrecanase are similar to those
required to deplete adipocyte ATP in vitro.181
To further examine the effect of GIcN in cartilage tissue explants, we tested the effects
of GIcN over a range of doses for inhibition of aggrecanase-induced degradation of
mechanical properties. We have previously shown that GIcN is capable of reversing IL1-induced degradation in terms of GAG loss, equilibrium modulus, dynamic stiffness,
streaming potential, and hydraulic permeability. 182
136
C
0
N =4
(Control N = 3)
0
0
0.
0
o 1.0
N =5
(Control N = 10)
T
-- -
0.
0
0 1.0 -
- -
-
-
-
-
CD
N
N
E
0
0
z 0.0
Z 0.0
0
0.01
A
0.04 0.16 0.63
2.5
(CnrN6
0
10
B
[Glucosamine] (mM)
0.01 0.04 0.16 0.63
2.5
10
[Glucosamine] (mM)
(Control N =8)
12
o
0
o
0. 1.0----
E
0
-------------------------
1.0---
----TA
NN
o
0
0
0.01 0.04 0.16 0.63
2.5
10
a
o
0
D
[Glucosamine] (mM)
C
0.01 0.04 0.16 0.63 2.5
[Glucosamine] (mM)
10
N =4
(Control N =8)
0
0L
L.
0
.' 1.0 -
-- -- --
N Sulfate
0 Proline
N
0
z 0.0
.
0
E
.. .
0.01 0.04 0.16 0.63
2.5
10
[Glucosamine] (mM)
Figure El. Effect of glucosamine on incorporation of radiolabeled sulfate and
proline into 0.5-mm-thick cartilage tissue explants. Cartilage was cultured in
medium with glucosamine and radiolabel on the first day post-mortem. Tissue
was removed after approximately 24 hours, washed, digested, and radioactivity
measured by liquid scintillation. Each panel (A-E) represents the results of a
separate experiment, each with cartilage harvested from a different animal. All
data is shown as mean ± SEM.
137
We present here an investigation of the effect of GlcN on incorporation of radiolabelled
sulfate and proline in newborn bovine articular cartilage tissue explants in a series of
five experiments (Fig. El). Clearly, the animal-to-animal variability in the response of
the tissue to GIcN was substantial. The data was therefore analysed with a mixedeffects model99 in SPlus to account for variation with animal (a random effect) and
estimate effect of GIcN dose (a fixed effect). Since the doses were applied over an
exponential range, GlcN dose was rank-transformed (i.e. converted from 0, 0.01, 0.04,
0.016, 0.63, 2.5, and 10 mM to 0, 1, 2, 3, 4, 5, 6). The rank-transformed dose was
included in the fixed effects model as both a linear term, to test for an increase in
incorporation at lower doses of GlcN, and a quadratic term, to test for a decrease in
incorporation at higher doses of GIcN (Fig. E2).
The linear rank-transformed
dose term was positive and statistically significant,
demonstrating that at low concentrations, GIcN stimulated sulfate incorporation. This
effect was estimated to be relatively modest, however (<20% increase over controls),
and was not observed in all animals. The quadratic rank-transformed dose term was
negative and also statistically significant, confirming the bimodal response of the tissue.
This inhibition of sulfate incorporation by 10 mM GIcN was observed consistently in all
animals and was quite pronounced (greater than 30% mean inhibition for each animal).
We therefore confirm a modest stimulation of proteoglycan synthesis by 10-50 pM GIcN
in newborn bovine cartilage tissue explants that varied widely with animal. However,
higher concentrations of GIcN produced strong inhibition of sulfate and proline
synthesis.
138
-
1.6
.2
1.4
-0
o
o.
8a
1.2
1.0
0.8
0
9
8
U.
'I)
-8
0.4
Z0
0.2
0.00
0.01
0.04
0.16
0.63
2.50
10.00
Glucosamine Concentration (mM)
Figure E2. Summary of effect of glucosamine on incorporation of radiolabeled
sulfate into cartilage tissue explants. Means of sulfate incorporation values,
normalized to incorporation of control cartilage (no GIcN), are shown
separately for each animal (circular markers). Effect of GIcN was estimated by
a linear mixed-effects model (solid line). The first-order rank-transformed dose
term was positive and statistically significant (0.16 ± 0.04, p < 0.001). The
quadratic rank-transformed dose term was negative and also statistically
significant (-0.038 ± 0.006, p < 0.001).
The inhibitory effect of GIcN on cartilage synthetic activity is consistent with cellular
depletion of ATP, although at 10 mM, the effects of GIcN on the hexosamine pathway
are also likely to be important. This result may be important for interpretation of the
results of studies using at these higher concentrations. Since the conclusion of these
studies, de Mattei et al. 1 83 have reported similar results and clearly emphasized the
need to consider the strong inhibitory effect of GIcN on biosynthesis at concentrations
over 1 mM.
It seems likely that the GIcN-mediated inhibition of aggrecanase activity, while useful for
in vitro studies of aggrecanase mechanisms, is not relevant to the clinical effect of GIcN.
The data of Fenton et al.,1 8 4 showing that GluN reduces LPS-induced release of NO and
PG from cartilage at 12 mM GluN, accompanied by deep inhibition of sulfate
incorporation into the tissue, are again consistent with our results and the hypothesis of
intracellular ATP-depletion and resultant loss of enzymatic activity. Although these
139
researchers do not comment on the inhibition of sulfate incorporation in their study, it
would seem likely that the results are not relevant to the clinical mechanism of GlcN
effect. One can predict that, should glucosamine be verified as a clinically-effective
treatment for knee OA, the precise mechanism of its action will continue to be the object
of further investigations.
140
Appendix F
Miscellaneous Experimental Results
40-
N=4
T
"I.
30 -
0
20 -
10 -
0
0
0.2
____F__
___T_
---- F__
1
[sol IL-1 receptor] (ug/m1)
5
--- 7
No IL-1
control
Figure F1 IL-1sr Dose-response. Newborn bovine cartilage was incubated in 10
ng/ml hrlL-1a plus a range of hrlL-1sr. Control cartilage was incubated without
IL-1 or IL-1 sr. GAG release to the medium was measured after 3 days.
141
50 -
I
40 IM
0
-j
N = 4
(N = 2 for 0 ng/ml)
30 -
20 -
10 -
--
0
0
200
100
50
[IL-1 receptor antagonist] (ng/ml)
Figure F2 IL-1 RA Dose-response. Newborn bovine cartilage was pre-incubated
for 24 hours in varying doses of IL-1RA and then incubated in 1 ng/ml IL-1a.
GAG release to the medium was measured after 3 days.
60 0)
0.
