Identification of important interactions between subchondral bone ...

F A C U L T Y O F S C I E N C E
U N I V E R S I T Y O F C O P E N H A G E N
Identification of important interactions
between subchondral bone turnover and
cartilage degradation in the pathogenesis of
osteoarthritis
PhD thesis by Suzi Høgh Madsen M.Sc
November 2011

Front page pictures
Digital drawing of a knee symbolizing pain (http://www.spa-resorts.cz/eng/kaleidoskop/osteoarthritis-treatment-1196.html).
Logo from the sponsor of my industrial PhD; The ministry of Science, Technology and Innovation.
Logo from the company where I performed my industrial PhD; Nordic Bioscience A/S.
Drawing of the three cell types investigated in the present thesis; Osteoclast (red), osteoblast (brawn), and
chondrocyte (blue).

Supervisors
University of Copenhagen:
Supervisors
Søren Tvorup Christensen, PhD, Associate Professor
Department of Biology
Universitetsparken 13
DK-2100 Copenhagen OE
stchristensen@bio.ku.dk
Nordic Bioscience A/S:
Morten Asser Karsdal, PhD, CEO
Herlev Hovedgade 207
DK-2730 Herlev
mk@nordicbioscience.com
Anne-Christine Bay-Jensen, PhD
Herlev Hovedgade 207
DK-2730 Herlev
acbj@nordicbioscience.com
Kim Henriksen, PhD
Herlev Hovedgade 207
DK-2730 Herlev
kh@nordicbioscience.com
1

Acknowledgements
Acknowledgements
First, I would like to acknowledge the Ministry of Science, Technology and Innovation (VTU)
and Nordic Bioscience A/S, who funded this industrial Ph.D project.
I would like to thank my supervisors Morten Asser Karsdal from Nordic
Bioscience and Søren Tvorup Christensen from the University of Copenhagen, who have both
provided great scientific guidance and enthusiasm throughout the project and writing process. I
would also like to thank the co-supervisors, Anne-Christine Bay-Jensen and Kim Henriksen, for
their support, excellent coaching, and helpful scientific discussions throughout my time at Nordic
Bioscience. Furthermore, I would like to thank all my colleagues at Nordic Bioscience for helpful
comments, technical help, scientific discussions, good times, and for creating a warm and
pleasant work environment.
Finally, I would like to thank my family and friends who have always supported my
dreams and ambitions. Especially, I would like to thank Rasmus Klinck, for critical proofreading
of my thesis, as well as for good friendship and support throughout my thesis.
__________________________________
Suzi Høgh Madsen
November 2011
2

Preface
Preface
The work presented in this thesis was performed in collaboration between Nordic Bioscience
A/S and the University of Copenhagen under the industrial PhD agreement and with the
supervision of Prof. Morten Asser Karsdal (company supervisor), Dr. Anne-Christine Bay Jensen
(project supervisor, company), Dr. Kim Henriksen (project supervisor, company), and Associate
Prof. Søren Tvorup Christensen (university supervisor). The work was carried out in the period
ranging from October 2008 to November 2011.
The PhD work includes four original research papers listed below. Three papers
have been accepted and published, and one is under review. The individual articles are referred to
in the present thesis by their roman numerals.
The PhD thesis consists of an introduction to the topics of the four papers, the
four papers, and a general conclusion.
Paper I:
Characterization of an Ex vivo Femur Head Model Assessed by Markers of Bone and
Cartilage Turnover
Suzi Hoegh Madsen, Anne Sofie Goettrup, Gedske Thomsen, Søren Tvorup Christensen,
Nikolaj Schultz, Kim Henriksen, Anne-Christine Bay-Jensen, and Morten Asser Karsdal
Cartilage 2011;2(3):265–278
Paper II:
Glucocorticoids exert context-dependent effects on cells of the joint in vitro
Suzi H. Madsen, Kim V. Andreassen, Søren T. Christensen, Morten A. Karsdal,
Francis M. Sverdrup, Anne-Christine Bay-Jensen, and Kim Henriksen
Steroids 2011;76(13):1474-1482
Paper III:
Aggrecanase- and matrix metalloproteinase-mediated aggrecan degradation is associated
with different molecular characteristics of aggrecan and separated in time ex vivo
Suzi Hoegh Madsen, Eren Ufuk Sumer, Anne-Christine Bay-Jensen, Bodil-Cecilie Sondergaard,
Per Qvist, and Morten Asser Karsdal
Biomarkers 2010;15(3):266-276
Paper IV:
Metalloproteinase and cysteine protease profiling in normal and Osteoarthritic cartilage
Suzi Hoegh Madsen, Kim Henriksen, Søren Tvorup Christensen, Morten Asser Karsdal, and
Anne Christine Bay-Jensen
In review in Proteome Science
The industrial PhD project has been supported by the Ministry of Science, Technology and
Innovation (VTU) (08-036109) and Nordic Bioscience A/S.
3

Suzi Høgh Madsen, PhD thesis
4

Contents
Contents
Abstract .......................................................................................................................................................... 7
Sammendrag på dansk (abstract in Danish) ............................................................................................. 9
CHAPTER 1
Objectives .................................................................................................................................................... 11
CHAPTER 2
Introduction ................................................................................................................................................ 13
2.1 Epidemiology and etiology of osteoarthritis ................................................................................ 13
2.2 Biology of the joint .......................................................................................................................... 15
2.2.1 Development of bone from a cartilage template ................................................................. 15
2.2.2 Adult bone - macroscopic and microscopic organization .................................................. 17
2.2.3 The cells of bone and their communications ....................................................................... 19
2.2.4 Articular cartilage ...................................................................................................................... 22
2.3 The pathology of Osteoarthritis .................................................................................................... 25
2.3.1 A chondrocyte is not just a chondrocyte .............................................................................. 26
2.3.2 The extracellular matrix of cartilage in OA .......................................................................... 28
2.3.3 The subchondral bone in OA ................................................................................................. 29
2.3.4 Diagnostics - biochemical markers as a tool ........................................................................ 30
2.4 The effect of Glucocorticoids ........................................................................................................ 33
2.5 References for CHAPTER 2 ......................................................................................................... 35
CHAPTER 3
Overview of OA models ........................................................................................................................... 43
3.1 Pre-clinical models ........................................................................................................................... 43
3.2 Advantages/disadvantages of pre-clinical models ...................................................................... 44
3.2.1 In vitro .......................................................................................................................................... 44
3.2.2 Ex vivo......................................................................................................................................... 44
3.2.3 In vivo........................................................................................................................................... 46
3.3 Ex vivo controls ................................................................................................................................ 47
3.4 References for CHAPTER 3 ......................................................................................................... 48
5

Contents
CHAPTER 4
PAPER I ...................................................................................................................................................... 51
CHAPTER 5
PAPER II .................................................................................................................................................... 67
CHAPTER 6
PAPER III ................................................................................................................................................... 77
CHAPTER 7
PAPER IV ................................................................................................................................................... 89
CHAPTER 8
General discussion and conclusions ...................................................................................................... 101
Future perspectives .................................................................................................................................. 103
Abbreviations ............................................................................................................................................ 104
Appendix I ................................................................................................................................................. 105
Co-author declarations for paper I - IV .......................................................................................... 105
Appendix II ............................................................................................................................................... 112
Publication list ...................................................................................................................................... 112
6

Abstract
Abstract
Osteoarthritis (OA) is the most common form of arthritis and a major cause of pain and
disability, with prevalence increasing with age. At present there are no structure modifying
treatments accepted by the Food and Drug Administration (FDA), emphasizing the need for
understanding the pathogenic processes leading to OA. OA is a complex disease of the entire
joint, including bone, cartilage and the synovium, and it is characterized by the progressive
degradation of articular cartilage, mild synovial inflammation, and alterations of the subchondral
bone. Currently, it is not clear whether the pathogenesis of OA originates in the bone or cartilage
compartment, and little is still known on how these compartments drive disease progression. An
increased amount of evidence suggests a strong coupling between the subchondral bone and the
articular cartilage turnover, with pathological processes occurring simultaneously in both
compartments. Therefore, an optimal intervention strategy for OA likely includes targeting both
bone and cartilage compartments.
The aim of the thesis was to investigate the pathogenesis of OA with respect to
bone and cartilage, and the importance of the interaction between the two tissues. Understanding
the interactions between osteoclasts, osteoblast and chondrocytes may be essential for the
development of future treatments for OA.
Before the launch of this thesis, there were no validated pre-clinical models that
allowed investigation of whole tissue pathology of OA in one single system, which combined the
metabolism of the implicated cell types. Thus, the present thesis focused on the development and
characterization of a novel murine ex vivo femur head model, comprising the three major cell
types involved in the deterioration of joint structure; the osteoblasts, osteoclasts, and
chondrocytes. We have established the ex vivo femur head model with positive and negative
controls, which allow the investigation of the interaction between bone and cartilage.
Furthermore, by using the novel ex vivo model and other pre-clinical models, we evaluated the
effect of glucocorticoids on bone and cartilage and found that the effect of glucocorticoids highly
depend on the activation and differential stage of the cell targeted in the joint.
The thesis also studied the degradation processes of cartilage, focusing on the two
predominant proteins; aggrecan and collagen type II. We investigated the molecular differences
between matrix metalloproteinase- and aggrecanase-mediated aggrecan degradation as a
consequence of their distinct time-dependent degradation profiles. We found that the aggrecan
molecule may be divided into two different protease-specific pools. Furthermore, we investigated
the differences between normal and OA cartilage with respect to endogenous proteases (assessed
by biochemical markers) and found that the degradation markers are released in part by different
proteolytic pathways.
7

Abstract
In conclusion, we have established a pre-clinical whole-tissue model for investigating cell-cell
interactions that reflect parts of the processes in the pathogenesis of joint degenerative diseases.
This model is also an excellent screening model for potential OA drugs. The model confirmed
beneficial effects of glucocorticoids previously observed on articular cartilage, but not bone. The
loss of sulphated glycosaminoglycans (sGAGs) from cartilage in early OA is not equivalent to
loss of aggrecan, since there may be different pools of aggrecan that are sensitive to different
types of proteases. Furthermore, the role of endogenous proteases probably depends on the stage
of OA, as we did not observe the same release of biochemical markers in normal versus OA
cartilage. This highlights the importance of developing different treatments specific for the
different stages of OA.
8