I
N = 4 -6
50 40 -
_U)
*)
30 -
0
20 10 -
0
-7--T
0
25
50
100
[IL-1 receptor antagonist] (ng/ml)
--
I
No IL-1
control
Figure F3 IL-1 RA Dose-response 2. Newborn bovine cartilage was incubated in
1 ng/ml IL-1a plus a range of IL-1RA (there was no pre-incubation in IL-IRA).
Control cartilage was incubated in the absence of IL-1 and IL-1 RA. GAG release
to the medium was measured after 3 days.
142
40 Plus ManN
N
Ad
0)
Mean ± SEM
N =5
30 _
20 -
0
0J
10 -
U
Control Co-culture
Injury
Injury plus
Co-culture
IL-1
Figure F4
Effect of mannosamine (ManN) on injury and joint capsule
coincubation. Newborn bovine cartilage was incubated with or without 1.5 mM
ManN. Injury was 50% compression at 1 mm/s, producing mean ± SD peak
stresses of 21± 4 MPa. Co-culture was coincubation of cartilage with a joint
capsule disk in 0.5 ml medium. As a positive control for the effect of ManN, a
group of cartilage was incubated in 1 ng/ml rhIL-1a. GAG release to the medium
was measured after 3 days. ManN appeared to have had no effect.
30 *
Plus ManN
Mean ± SEM
I
N =5
20 -
-- -
U)
0
T==
10 -
U
-
-
Control
-
-
.
Co-culture
Injury
Injury plus
Co-culture
IL-1
Figure F5
Effect of ManN on injury and coincubation. See Fig. F4 for
description. ManN here inhibited the GAG loss due to IL-1, and a trend was seen
for decreased GAG loss in other conditions, suggesting the involvement of an
aggrecanase. The variable effect of ManN has been noted previously.1 8 2
143
20 N=4
T
U)
10
0
0 0
10
1
100
SB703704 Concentration (pM)
No IL-1
control
Figure F6 SB703704 dose-response. Cartilage was incubated in 10 ng/ml IL-1
plus varying doses of the aggrecanase inhibitor SB703704. Control cartilage was
incubated without IL-1 or SB703704. GAG release to the medium was measured
after 3 days.
*
30 -
10 pM SB703704
Mean ± SEM
N=4
20 0
-J
10
U
-
--
-Inju ry
Control Co-culture
Injury plus
Co-culture
IL-1
Figure F7 Effect of aggrecanase inhibition with SB703704 on injury and joint
capsule coincubation. Cartilage disks were incubated with or without 10 pM
ManN. Injury was 50% compression at 1 mm/s. Co-culture was coincubation of
cartilage with a joint capsule disk in 0.25 ml medium. As a positive control for
inhibition, a group of cartilage was incubated in 1 ng/ml rhIL-1a. GAG release to
the medium was measured after 3 days.
144
60
N=4
50
40
..
IM
30
0
20
20
0
o 0
Controls
0
1
10
100
LPS Concentration for Stimulation of
Joint Capsule (pg/ml)
Figure F8 Lipopolysaccharide (LPS) dose-response in joint capsule. Joint
capsule was incubated in varying doses of LPS (from E. coli serotype 055:B5,
Sigma) for 3 days. Capsule was removed and cartilage was then incubated in the
conditioned medium. Controls (not location-matched) were incubated in 3-dayold standard medium. GAG release to the medium was measured after 3 days.
145
50
N =5
S40
30
U)
o
20
D
10
20
0
0
1
10
100
Concentration of Lipopolysaccharide (pg/ml)
Figure F9 Lipopolysaccharide (LPS) dose-response in cartilage. Newborn
bovine cartilage was incubated in varying doses of LPS. GAG release to the
medium was measured after 3 days.
146
1.2
0
1.0
IL-1 sol.
receptor
0
0.
Mean ± SEM
N =6
0.
U
-9
0.6
0
0.0
Control
LPS
IL-1
Figure F10 Effect of recombinant human IL-1 soluble receptor (rhlL-1sr) on
LPS-induced inhibition of bovine cartilage biosynthetic activity. Since LPS
stimulates release of IL-1 from several cell types and has suggested to stimulate
release of IL-1 from cartilage,185 we tested whether rhIL-1sr blocks LPS-induced
effects, in an attempt to verify whether the rhlL-1sr blocks bovine IL-1. After
equilibration in standard medium with 10% FBS, cartilage was transferred to
serum-free medium with or without 10 pg/ml LPS and with or without 5 pg/ml IL1sr. As a positive control for blockade with IL-1sr, additional groups of cartilage
disks were incubated with 1 ng/ml rhlL-1 a. After three days of treatment, cartilage
was radiolabeled with 35S-sulfate for 12 hours. LPS induced a strong inhibition of
cartilage biosynthesis which was not reversed by addition of IL-1sr. It remains
unclear whether the lack of effect of IL-1sr is because of lack of cross-species
reactivity or because LPS-mediated effects are not mediated by IL-1 here.
147
30 -
T
25 -
E]Unloaded
20 U)
.*
-i-
T
15
UInjury (40%)
-
N=4
U)
10 -
5
0
--- |
Control
1 ng/ml IL-1
Figure F11 Effect of IL-1 and Injury. Newborn bovine cartilage was obtained and
cultured as described in Chapter 3. Injury consisted of compression to 40% strain
at 1 mm/s. Cartilage disks were then incubated in 1 % ITS medium with or without
1 ng/ml rhIL-1a. GAG content of conditioned medium was assayed 3 days after
intervention. There was no significant synergistic effect between IL-1 and injury
to 40% strain (p = 0.75 for interaction by two-way ANOVA).
148
60
-]no IL-1
ng/mI IL-1
fl
J~dN
=4
40
0
0
0
0
40
35
45
50
Strain (%)
Figure F12 Effect of strain on interaction of IL-1 and injury. Injury consisted of
compression to 35-50% strain at 1 mm/s. Cartilage disks were then incubated in
1% ITS medium with or without 1 ng/ml rhIL-1a. GAG content of conditioned
medium was assayed 3 days after intervention.
149
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Davis MA, Ettinger WH, Neuhaus JM, Cho SA, Hauck WW. The association of knee
injury and obesity with unilateral and bilateral osteoarthritis of the knee. Am J Epidemiol
1989; 130:278-88.
Gelber AC, Hochberg MC, Mead LA, Wang NY, Wigley FM, Klag MJ. Joint injury in
young adults and risk for subsequent knee and hip osteoarthritis. Ann Intern Med 2000;
133:321-8.
Roos H, Adalberth T, Dahlberg L, Lohmander LS. Osteoarthritis of the knee after injury
to the anterior cruciate ligament or meniscus: the influence of time and age.
Osteoarthritis Cartilage 1995; 3:261-7.
Daniel DM, Stone ML, Dobson BE, Fithian DC, Rossman DJ, Kaufman KR. Fate of the
ACL-injured patient. A prospective outcome study. Am J Sports Med 1994; 22:632-44.
Lohmander LS, Hoerrner LA, Dahlberg L, Roos H, Bjornsson S, Lark MW. Stromelysin,
tissue inhibitor of metalloproteinases and proteoglycan fragments in human knee joint
fluid after injury. J Rheumatol 1993; 20:1362-8.
Lohmander LS, Neame PJ, Sandy JD. The structure of aggrecan fragments in human
synovial fluid. Evidence that aggrecanase mediates cartilage degradation in
inflammatory joint disease, joint injury, and osteoarthritis. Arthritis Rheum 1993;
36:1214-22.
Lohmander LS, Atley LM, Peitka TA, Eyre DR. The release of cross-linked
proteoglycans from type 11 collagen into joint fluid and serum is increased in
osteoarthritis and after joint injury. Transactions of the Orthopaedic Research Society
2000; 25:236.
Price JS, Till SH, Bickerstaff DR, Bayliss MT, Hollander AP. Degradation of cartilage
type I collagen precedes the onset of osteoarthritis following anterior cruciate ligament
rupture. Arthritis Rheum 1999; 42:2390-8.
Radin EL, Parker HG, Pugh JW, Steinberg RS, Paul IL, Rose RM. Response of joints to
impact loading. 3. Relationship between trabecular microfractures and cartilage
degeneration. J Biomech 1973; 6:51-7.
Radin EL, Ehrlich MG, Chernack R, Abernethy P, Paul IL, Rose RM. Effect of repetitive
impulsive loading on the knee joints of rabbits. Clin Orthop 1978:288-93.
Radin EL, Martin RB, Burr DB, Caterson B, Boyd RD, Goodwin C. Effects of mechanical
loading on the tissues of the rabbit knee. J Orthop Res 1984; 2:221-34.
Newberry WN, Garcia JJ, Mackenzie CD, Decamp CE, Haut RC. Analysis of acute
mechanical insult in an animal model of post-traumatic osteoarthrosis. J Biomech Eng
1998; 120:704-9.
Chen CT, Burton-Wurster N, Lust G, Bank RA, Tekoppele JM. Compositional and
metabolic changes in damaged cartilage are peak-stress, stress-rate, and loadingduration dependent. J Orthop Res 1999; 17:870-9.
Farquhar T, Xia Y, Mann K, et al. Swelling and fibronectin accumulation in articular
cartilage explants after cyclical impact. Journal of Orthopaedic Research 1996; 14:417423.
Steinmeyer J, Knue S. The proteoglycan metabolism of mature bovine articular cartilage
explants superimposed to continuously applied cyclic mechanical loading. Biochem
Biophys Res Commun 1997; 240:216-21.
Repo RU, Finlay JB. Survival of articular cartilage after controlled impact. Journal of
Bone and Joint Surgery 1977; 59A:1068-1076.
Jeffrey JE, Gregory DW, Aspden RM. Matrix damage and chondrocyte viability following
a single impact load on articular cartilage. Arch Biochem Biophys 1995; 322:87-96.
150
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
Torzilli PA, Grigiene R, Borrelli J, Jr., Helfet DL. Effect of impact load on articular
cartilage: cell metabolism and viability, and matrix water content. J Biomech Eng 1999;
121:433-41.
Kurz B, Jin M, Patwari P, Cheng DM, Lark MW, Grodzinsky AJ. Biosynthetic response
and mechanical properties of articular cartilage after injurious compression. J Orthop
Res 2001; 19:1140-6.
Jeffrey JE, Thomson LA, Aspden RM. Matrix loss and synthesis following a single
impact load on articular cartilage in vitro. Biochim Biophys Acta 1997; 1334:223-32.
Palmoski MJ, Brandt KD. Effects of static and cyclic compressive loading on articular
cartilage plugs in vitro. Arthritis Rheum 1984; 27:675-81.
Parkkinen JJ, Lammi MJ, Helminen HJ, Tammi M. Local stimulation of proteoglycan
synthesis in articular cartilage explants by dynamic compression in vitro. J Orthop Res
1992; 10:610-20.
Sah RL, Kim YJ, Doong JY, Grodzinsky AJ, Plaas AH, Sandy JD. Biosynthetic response
of cartilage explants to dynamic compression. J Orthop Res 1989; 7:619-36.
Blanco FJ, Guitian R, Vazquez-Martul E, de Toro FJ, Galdo F. Osteoarthritis
chondrocytes die by apoptosis. A possible pathway for osteoarthritis pathology. Arthritis
Rheum 1998; 41:284-9.
Hashimoto S, Ochs RL, Komiya S, Lotz M. Linkage of chondrocyte apoptosis and
cartilage degradation in human osteoarthritis. Arthritis Rheum 1998; 41:1632-8.
Kirsch T, Swoboda B, Nah H. Activation of annexin I and V expression, terminal
differentiation, mineralization and apoptosis in human osteoarthritic cartilage.
Osteoarthritis Cartilage 2000; 8:294-302.
Hashimoto S, Ochs RL, Rosen F, et al. Chondrocyte-derived apoptotic bodies and
calcification of articular cartilage. Proc Natl Acad Sci U S A 1998; 95:3094-9.
Loening AM, James IE, Levenston ME, et al. Injurious mechanical compression of
bovine articular cartilage induces chondrocyte apoptosis. Arch Biochem Biophys 2000;
381:205-12.
D'Lima DD, Hashimoto S, Chen PC, Colwell CW, Jr., Lotz MK. Impact of mechanical
trauma on matrix and cells. Clin Orthop 2001:S90-9.
Frank EH, Jin M, Loening AM, Levenston ME, Grodzinsky AJ. A versatile shear and
compression apparatus for mechanical stimulation of tissue culture explants. J Biomech
2000; 33:1523-7.
Ragan PM, Badger AM, Cook M, et al. Down-regulation of chondrocyte aggrecan and
type-Il collagen gene expression correlates with increases in static compression
magnitude and duration. J Orthop Res 1999; 17:836-42.
Martin JA, Heiner AD, Brown KD, Schroeder A, Brand RA, Buckwalter JA. Mechanical
stress induces proMMP-3 protein expression in human articular cartilage, Transactions
of the Orthopaedic Research Society, 1999. Vol. 24.
Chubinskaya S, Kuettner KE, Cole AA. Expression of matrix metalloproteinases in
normal and damaged articular cartilage from human knee and ankle joints. Lab Invest
1999; 79:1669-77.
Lark MW, Bayne EK, Flanagan J, et al. Aggrecan degradation in human cartilage.