Sammendrag på dansk
Sammendrag på dansk (abstract in Danish)
Slidgigt er den mest almindelige form for gigt der findes og en væsentlig årsag til smerte og
invalidering. Risikoen for at få slidgigt øges betydeligt med alderen. På nuværende tidspunkt
findes der ingen godkendte behandlinger for slidgigt fra den amerikanske Food and Drug
Administration (FDA), hvilket understreger behovet for at forstå de patogene processer, der
fører til sygdommen. Slidgigt er en kompleks sygdom, der involverer hele leddet, hvilket
inkluderer bl.a. knogle, brusk og synovium. Slidgigt kendetegnes ved en gradvis nedbrydning af
ledbrusken, en mild inflammation af synovial-vævet og ændringer af den subchondrale knogle. I
øjeblikket diskuteres det om patogenesen af slidgigt starter i knogle- eller brusk-vævet, og om
vævene hver især er med til at gøre sygdomsforløbet værre. En del forskning tyder på at der i
slidgigt er sammenhæng mellem ændringer i subchondral knogle, ændringer i ledbrusken, og de
patogene processer der forekommer samtidigt i begge væv. Derfor vil den bedste behandling af
slidgigt sandsynligvis være en behandling der både involverer knogler og brusk.
Formålet med PhD-afhandlingen var at undersøge patogenesen af slidgigt med
fokus på knogle- og brusk-delen, og at undersøge hvor vigtig interaktionen er mellem de to væv.
Det kan være afgørende for udviklingen af fremtidige behandlinger for slidgigt at forstå det
samspil, der er mellem osteoklaster, osteoblaster og kondrocytter.
Før jeg startede på denne afhandling, var der ingen etablerede prækliniske modeller,
der kunne undersøge de patologiske processer, der forekommer i slidgigt i ét enkelt system, og
som indeholder de implicerede celletyper fra leddet. Derfor fokuserede denne afhandling på
udviklingen og karakteriseringen af en ny ex vivo femur hoved muse-model, der indeholder de tre
hoved-celletyper, som er involveret i degenereringen af leddet; osteoblaster, osteoklaster og
kondrocytter. Vi har etableret en ny ex vivo femur hoved model med positive og negative
kontroller, der kan undersøge interaktionen mellem knogle og brusk. Herudover, ved at bruge
den nye ex vivo model og andre prækliniske modeller, undersøgte vi effekten af glukokortikoider
på knogle og brusk, og fandt at effekten af glukokortikoider afhænger betydeligt af aktiveringen
og differentieringsstadiet af den undersøgte celle fra leddet.
Afhandlingen undersøgte også de nedbrydningsprocesser vi finder i brusk, med
fokus på de to hoved-proteiner; aggrecan og kollagen type II. Vi undersøgte de molekylære
forskelle mellem matrix metalloproteinase- og aggrecanase-genereret aggrecan nedbrydning, på
grund af deres forskellige nedbrydningsprofiler over tid. Vi fandt at aggrecanmolekylet måske er
opdelt i to forskellige proteasespecifikke puljer. Desuden undersøgte vi de forskelle, der er
mellem normalt og sygt brusk mht. endogene proteaser (målt med biokemiske markører) og
fandt at nedbrydnings-markørerne til dels er frigivet via forskellige proteolytiske veje.
9

Sammendrag på dansk
Samlet set har vi etableret en præklinisk fler-vævs model, der kan undersøge celle-celle
interaktioner, og som afspejler dele af processerne, man ser i patogenesen af degenerative led-
sygdomme. Denne model er også en udmærket screeningsmodel for potentielle medikamenter
for slidgigt. Modellen bekræftede de gavnlige virkninger af glukokortikoider man tidligere har set
på ledbrusk, men ikke på knogle. Udtømning af sulfat-glykosaminoglycaner (sGAG’s) fra brusk i
begyndende slidgigt er ikke ensbetydende med udtømning af aggrecaner, da vi foreslår, at der er
forskellige puljer af aggrecaner, der er følsomme over for forskellige typer af proteaser.
Derudover afhænger rollerne for de endogene proteaser formentlig af det slidgigt-stadie man er i,
da vi fandt forskellige frigivelsesmønstre af biokemiske markører i normalt brusk versus slidgigt
brusk. Dette understreger betydningen af at udvikle forskellige behandlinger, der er specifikke for
de forskellige stadier af slidgigt.
10

CHAPTER 1
Objectives
CHAPTER 1: Objectives
The aim of the PhD project was to gain further insight into the pathogenesis of osteoarthritis
(OA) with respect to bone and cartilage, which are believed to be the primary tissues involved in
joint turnover. In particular, the importance of the interaction between bone and cartilage was
investigated to assess the importance of bone in the pathogenesis of OA. To this end, we wanted
to develop a pre-clinical model that includes the two key tissues, subchondral bone and articular
cartilage. Furthermore, as articular cartilage degradation is believed to be a hallmark of OA
pathogenesis, we investigated the proteolytic processes of pathogenic cartilage in order to
understand the OA-induced changes, which ultimately lead to cartilage loss.
The PhD project had two main objectives:
1) The commercial objective.
The first objective was to develop and validate a pre-clinical model system that allows
investigation of whole tissue pathology in the pathogenesis of OA. This model should contain
the three major cell types involved in the deterioration of joint structure, the osteoblasts,
osteoclasts, and chondrocytes, including their cellular signalling processes (PAPER I). This model
should allow testing of potential OA drugs in a more pathological relevant model. Thus, the
effects of glucocorticoids as a potential drug for OA were tested in this model to evaluate the
usefulness of the model and further to evaluate the effect of glucocorticoids on bone and
cartilage cells (PAPER II).
2) The basic research objective.
The second objective focused on the degradation of collagen type II and aggrecan in OA
cartilage. We hypothesized that the loss of aggrecan in early and late OA was dependent on
different proteases due to different molecular structures of the aggrecan. We expected to find
different pools of aggrecan being degraded by appertaining proteases (so a specific protease
should be targeted depending on the disease stage) (PAPER III). We also hypothesized that the
ratio between endogenous proteases in OA cartilage, depends on the stage of the disease.
Furthermore, we wanted to investigate if the biochemical markers were able to assess the
different endogenous proteases in normal vs. OA cartilage (PAPER IV).
11

Suzi Høgh Madsen, PhD thesis
12

CHAPTER 2
Introduction
CHAPTER 2: Introduction
This Introduction aims to introduce the knowledge that forms the basis for my thesis work. The
scope of the introduction is: i) to provide an overview of the development and biology of the
joint, ii) the degeneration of the joint, with emphasis on osteoarthritis (OA), iii) and to introduce
the current knowledge on the mode of action of glucocorticoids.
2.1 Epidemiology and etiology of osteoarthritis
There are more than 100 different types of arthritis 1 . The most common type is OA, also known
as degenerative joint disease or degenerative arthritis. It is a condition that affects more than 46
million people in the USA, and it is expected to affect 67 million people by the year 2030 1 .
According to Gigtforeningen (The Danish Arthritis Society), more than 210.000 people in Denmark
are diagnosed with OA, but experts estimate that the actual number is much higher. Although the
disease occurs in people of all ages, it is most common in the older population; 50% of people
over the age of 40 years and 100% of people over the age of 60 years have OA in one or more
joints 2 . Epidemiologic studies further suggest that there are clear sex-specific differences 3 . Before
50 years of age, the prevalence of OA in most joints is higher in men than in women. After the
age of 50 years, women are more often affected with hand-, foot-, and knee OA than men 4 .
The etiology of OA is unknown. The general perception of OA is that it is a
cartilage disease. However, OA is not only affecting the articular cartilage, but the entire joint,
including the subchondral bone, ligaments, joint capsule, synovial membrane and periarticular
muscles. However, loss of cartilage is the obvious central hallmark of OA 5 . An increasing line of
evidence suggests that both bone and cartilage contribute to the onset of the disease. One of the
first indications came from clinical observations suggesting that osteopenic women generally do
not develop severe OA 6 . Likewise, subjects with OA of the hip have greater bone mass than
normal subjects or subjects with osteoporosis 7 , suggesting that OA could initially be a disease of
the bone 8 . To date, the majority of clinical studies support the idea that OA is associated with
increased bone density, but the subject remains controversial 9 .
OA commonly affects hands, feet, spine, and large weight-bearing joints, such as
the hip and knees. In most cases, the mechanisms underlying the development of OA are
unknown. These cases are classified as primary OA and are mostly related to aging or heredity.
When the cause of OA is known, as after an injury, trauma, anatomic abnormalities or obesity,
the condition is referred to as secondary OA 5,10,11 . OA is characterized by gradual loss of articular
cartilage, alterations of the subchondral bone, and development of osteophytes. These alterations
13

CHAPTER 2: Introduction
lead to pain and limitation of joint mobility, which decrease the quality of life for the patients and
become a burden on the society 5 . Today, treatment of OA is based on anti-inflammatory drugs
(e.g. NSAID), analgesics or arthroplastic surgery 12 , which neither slow down nor cure the disease.
Moreover, no disease modifying osteoarthritic drugs (DMOADS) have been developed and
approved by the American Food and Drug Administration (FDA) yet. Therefore, the discovery
and the development of effective disease-modifying drugs are urgently needed.
The large number of people affected by OA is increasing
very fast. The increased life expectancy of the elderly
population and the obesity epidemic, affirms the necessity
to find a treatment for OA as fast a possible: Both to
prevent an “OA-epidemic”, but also to avoid the great
social-economical costs in the future.
14

2.2 Biology of the joint
CHAPTER 2: Introduction
In order to get insight into the underlying processes of OA, it is important to understand the
physiology and biology of healthy joints. Joints that are able to move freely are termed synovial
joints and together with muscles and bone they represent the machinery that enables us to move.
The knee is a synovial weight-bearing joint, which comprises many tissues that maintain the
mobile function; articular cartilage, subchondral bone, ligaments, capsule, synovial membrane and
supporting muscles (fig. 1A). The ends of the adjoining bones are covered with articular cartilage,
which functions as a shock-absorber and provides a friction-free surface that enables the bones
to move smoothly against each other. The joint capsule, tendons, muscles and ligaments provide
stability of the joint. The bones are connected by strong ligaments and the tendons provide the
attachment of muscles to the bone. The inside of the joint is covered with a synovial membrane,
which provides and holds the lubricating synovial fluid that enables the joint to move easily (see
fig. 1B) 13,14 .
Fig. 1. Illustration of a healthy synovial knee joint. The different components of a knee
joint are outlined in figure A and B. A) The components responsible for stability and
movement are shown. B) The components of the articular capsule, which provide an easy
friction, are shown in details. The figure was produced by Madsen, S.H.
2.2.1 Development of bone from a cartilage template
Although OA is a disease of the whole joint, this thesis focuses on bone and cartilage, which are
the most affected tissues in OA. A description of the development of the joint, with respect to
bone and cartilage, will help the reader to understand the tight connections between these two
tissues.
15