Evidence for both matrix metalloproteinase and aggrecanase activity in normal,
osteoarthritic, and rheumatoid joints. J Clin Invest 1997; 100:93-106.
Quinn TM, Grodzinsky AJ, Hunziker EB, Sandy JD. Effects of injurious compression on
matrix turnover around individual cells in calf articular cartilage explants. J Orthop Res
1998; 16:490-9.
Bonassar LJ, Sandy JD, Lark MW, Plaas AH, Frank EH, Grodzinsky AJ. Inhibition of
cartilage degradation and changes in physical properties induced by IL-1 beta and
151
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
retinoic acid using matrix metalloproteinase inhibitors. Arch Biochem Biophys 1997;
344:404-12.
Maroudas A. Balance between swelling pressure and collagen tension in normal and
degenerate cartilage. Nature 1976; 260:808-809.
Jubb RW, Fell HB. The effect of synovial tissue on the synthesis of proteoglycan by the
articular cartilage of young pigs. Arthritis Rheum 1980; 23:545-55.
Vankemmelbeke MN, Ilic MZ, Handley CJ, Knight CG, Buttle DJ. Coincubation of bovine
synovial or capsular tissue with cartilage generates a soluble "Aggrecanase" activity.
Biochem Biophys Res Commun 1999; 255:686-91.
Schumacher HR, Jr. Synovial membrane and fluid morphologic alterations in early
rheumatoid arthritis: microvascular injury and virus-like particles. Ann N Y Acad Sci
1975; 256:39-64.
Chu CQ, Field M, Feldmann M, Maini RN. Localization of tumor necrosis factor alpha in
synovial tissues and at the cartilage-pannus junction in patients with rheumatoid arthritis.
Arthritis Rheum 1991; 34:1125-32.
McCachren SS. Expression of metalloproteinases and metalloproteinase inhibitor in
human arthritic synovium. Arthritis Rheum 1991; 34:1085-93.
Ilic MZ, Vankemmelbeke MN, Holen 1,Buttle DJ, Clem Robinson H, Handley CJ. Bovine
joint capsule and fibroblasts derived from joint capsule express aggrecanase activity.
Matrix Biol 2000; 19:257-65.
Mankin HJ. Localization of tritiated thymidine in articular cartilage of rabbits. III: Mature
articular cartilage. J Bone Joint Surg Am 1963; 45:529-540.
Aigner T, Hemmel M, Neureiter D, et al. Apoptotic cell death is not a widespread
phenomenon in normal aging and osteoarthritis human articular knee cartilage: a study
of proliferation, programmed cell death (apoptosis), and viability of chondrocytes in
normal and osteoarthritic human knee cartilage. Arthritis Rheum 2001; 44:1304-12.
Grogan SP, Aklin B, Frenz M, Brunner T, Schaffner T, Mainil-Varlet P. In vitro model for
the study of necrosis and apoptosis in native cartilage. J Pathol 2002; 198:5-13.
Doi T, Nishida K, Matsuo M, Yoshida A, Murakami T, Inoue H. Evidence of oncotic cell
death and DNA fragmentation in human hypertrophic chondrocytes in chondroosteophyte. Osteoarthritis Cartilage 2002; 10:270-6.
Chen CT, Burton-Wurster N, Borden C, Hueffer K, Bloom SE, Lust G. Chondrocyte
necrosis and apoptosis in impact damaged articular cartilage. J Orthop Res 2001;
19:703-11.
Aigner T, Kim HA. Apoptosis and cellular vitality: issues in osteoarthritic cartilage
degeneration. Arthritis Rheum 2002; 46:1986-96.
Colicos MA, Dash PK. Apoptotic morphology of dentate gyrus granule cells following
experimental cortical impact injury in rats: possible role in spatial memory deficits. Brain
Res 1996; 739:120-31.
Sperandio S, de Belle I, Bredesen DE. An alternative, nonapoptotic form of programmed
cell death. Proc Natl Acad Sci U S A 2000; 97:14376-81.
D'Lima DD, Hashimoto S, Chen PC, Colwell CW, Jr., Lotz MK. Human chondrocyte
apoptosis in response to mechanical injury. Osteoarthritis Cartilage 2001; 9:712-9.
Patwari P, Cook MN, DiMicco MA, et al. Proteoglycan degradation after injurious
compression of bovine and human articular cartilage in vitro: interaction with exogenous
cytokines. Arthritis Rheum 2003; 48:1292-1301.
Patwari P, Kurz B, Berger E, et al. Comparative evaluation of apoptotic and necrotic cell
death after injurious compression of bovine articular cartilage using electron microscopy
and tunel staining. Transactions of the Orthopaedic Research Society 2002; 27:108.
152
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
Clements KM, Bee ZC, Crossingham GV, Adams MA, Sharif M. How severe must
repetitive loading be to kill chondrocytes in articular cartilage? Osteoarthritis Cartilage
2001; 9:499-507.
Tew SR, Kwan AP, Hann A, Thomson BM, Archer CW. The reactions of articular
cartilage to experimental wounding: role of apoptosis. Arthritis Rheum 2000; 43:215-25.
Sloop GD, Roa JC, Delgado AG, Balart JT, Hines MO, 3rd, Hill JM. Histologic sectioning
produces TUNEL reactivity. A potential cause of false-positive staining. Arch Pathol Lab
Med 1999; 123:529-32.
Lemke A, Patwari P, Grodzinsky AJ, Domm C, SchOnke M, Kurz B. Influence of age on
mechanical induction of apoptosis in articular cartilage, Transactions of the Orthopaedic
Research Society, 2003. Vol. 28.
Agar A, Yip SS, Hill MA, Coroneo MT. Pressure related apoptosis in neuronal cell lines.
J Neurosci Res 2000; 60:495-503.
Sho E, Sho M, Singh TM, Xu C, Zarins CK, Masuda H. Blood flow decrease induces
apoptosis of endothelial cells in previously dilated arteries resulting from chronic high
blood flow. Arterioscler Thromb Vasc Biol 2001; 21:1139-45.
Beattie MS, Hermann GE, Rogers RC, Bresnahan JC. Cell death in models of spinal
cord injury. Prog Brain Res 2002; 137:37-47.
Hong SY, Lee H, You WK, Chung KH, Kim DS, Song K. The snake venom disintegrin
salmosin induces apoptosis by disassembly of focal adhesions in bovine capillary
endothelial cells. Biochem Biophys Res Commun 2003; 302:502-8.
Lucchinetti E, Adams CS, Horton WE, Jr., Torzilli PA. Cartilage viability after repetitive
loading: a preliminary report. Osteoarthritis Cartilage 2002; 10:71-81.
Felson DT, Lawrence RC, Dieppe PA, et al. Osteoarthritis: new insights. Part 1: the
disease and its risk factors. Ann Intern Med 2000; 133:635-46.
Sah RL, Doong JY, Grodzinsky AJ, Plaas AH, Sandy JD. Effects of compression on the
loss of newly synthesized proteoglycans and proteins from cartilage explants. Arch
Biochem Biophys 1991; 286:20-9.
Ewers BJ, Dvoracek-Driksna D, Orth MW, Haut RC. The extent of matrix damage and
chondrocyte death in mechanically traumatized articular cartilage explants depends on
rate of loading. J Orthop Res 2001; 19:779-84.
Quinn TM, Maung AA, Grodzinsky AJ, Hunziker EB, Sandy JD. Physical and biological
regulation of proteoglycan turnover around chondrocytes in cartilage explants.
Implications for tissue degradation and repair. Ann N Y Acad Sci 1999; 878:420-41.
Thibault M, Poole AR, Buschmann MD. Cyclic compression of cartilage/bone explants in
vitro leads to physical weakening, mechanical breakdown of collagen and release of
matrix fragments. J Orthop Res 2002; 20:1265-73.
Pelletier JP, Martel-Pelletier J, Abramson SB. Osteoarthritis, an inflammatory disease:
potential implication for the selection of new therapeutic targets. Arthritis Rheum 2001;
44:1237-47.
Attur MG, Dave M, Akamatsu M, Katoh M, Amin AR. Osteoarthritis or osteoarthrosis: the
definition of inflammation becomes a semantic issue in the genomic era of molecular
medicine. Osteoarthritis Cartilage 2002; 10:1-4.
Patwari P, Fay J, Cook MN, et al. In vitro models for investigation of the effects of acute
mechanical injury on cartilage. Clin Orthop 2001:S61-71.
Muehleman C, Bareither D, Huch K, Cole AA, Kuettner KE. Prevalence of degenerative
morphological changes in the joints of the lower extremity. Osteoarthritis Cartilage 1997;
5:23-37.
Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation and discrimination of
sulphated glycosaminoglycans by use of dimethylmethylene blue. Biochimica et
Biophysica Acta 1986; 883:173-177.
153
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
Flannery CR, Little CB, Caterson B, Hughes CE. Effects of culture conditions and
exposure to catabolic stimulators (IL-1 and retinoic acid) on the expression of matrix
metalloproteinases (MMPs) and disintegrin metalloproteinases (ADAMs) by articular
cartilage chondrocytes. Matrix Biol 1999; 18:225-37.
Arner EC, Hughes CE, Decicco CP, Caterson B, Tortorella MD. Cytokine-induced
cartilage proteoglycan degradation is mediated by aggrecanase. Osteoarthritis Cartilage
1998; 6:214-28.
Borden P, Solymar D, Sucharczuk A, Lindman B, Cannon P, Heller RA. Cytokine control
of interstitial collagenase and collagenase-3 gene expression in human chondrocytes. J
Biol Chem 1996; 271:23577-81.
Reboul P, Pelletier JP, Tardif G, Cloutier JM, Martel-Pelletier J. The new collagenase,
collagenase-3, is expressed and synthesized by human chondrocytes but not by
synoviocytes. A role in osteoarthritis. J Clin Invest 1996; 97:2011-9.
Vincent T, Hermansson M, Bolton M, Wait R, Saklatvala J. Basic FGF mediates an
immediate response of articular cartilage to mechanical injury. Proc Natl Acad Sci U S A
2002; 99:8259-64.
Vaatainen U, Lohmander LS, Thonar E, et al. Markers of cartilage and synovial
metabolism in joint fluid and serum of patients with chondromalacia of the patella.
Osteoarthritis Cartilage 1998; 6:115-24.
Lettesjo H, Nordstrom E, Strom H, et al. Synovial fluid cytokines in patients with
rheumatoid arthritis or other arthritic lesions. Scand J Immunol 1998; 48:286-92.
Radin EL, Burr DB, Caterson B, Fyhrie D, Brown TD, Boyd RD. Mechanical
determinants of osteoarthrosis. Semin Arthritis Rheum 1991; 21:12-21.
Treppo S, Koepp H, Quan EC, Cole AA, Kuettner KE, Grodzinsky AJ. Comparison of
biomechanical and biochemical properties of cartilage from human knee and ankle pairs.
J Orthop Res 2000; 18:739-48.
Eger W, Schumacher BL, Mollenhauer J, Kuettner KE, Cole AA. Human knee and ankle
cartilage explants: catabolic differences. J Orthop Res 2002; 20:526-34.
Aigner T, Kurz B, Fukui N, Sandell L. Roles of chondrocytes in the pathogenesis of
osteoarthritis. Curr Opin Rheumatol 2002; 14:578-84.
Goldring MB. The role of the chondrocyte in osteoarthritis. Arthritis Rheum 2000;
43:1916-26.
Wilder FV, Hall BJ, Barrett JP, Jr., Lemrow NB. History of acute knee injury and
osteoarthritis of the knee: a prospective epidemiological assessment. The Clearwater
Osteoarthritis Study. Osteoarthritis Cartilage 2002; 10:611-6.
Buckwalter JA. Articular cartilage injuries. Clin Orthop 2002:21-37.
Irie K, Uchiyama E, Iwaso H. Intraarticular inflammatory cytokines in acute anterior
cruciate ligament injured knee. Knee 2003; 10:93-6.
Roy S. Ultrastructure of synovial membrane in osteo-arthritis. Ann Rheum Dis 1967;
26:517-27.
Lindblad S, Hedfors E. Arthroscopic and immunohistologic characterization of knee joint
synovitis in osteoarthritis. Arthritis Rheum 1987; 30:1081-8.
Smith MD, Triantafillou S, Parker A, Youssef PP, Coleman M. Synovial membrane
inflammation and cytokine production in patients with early osteoarthritis. J Rheumatol
1997; 24:365-71.
Walker ER, Boyd RD, Wu DD, Lukoschek M, Burr DB, Radin EL. Morphologic and
morphometric changes in synovial membrane associated with mechanically induced
osteoarthrosis. Arthritis Rheum 1991; 34:515-24.
Marijnissen AC, van Roermund PM, Verzijl N, Tekoppele JM, Bijlsma JW, Lafeber FP.
Steady progression of osteoarthritic features in the canine groove model. Osteoarthritis
Cartilage 2002; 10:282-9.
154
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
Myers SL, Brandt KD, O'Connor BL, Visco DM, Albrecht ME. Synovitis and osteoarthritic
changes in canine articular cartilage after anterior cruciate ligament transection. Effect of
surgical hemostasis. Arthritis Rheum 1990; 33:1406-15.
Ray A, Kuroki K, Cook JL, et al. Induction of matrix metalloproteinase 1 gene expression
is regulated by inflammation-responsive transcription factor SAF-1 in osteoarthritis.