CHAPTER 2: Introduction
Bone is a specialized tissue, which together with cartilage make up the human skeleton. The
bones have three main functions; 1) providing mechanical support for the whole body and acting
as muscle attachment for locomotion, 2) providing protective shields for the vital organs and the
bone marrow, and 3) providing a reserve for ions, mainly calcium and phosphate, and for growth
factors and cytokines 15,16 .
There are several subtypes of bones; long bone, flat bones, short bones and
irregular bones 17 , of which only long bones - e.g. femur and tibia, which constitute the knee - will
be discussed further in this thesis. Long bones are developed by endochondral ossification as
shown in fig. 2A 18 . This whole process starts with a condensation of mesenchymal cells to form
the cartilage anlage of the future bone. The mesenchymal cells further proliferate and
differentiate into chondroblasts. The chondroblasts produce a cartilaginous matrix, where the
predominant collagen is collagen type II. The chondroblasts become progressively embedded
within their own matrix, where they lie within lacunae, and they are then called chondrocytes (the
only cell type in articular cartilage). Eventually, the chondrocytes may further differentiate into
hypertrophic chondrocytes whereby the cartilage becomes calcified before the hypertrophic
chondrocytes undergo apoptosis and then disappear from the lacunae. This generates the primary
ossification-centre, which splits into two opposite growth plates that moves from the centre of
the long bone, towards the ends. At this stage of bone development, the blood vessels penetrate
the calcified matrix supplying osteoclast precursors. Differentiated osteoclasts (bone resorbing
cells) use the lacunas in the calcified cartilage matrix as templates to initiate resorption.
Fig. 2. Development of long bone by endochondral ossification. Schematic diagram showing the initial
stages of endochondral ossification. A) Endochondral ossification is initiated from accumulated mesenchymal
cells, which continue to differentiate and starts forming the primary, and then later the two secondary
ossification centres. Trabecular bone is present after several cycles of remodelling 16 . B) The growth plate
between the epiphysis and diaphysis is enlarged and divided in the different zones. The growth plate
demonstrates the different stages of chondrocyte differentiation involved in endochondral bone formation.
The figure was produced by Madsen, S.H.
16

CHAPTER 2: Introduction
Also osteoblasts (bone forming cells) arrive with the bloodstream and form a layer of woven
bone on top of the remaining cartilage. The septae of calcified cartilage separating the lacunae are
first resorbed by osteoclasts, while the remaining septae that extend into the diaphysis are used
for osteoblasts to deposit bone matrix, resulting in the vascular bone marrow cavity (fig. 2A) 19 .
The bone continues to grow through both the growth plates in appertaining epiphyses.
In the growth plate (see fig. 2B) resting chondrocytes resume proliferation
organized in columns toward the diaphysis and then mature into hypertrophic cells that express
specific genes, such as collagen type X, and calcify the matrix. After apoptosis of the
hypertrophic chondrocytes, the calcified matrix is resorbed by osteoclasts followed by bone
deposit by osteoblasts. This process is known as the bone modelling process, where bone
resorption and formation occur independently of each other. This modelling continues as long as
needed, and eventually, secondary ossification-centres begin to form at the epiphyseal ends of the
long bone.
At the end of human adolescents, the proliferation of chondrocytes in the growth
plate slows down and eventually stops. The continuous replacement of cartilage by bone results
in the obliteration of the growth plate. Only articular cartilage remains. During primary bone
formation, woven bone structures are first formed, which comprise some amount of calcified
cartilage. The woven bone will later be replaced by the more robust lamellar bone without
calcified cartilage, during remodelling of the bone matrix (described in section 2.2.3), which is a
different process than the modelling process. Mineralization of articular cartilage and its
replacement by bone continues in the adult, however, at a much reduced rate than in growing
bone 19,20 .
2.2.2 Adult bone - macroscopic and microscopic organization
After a long bone is fully developed, it has two epiphyses covered with a layer of articular
cartilage, a cylindrical hollow portion in the middle called the diaphysis, and a transition zone
between them called the metaphysis (see fig. 3) 16 .
The long bone is divided in two types of structural bone (fig. 3):
I) The external part of the long bone is formed by a thick and dense layer of calcified tissue called
the cortex, which encloses the medullary cavity where the hematopoietic bone marrow is
housed 17 .
II) Toward the metaphysis and the epiphysis, the cortex becomes progressively thinner and the
internal space is filled with a network of thin, calcified trabecular bone. The spaces enclosed by
these thin trabeculae are also filled with hematopoietic bone marrow and are continuous with the
diaphyseal medullary cavity 17 .
17