Arthritis Rheum 2003; 48:134-45.
Hedbom E, Hauselmann HJ. Molecular aspects of pathogenesis in osteoarthritis: the
role of inflammation. Cell Mol Life Sci 2002; 59:45-53.
Yuan GH, Masuko-Hongo K, Kato T, Nishioka K. Immunologic intervention in the
pathogenesis of osteoarthritis. Arthritis Rheum 2003; 48:602-11.
Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation and discrimination of
sulphated glycosaminoglycans by use of dimethylmethylene blue. Biochim Biophys Acta
1986; 883:173-7.
Diggle P, Liang K-Y, Zeger SL. Analysis of longitudinal data. Oxford statistical science
series; 13. New York: Clarendon Press, 1994:253.
Fell HB, Jubb RW. The effect of synovial tissue on the breakdown of articular cartilage in
organ culture. Arthritis Rheum 1977; 20:1359-71.
Dingle JT, Saklatvala J, Hembry R, Tyler J, Fell HB, Jubb R. A cartilage catabolic factor
from synovium. Biochem J 1979; 184:177-80.
Yodlowski ML, Hubbard JR, Kispert J, Keller K, Sledge CB, Steinberg JJ. Antibody to
interleukin 1 inhibits the cartilage degradative and thymocyte proliferative actions of
rheumatoid synovial culture medium. J Rheumatol 1990; 17:1600-7.
Neidhart M, Gay RE, Gay S. Anti-interleukin-1 and anti-CD44 interventions producing
significant inhibition of cartilage destruction in an in vitro model of cartilage invasion by
rheumatoid arthritis synovial fibroblasts. Arthritis Rheum 2000; 43:1719-28.
Decker RS, Dingle JT. Cardiac catabolic factors: the degradation of heart valve
intercellular matrix. Science 1982; 215:987-9.
Rowan AD, Koshy PJ, Shingleton WD, et al. Synergistic effects of glycoprotein 130
binding cytokines in combination with interleukin-1 on cartilage collagen breakdown.
Arthritis Rheum 2001; 44:1620-32.
Flannery CR, Little CB, Hughes CE, Curtis CL, Caterson B, Jones SA. IL-6 and its
soluble receptor augment aggrecanase-mediated proteoglycan catabolism in articular
cartilage. Matrix Biol 2000; 19:549-53.
Legendre F, Dudhia J, Pujol JP, Bogdanowicz P. JAK/STAT but not ERK1/ERK2
pathway mediates interleukin (IL)-6/soluble IL-6R down-regulation of type II collagen,
aggrecan core, and link protein transcription in articular chondrocytes. Association with
a down-regulation of SOX9 expression. J Biol Chem 2003; 278:2903-12.
Scaffidi P, Misteli T, Bianchi ME. Release of chromatin protein HMGB1 by necrotic cells
triggers inflammation. Nature 2002; 418:191-5.
Taniguchi N, Kawahara K, Yone K, et al. High mobility group box chromosomal protein 1
plays a role in the pathogenesis of rheumatoid arthritis as a novel cytokine. Arthritis
Rheum 2003; 48:971-81.
Sandy JD, Neame PJ, Boynton RE, Flannery CR. Catabolism of aggrecan in cartilage
explants. Identification of a major cleavage site within the interglobular domain. Journal
of Biological Chemistry 1991; 266:8683-5.
Ilic MZ, Handley CJ, Robinson HC, Mok MT. Mechanism of catabolism of aggrecan by
articular cartilage. Arch Biochem Biophys 1992; 294:115-22.
Loulakis P, Shrikhande A, Davis G, Maniglia CA. N-terminal sequence of proteoglycan
fragments isolated from medium of interleukin-1-treated articular-cartilage cultures.
Putative site(s) of enzymic cleavage. Biochem J 1992; 284 ( Pt 2):589-93.
155
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
Tortorella MD, Burn TC, Pratta MA, et al. Purification and cloning of aggrecanase-1: a
member of the ADAMTS family of proteins. Science 1999; 284:1664-6.
Sandy JD, Flannery CR, Neame PJ, Lohmander LS. The Structure of Aggrecan
Fragments in Human Synovial Fluid: Evidence for the Involvement in Osteoarthritis of a
Novel Proteinase Which Cleaves the Glu 373-Ala 374 Bond of the Interglobular Domain.
Journal of Clinical Investigation 1992; 89:1512-1516.
Gao G, Westling J, Thompson VP, Howell TD, Gottschall PE, Sandy JD. Activation of
the proteolytic activity of ADAMTS4 (aggrecanase-1) by C-terminal truncation. J Biol
Chem 2002; 277:11034-41.
Flannery CR, Zeng W, Corcoran C, et al. Autocatalytic cleavage of ADAMTS-4
(aggrecanase-1) reveals multiple glycosaminoglycan-binding sites. J Biol Chem 2002;
277:42775-80.
Tortorella MD, Malfait AM, Deccico C, Arner E. The role of ADAM-TS4 (aggrecanase-1)
and ADAM-TS5 (aggrecanase-2) in a model of cartilage degradation. Osteoarthritis
Cartilage 2001; 9:539-52.
Sandy JD, Thompson V, Verscharen C, Gamett D. Chondrocyte-mediated catabolism of
aggrecan: evidence for a glycosylphosphatidylinositol-linked protein in the aggrecanase
response to interleukin-1 or retinoic acid. Arch Biochem Biophys 1999; 367:258-64.
Sandy JD, Gamett D, Thompson V, Verscharen C. Chondrocyte-mediated catabolism of
aggrecan: aggrecanase-dependent cleavage induced by interleukin-1 or retinoic acid
can be inhibited by glucosamine. Biochem J 1998; 335 ( Pt 1):59-66.
Sandy JD, Verscharen C. Analysis of aggrecan in human knee cartilage and synovial
fluid indicates that aggrecanase (ADAMTS) activity is responsible for the catabolic
turnover and loss of whole aggrecan whereas other protease activity is required for Cterminal processing in vivo. Biochem J 2001; 358:615-26.
Tortorella MD, Pratta M, Liu RQ, et al. Sites of aggrecan cleavage by recombinant
human aggrecanase-1 (ADAMTS-4). J Biol Chem 2000; 275:18566-73.
Patwari P, Kurz B, Sandy JD, Grodzinsky AJ. Mannosamine inhibits aggrecanasemediated changes in the physical properties and biochemical composition of articular
cartilage. Arch Biochem Biophys 2000; 374:79-85.
Hering TM, Kollar J, Huynh TD. Complete coding sequence of bovine aggrecan:
comparative structural analysis. Arch Biochem Biophys 1997; 345:259-70.