CHAPTER 2: Introduction
18
Fig. 3. Schematic drawing of a long
bone. A femur bone divided in sections:
epiphyses, metaphyses, and a diaphysis.
The different types of bone are shown;
cortex (hard) and trabecular (spongy)
bone. The proximal epiphysis (femur
head) and the distal epiphysis are
covered with cartilage to provide a
smooth friction (during movement). The
figure was produced by Madsen, S.H.
The weight bearing bones, such as the femur and tibia that comprises the knee, have two
opposite and important functions. The first is to resist loading, thus the bone must be stiff. The
second is to be flexible to withstand shock, compression and tension by energy absorption and
deformation 21 . The stiffness is achieved by mineralization and flexibility is achieved by a suitable
organic matrix composition. However, if the bone matrix is too mineralized, it will become brittle
and more prone to fractures, whereas if it is under-mineralized, the bone will be too flexible and
will crack more readily. Therefore, the functionality of the bone highly depends on the ratio of
organic and inorganic matrices 21 .
In mammalian bone, the inorganic part of the matrix (50-70%) consist of
hydroxyapatite crystals [3Ca 3(PO 4) 2·(OH) 2], which are layered on and within collagen fibres,
ensuring bone matrix stiffness (mineralization). The crystals tend to orient in the same direction
as the collagen fibres. The organic part of the matrix (20-40%) consists mainly of collagen type I
(approximately 90%), but also glycoproteins, proteoglycans, and non-collagenous proteins. The
collagen type I molecule is a triple helix (described further in section 2.3.4), which is assembled
into fibres, providing the tensile strength of the matrix. The orientation of the collagen fibres
determines the microscopic structure of the bone. During bone development or fracture healing,
where bone is formed very rapidly, there is no preferential organization of the collagen fibres; this
type of bone is called woven bone (fig. 4) 17 . Woven bone is characterized by irregular bundles of
collagen fibres and calcification, which occur in irregularly distributed patches. Woven bone is
progressively replaced by mature lamellar bone (fig. 4) during the remodelling process that

CHAPTER 2: Introduction
follows normal development or healing. Lamellar bone is characterized by organized fibres,
which are forming arches for optimal bone strength. The lamellae can be parallel to each other if
deposited along a flat surface (trabecular bone) or concentric if deposited on a surface
surrounding a channel centred on a blood vessel 17,22,23 .
2.2.3 The cells of bone and their communications
19
Fig. 4. Lamellar and woven bone.
Histological cut from fibula of rabbit
showing details of lamellar bone
concentrically organized and woven
bone mixed with cartilage and calcified
cartilage tissues 24 .
Bone contains several types of cells, which are important during development, repair, and
maintenance. During bone development, the woven bone is modelled by resorption of calcified
cartilage by osteoclasts and formation of bone by osteoblasts, independently of each other (not
directly coupled) 16,25,26 . This is different in adult bone, where the tissue is constantly being
remodelled by a communication/coupling between osteoclasts and osteoblasts, which enable
bone to regenerate old or damaged tissue itself, adapt to changing stress environments, and
control calcium homeostasis.
Activation of the bone remodelling cycle is believed to be initiated by the
osteocytes 15,21,27,28 , which comprises the most abundant cell type in bone. Osteocytes are cells
trapped in the bone matrix. They rest in fluid-filled lacunae and communicate with surrounding
cells via extensions called canaliculi 23,29 . When bone ages, the tissue eventually suffers
microdamage and the surrounding osteocytes undergo apoptosis, which leads to signalling to the
nearest bone lining cells (fig. 5A). This results in increased expression of receptor activator of
nuclear factor kappa-B ligand (RANKL), which attracts osteoclast precursors to the damaged
area and initiates osteoclastogenesis. Osteoclasts derived from hematopoietic stem cells that are
differentiated into monocytes and subsequently pre-osteoclasts, which then fuse into large
multinucleated osteoclasts. The two cytokines, macrophage colony stimulating factor (M-CSF)
and RANKL, are produced by the cells of the osteoblast lineage, and are very important for the
differentiation of the osteoclasts (fig. 5A).

CHAPTER 2: Introduction
20
Fig. 5. The remodelling cycle on trabecular
bone. A) A microcrack or old bone causes
osteocytic apoptosis, which induces RANKL
expression by bone lining cells. This attracts
monocytes to the damaged bone area, and
osteoclastogenesis starts. The osteoclasts will
start to resorb bone, which is followed by
formation of new bone matrix by the osteoblasts.
Osteoblasts that are trapped in the remodelled
matrix become osteocytes, whereas the rest
either die or become flattened osteoblast lining
cells. B) The resorbing osteoclast seals tightly to
the bone surface to create a closed local
environment known as the resorption lacuna.
This is where the bone resorption takes place, by
lowering the pH (secretion of H + ) and by
releasing proteases by exocytosis. The figures
were produced by Madsen, S.H.
Active osteoclasts can be characterized by their ruffled cell borders, which seal and degrade the
underlying bone in the resorption pit, through the release of hydrogen ions (lowering of pH) and
proteases (such as cathepsin K and matrix metalloproteases). The resorbed bone tissue is
transported across the osteoclast in small vesicles and is released into the extracellular fluid (fig.
5B.) 21 .