Tortorella MD, Liu RQ, Burn T, Newton RC, Arner E. Characterization of human
aggrecanase 2 (ADAM-TS5): substrate specificity studies and comparison with
aggrecanase 1 (ADAM-TS4). Matrix Biol 2002; 21:499-511.
Sandy JD, Westling J, Kenagy RD, et al. Versican V1 proteolysis in human aorta in vivo
occurs at the Glu441-Ala442 bond, a site that is cleaved by recombinant ADAMTS-1 and
ADAMTS-4. J Biol Chem 2001; 276:13372-8.
Kuno K, Okada Y, Kawashima H, et al. ADAMTS-1 cleaves a cartilage proteoglycan,
aggrecan. FEBS Lett 2000; 478:241-5.
Pratta MA, Scherle PA, Yang G, Liu RQ, Newton RC. Induction of aggrecanase 1
(ADAM-TS4) by interleukin-1 occurs through activation of constitutively produced
protein. Arthritis Rheum 2003; 48:119-33.
Sharma L, Song J, Felson DT, Cahue S, Shamiyeh E, Dunlop DD. The role of knee
alignment in disease progression and functional decline in knee osteoarthritis. JAMA
2001; 286:188-95.
Brandt KD. Transection of the anterior cruciate ligament in the dog: a model of
osteoarthritis. Seminars in Arthritis and Rheumatology 1991; 21:22-32.
Stockwell RA, Meachim G. The Chondrocytes. In: Freeman MAR, ed. Adult Articular
Cartilage. New York, NY: Grune & Stratton, 1979:69-104.
156
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
Williamson AK, Chen AC, Sah RL. Compressive properties and function-composition
relationships of developing bovine articular cartilage. J Orthop Res 2001; 19:1113-21.
Hudelmaier M, Glaser C, Hohe J, et al. Age-related changes in the morphology and
deformational behavior of knee joint cartilage. Arthritis Rheum 2001; 44:2556-61.
Kim HT, Lo MY, Pillarisetty R. Chondrocyte apoptosis following intraarticular fracture in
humans. Osteoarthritis Cartilage 2002; 10:747-9.
Hockenbery DM, Oltvai ZN, Yin XM, Milliman CL, Korsmeyer SJ. BcI-2 functions in an
antioxidant pathway to prevent apoptosis. Cell 1993; 75:241-51.
Kane DJ, Sarafian TA, Anton R, et al. Bcl-2 inhibition of neural death: decreased
generation of reactive oxygen species. Science 1993; 262:1274-7.
Sandstrom PA, Mannie MD, Buttke TM. Inhibition of activation-induced death in T cell
hybridomas by thiol antioxidants: oxidative stress as a mediator of apoptosis. J Leukoc
Biol 1994; 55:221-6.
Cai J, Jones DP. Superoxide in apoptosis. Mitochondrial generation triggered by
cytochrome c loss. J Biol Chem 1998; 273:11401-4.
Ohyama K, Yuan B, Bessho T, Yamakawa T. Progressive apoptosis in chorion laeve
trophoblast cells of human fetal membrane tissues during in vitro incubation is
suppressed by antioxidative reagents. Eur J Biochem 2001; 268:6182-9.
Malassagne B, Ferret PJ, Hammoud R, et al. The superoxide dismutase mimetic
MnTBAP prevents Fas-induced acute liver failure in the mouse. Gastroenterology 2001;
121:1451-9.
Sugawara T, Lewen A, Gasche Y, Yu F, Chan PH. Overexpression of SOD1 protects
vulnerable motor neurons after spinal cord injury by attenuating mitochondrial
cytochrome c release. Faseb J 2002; 16:1997-9.
Del Carlo M, Jr., Loeser RF. Nitric oxide-mediated chondrocyte cell death requires the
generation of additional reactive oxygen species. Arthritis Rheum 2002; 46:394-403.
Kurz B, Jost B, Schunke M. Dietary vitamins and selenium diminish the development of
mechanically induced osteoarthritis and increase the expression of antioxidative
enzymes in the knee joint of STR/1 N mice. Osteoarthritis Cartilage 2002; 10:119-26.
Faulkner KM, Liochev SI, Fridovich 1.Stable Mn(lll) porphyrins mimic superoxide
dismutase in vitro and substitute for it in vivo. J Biol Chem 1994; 269:23471-6.
Gardner PR, Nguyen DD, White CW. Superoxide scavenging by Mn(II/lll) tetrakis (1methyl-4-pyridyl) porphyrin in mammalian cells. Arch Biochem Biophys 1996; 325:20-8.
Klann E. Cell-permeable scavengers of superoxide prevent long-term potentiation in
hippocampal area CA1. J Neurophysiol 1998; 80:452-7.
Kurz B, Schunke M. Articular chondrocytes and synoviocytes in culture: influence of
antioxidants on lipid peroxidation and proliferation. Anat Anz 1997; 179:439-46.
Agresti A. Categorical data analysis. Wiley series in probability and mathematical
statistics. Applied probability and statistics. New York: Wiley, 1990:558.
Earnshaw WC. Nuclear changes in apoptosis. Curr Opin Cell Biol 1995; 7:337-43.
Dini L, Coppola S, Ruzittu MT, Ghibelli L. Multiple pathways for apoptotic nuclear
fragmentation. Exp Cell Res 1996; 223:340-7.
Maly FE, Nakamura M, Gauchat JF, et al. Superoxide-dependent nitroblue tetrazolium
reduction and expression of cytochrome b-245 components by human tonsillar B
lymphocytes and B cell lines. J Immunol 1989; 142:1260-7.
Mankin HJ. Localization of tritiated thymidine in articular cartilage of rabbits. I: Growth in
immature cartilage. J Bone Joint Surg Am 1962; 44:682-688.
Li KW, Williamson AK, Wang AS, Sah RL. Growth responses of cartilage to static and
dynamic compression. Clin Orthop 2001:S34-48.
Kaiki G, Tsuji H, Yonezawa T, et al. Osteoarthrosis induced by intraarticular hydrogenperoxide injection and running load. J Orthop Res 1990; 8:731-740.
157
154.
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
Ferrer-Sueta G, Batinic-Haberle 1,Spasojevic I, Fridovich I, Radi R. Catalytic scavenging
of peroxynitrite by isomeric Mn(Ill) N-methylpyridylporphyrins in the presence of
reductants. Chem Res Toxicol 1999; 12:442-9.
Dwyer BE, Lu SY, Laitinen JT, Nishimura RN. Protective properties of tin- and
manganese-centered porphyrins against hydrogen peroxide-mediated injury in rat
astroglial cells. J Neurochem 1998; 71:2497-504.
Murray M, Vrahas M. Chondrocyte apoptosis after articular fracture in humans.
Transactions of the Orthopaedic Research Society 2002; 27:424.
Colwell CW, Jr., D'Lima DD, Hoenecke HR, et al. In vivo changes after mechanical
injury. Clin Orthop 2001:S116-23.
Nuttall ME, Nadeau DP, Fisher PW, et al. Inhibition of caspase-3-like activity prevents
apoptosis while retaining functionality of human chondrocytes in vitro. J Orthop Res
2000; 18:356-63.
Hooiveld M, Roosendaal G, Wenting M, Van Den Berg M, Bijlsma J, Lafeber F. Shortterm exposure of cartilage to blood results in chondrocyte apoptosis. Am J Pathol 2003;
162:943-51.
Fosang AJ, Last K, Gardiner P, Jackson DC, Brown L. Development of a cleavage-sitespecific monoclonal antibody for detecting metalloproteinase-derived aggrecan
fragments: detection of fragments in human synovial fluids. Biochem J 1995; 310 ( Pt
1):337-43.
Flechtenmacher J, Huch K, Thonar EJ, et al. Recombinant human osteogenic protein 1
is a potent stimulator of the synthesis of cartilage proteoglycans and collagens by human
articular chondrocytes. Arthritis Rheum 1996; 39:1896-904.
Chen P, Vukicevic S, Sampath TK, Luyten FP. Osteogenic protein-1 promotes growth
and maturation of chick sternal chondrocytes in serum-free cultures. J Cell Sci 1995; 108
( Pt 1):105-14.
Koepp HE, Sampath KT, Kuettner KE, Homandberg GA. Osteogenic protein-1 (OP-1)
blocks cartilage damage caused by fibronectin fragments and promotes repair by
enhancing proteoglycan synthesis. Inflamm Res 1999; 48:199-204.
Chubinskaya S, Kumar B, Merrihew C, Heretis K, Rueger DC, Kuettner KE. Age-related
changes in cartilage endogenous osteogenic protein-1 (OP-1). Biochim Biophys Acta
2002; 1588:126-34.
Drovanti A, Bignamini AA, Rovati AL. Therapeutic activity of oral glucosamine sulfate in
osteoarthrosis: a placebo-controlled double-blind investigation. Clin Ther 1980; 3:26072.
Pujalte JM, Llavore EP, Ylescupidez FR. Double-blind clinical evaluation of oral
glucosamine sulphate in the basic treatment of osteoarthrosis. Curr Med Res Opin 1980;
7:110-14.
Vajaradul Y. Double-blind clinical evaluation of intra-articular glucosamine in outpatients
with gonarthrosis. Clin Ther 1981; 3:336-43.
D'Ambrosio E, Casa B, Bompani R, Scali G, Scali M. Glucosamine sulphate: a controlled
clinical investigation in arthrosis. Pharmatherapeutica 1981; 2:504-8.
Reichelt A, Forster KK, Fischer M, Rovati LC, Setnikar I. Efficacy and safety of
intramuscular glucosamine sulfate in osteoarthritis of the knee. A randomised, placebocontrolled, double-blind study. Arzneimittelforschung 1994; 44:75-80.
Qiu GX, Gao SN, Giacovelli G, Rovati L, Setnikar I. Efficacy and safety of glucosamine
sulfate versus ibuprofen in patients with knee osteoarthritis. Arzneimittelforschung 1998;
48:469-74.
Reginster JY, Deroisy R, Rovati LC, et al. Long-term effects of glucosamine sulphate on
osteoarthritis progression: a randomised, placebo-controlled clinical trial. Lancet 2001;
357:251-6.
158
172.
173.
174.
175.
176.
177.
178.
179.
180.
181.
182.
183.
184.
185.
Houpt JB, McMillan R, Wein C, Paget-Dellio SD. Effect of glucosamine hydrochloride in
the treatment of pain of osteoarthritis of the knee. J Rheumatol 1999; 26:2423-30.
Rindone JP, Hiller D, Collacott E, Nordhaugen N, Arriola G. Randomized, controlled trial
of glucosamine for treating osteoarthritis of the knee. West J Med 2000; 172:91-4.
Hughes R, Carr A. A randomized, double-blind, placebo-controlled trial of glucosamine
sulphate as an analgesic in osteoarthritis of the knee. Rheumatology (Oxford) 2002;
41:279-84.
Marcus DM, Grollman AP. Botanical medicines--the need for new regulations. N EngI J
Med 2002; 347:2073-6.
NIH News Release. Glucosamine/Chondroitin Arthritis Intervention Trial (GAIT) Begins
Patient Recruitment. (Accessed April 21, 2003, at
http://nccam.nih.gov/news/19972000/121100/.)
Kim JJ, Conrad HE. Effect of D-glucosamine concentration on the kinetics of
mucopolysaccharide biosynthesis in cultured chick embryo vertebral cartilage. J Biol
Chem 1974; 249:3091-7.
Vidal y Plana RR, Bizzarri D, Rovati AL. Articular cartilage pharmacology: 1.In vitro
studies on glucosamine and non steroidal antiinflammatory drugs. Pharmacol Res
Commun 1978; 10:557-69.
Bassleer C, Rovati L, Franchimont P. Stimulation of proteoglycan production by
glucosamine sulfate in chondrocytes isolated from human osteoarthritic articular
cartilage in vitro. Osteoarthritis Cartilage 1998; 6:427-34.
Noyszewski EA, Wroblewski K, Dodge GR, et al. Preferential incorporation of
glucosamine into the galactosamine moieties of chondroitin sulfates in articular cartilage
explants. Arthritis Rheum 2001; 44:1089-95.
Hresko RC, Heimberg H, Chi MM, Mueckler M. Glucosamine-induced insulin resistance
in 3T3-L1 adipocytes is caused by depletion of intracellular ATP. J Biol Chem 1998;
273:20658-68.
Patwari P. Mannosamine inhibits aggrecanase-mediated degradation of the
mechanically functional portion of proteoglycans and of the physical properties of
articular cartilage. Dept. of Electrical Engineering and Computer Science. Cambridge,
MA: Massachusetts Institute of Technology, 2000:60.
de Mattei M, Pellati A, Pasello M, et al. High doses of glucosamine-HCI have detrimental
effects on bovine articular cartilage explants cultured in vitro. Osteoarthritis Cartilage
2002; 10:816-25.
Fenton JI, Chlebek-Brown KA, Peters TL, Caron JP, Orth MW. Glucosamine HCI
reduces equine articular cartilage degradation in explant culture. Osteoarthritis Cartilage
2000; 8:258-65.
Morales TI, Wahl LM, Hascall VC. The effect of bacterial lipopolysaccharides on the
biosynthesis and release of proteoglycans from calf articular cartilage cultures. J Biol
Chem 1984; 259:6720-9.
159