CHAPTER 2: Introduction
When the job is done, the osteoclasts undergo apoptosis, and osteoblast precursors are attracted
to the site and are differentiated into mature osteoblasts. Osteoblasts are mononucleated bone
forming cells, which derive from mesenchymal stem cells that are differentiated into
osteoprogenitor cell (in the presence of bone morphogenic protein (BMP)), and further into pre-
osteoblasts (fig. 5A). The mature osteoblasts deposit osteoid in which the collagens (mainly
collagen type I) are layered in an orderly fashion at sites of resorption, which converts into
lamellar bone when mineralized (a function also executed by the osteoblasts) 30,31,32,33,34 . When an
adequate amount of bone has been formed, the bone remodelling cycle ceases, and the
osteoblasts trapped in the bone matrix become osteocytes. The remaining cells undergo
apoptosis or differentiate into bone lining cells (fig. 5A) 27,35,36 .
The remodelling process showed in fig. 5A outlines important communication
processes from the osteoclasts to the osteoblasts, which is known as the coupling process that
ensures that resorbed bone is replaced by an equal amount of new bone 35,36,37,38 . There are several
theories about how this coupling process is regulated. These includes; the action of factors
released from the bone during resorption 27,35,36,39 , a closed bone remodelling
compartment 25,40,41,42,43 , a secretion of factors from the osteoclasts themselves or surface
signalling 35,44,45,46 . Reversely, osteoblasts are very important for the osteoclastogenesis, as they are
the source of M-CSF, RANKL and osteoprotegerin (OPG), in which the latter is a soluble decoy
receptor for RANKL 31,33,47 . Further, in the normal regulation of calcium homeostasis, parathyroid
hormone (PTH) stimulates osteoblasts to release RANKL and decreases the level of OPG. The
increased level of RANKL then stimulates osteoclasts to resorb bone 26,48 . These findings support
the conclusion that diverse and complex signalling systems coordinate the communication
between the bone cells.
The communication between osteoclasts and osteoclasts
is evident. The signal to osteoclasts through PTH via
osteoblasts is clear, whereas the coupling from
osteoclasts to osteoblasts is still debated.
No matter what... there is a communication between
the different cell types!
21

2.2.4 Articular cartilage
CHAPTER 2: Introduction
In a normal knee joint, the epiphyses of femur and tibia are covered with a layer of articular
cartilage, which functions as shock-absorber and provides a friction-free surface that enables the
bones to move smoothly against each other. Articular cartilage is avascular and is not innervated.
In this way, nutrients from the synovial fluid are transported to the chondrocytes (the only cell
type in cartilage), and waste products are transported back to the synovial fluid by diffusion.
Chondrocytes are cartilage specific and comprise 2-5% of the total cartilage 5,11,14,49 . Their function
is to produce and maintain the surrounding matrix. Hence, the physical condition of cartilage
ultimately depends on the health of these cells. Similar to osteoblasts, chondrocytes derive from
the mesenchymal stem cells 19,50 , and they differentiate into pre-hypertrophic and hypertrophic
chondrocytes before they undergo apoptosis. Even though chondrocytes are metabolically very
active, they normally do not divide after adolescence. Only small defects associated with minimal
loss of matrix components are regenerated. More extensive defects exceed the repair capacity,
and consequently the damage becomes permanent 51,52,53 .
Cartilage tissue has unique viscoelastic and compressive properties provided by the
extracellular matrix (ECM) (fig. 6), which consists mainly of collagen type II (15%) and forms a
fibrous network that entraps proteoglycans (10%), of which aggrecan is the predominant
molecule. The aggrecan molecule is a protein that covalently binds sulphated glycosaminoglycan
(sGAG) chains to the core protein, which is non-covalently attached to hyaluronic acid. The
aggrecans attract water molecules, due to the high negatively charged sGAGs, and allow the
tissue to withstand compressive forces (water comprises approximately 70% of
cartilage) 5,14,49,54,55,56 . The ECM also contains a number of less abundant molecules such as minor
collagens, growth factors and adhesive molecules, which are out of the scope of this thesis 14 .
22
Fig. 6. Cartilage components.
The articular cartilage is mainly
composed of collagen type II,
chondrocytes and aggrecan
attached to hyaluronic acid. The
figure was produced by
Madsen, S.H.

CHAPTER 2: Introduction
Cartilage is divided into different zones that have different chondrocyte phenotypes,
compositions, structures, and biomechanical properties (fig. 7). From the articular cartilage
surface and down to the subchondral bone, we find the superficial zone, the middle zone, the
deep zone and the calcified cartilage zone (fig. 7A). Since the superficial zone is exposed to
compression, tensile forces and diffusion, it has flattened chondrocytes and is rich in collagen
fibres that are parallel arranged to the surface (fig. 7B). In the middle zone, chondrocytes have a
spherical shape and are randomly distributed. The collagen fibres are arranged in an angulated
manner at which the compression load is better dispersed. The deep zone is responsible for
providing the greatest resistance to compressive forces. The collagen fibres are arranged in a
radial disposition, with the highest proteoglycan content, and the lowest water concentration. The
chondrocytes are still spherical, but are typically arranged in columnar orientation parallel to the
collagen fibres. The tide mark distinguishes the deep zone from the calcified cartilage zone. The
calcified layer plays an integral role in securing the cartilage to bone by anchoring the collagen
fibrils of the deep zone to the subchondral bone. In the calcified zone, the cell population is
scarce and chondrocytes are hypertrophic 57,58 .
Fig. 7. Schematic illustration of the zones observed within
articular cartilage. A) 3D view of articular cartilage divided in
different zones. B) The arrangements of collagens in the different
zones are illustrated in the boxes next to the cartilage zones. Adapted
with modifications from Sah et al. 59 .
23

CHAPTER 2: Introduction
In healthy mature cartilage, the normal turnover of the ECM is kept in equilibrium by cytokines
and growth factors, which stimulate the production of proteinases and matrix components.
Chondrocytes maintain a low turnover rate with collagen having a half-life of more than 100
years and aggrecan having a half-life ranging from 3-24 years. In OA, the turnover is moved
toward more degradation than formation of the cartilage 11,14 .
Chondrocytes and osteoblasts both derive from
the mesenchymal stem cells, suggesting a
possible connection or natural communication
between the cells.
24

2.3 The pathology of Osteoarthritis
CHAPTER 2: Introduction
OA is a complex disease of the entire joint. It is mainly characterized by the progressive
degradation of articular cartilage, formation of osteophytes (bone outgrowths) and alterations and
thickening of the subchondral bone (fig. 8) 60,61 . The main focus in the past decades has been on
the articular cartilage as the affected tissue and biomechanics as the causative agent. The
alterations in subchondral bone were believed to be secondary to the enforced inactivity.
However, several studies have shown that normal cartilage turnover is dependent on a healthy
subchondral bone 62,63 . Today, it is openly debated whether the pathogenesis of OA originates in
the bone or in the cartilage compartment, and it is not clear to which extend these two
compartments drives disease progression. An increased amount of evidence suggests a strong
coupling between the subchondral bone and the articular cartilage turnover with pathological
processes occurring concurrently in both compartments. Therefore, an optimal intervention
strategy for OA likely includes the targeting of both bone and cartilage compartments that affects
the pathophysiology of the whole joint by modulating and altering the interactions between the
different cell types.
Fig. 8. Illustration of a knee with severe osteoarthritis.
A severe chronic osteoarthritic knee, with eroded cartilage,
eroded meniscus, exposed bone and osteophytes (bone
outgrowths). Adapted with few modifications 64 .
25

CHAPTER 2: Introduction
A schematic overview of the histopathological progression of OA is shown in fig. 9 65 .
In very early OA, the superficial layer of cartilage is lost, the subchondral bone is getting
denser, and blood vessels are penetrating the calcified cartilage.
This affects the rest of the cartilage zones, which lose their organized matrix structures
(fibrillation and loss of proteoglycans) and chondrocyte order (cluster formation).
Simultaneously, the tidemark duplicates and is penetrated by blood vessels, and the
subchondral bone increases in size and density as the cartilage is calcified and replaced by
new bone.
In moderate OA, the organization of the collagen network and chondrocyte column
organization are lost and some of the chondrocytes have differentiated into hypertrophic
chondrocytes. The subchondral bone still increases in size and penetrate the cartilage
compartment.
In late OA, only some parts of the deep zone of cartilage is left and the subchondral bone has
increased in size and intermingled with cartilage (calcified zone has disappeared).
At the end stage OA, the cartilage is lost and the subchondral bone is exposed. This may lead
to synthesis of woven bone, which results in deformation of the joint.
Fig. 9. The pathological development of OA. The progression of OA is shown from the healthy stage to the
end stage of the disease, with focus on the matrix structure, cell number and phenotype, and ratio between
cartilage and subchondral bone. This figure illustrates the simultaneous processes of both bone and cartilage in
the pathogenesis of OA. Figure adapted from Bay-Jensen et al. 65 .
2.3.1 A chondrocyte is not just a chondrocyte
Most people develop OA in later stages of life. One of the main reasons lies within the adult
articular chondrocytes, which are ultimately responsible for remodelling and maintaining the
health of the cartilage, even though these cells possess only little regenerative capacity. Research
has shown that chondrocyte proliferation is rare in normal adult cartilage, and a reduction in the
26

CHAPTER 2: Introduction
number of active chondrocytes occurs with age along with a reduction in their response to
anabolic factors. Therefore, the increased loss of chondrocytes during the later stages of life and
thus the increasing lack of regeneration eventually result in damaged cartilage and the
development of OA 66,67,68,69,70 . However, in the pathogenesis of OA, chondrocytes proliferate to
form multiple cells within a chondron, which is a pivotal change from normal cartilage, where
more than 2–3 cells are unusual, even in the elderly 71 . These clusters of chondrocytes in OA are
not to be mistaken for the chondrocytes in healthy adult cartilage, as chondrocytes can change
into different phenotypes during the progression of OA:
An anabolic phenotype that expresses ECM proteins; e.g. collagen type II and aggrecan 72
A catabolic phenotype that expresses proteolytic enzymes that degrade the ECM 11,73,74
A Hypertrophic phenotype that expresses collagen type X before it undergoes apoptosis 75
A dedifferentiated phenotype that expresses collagen type I (often in repair areas after damage 76,77
A chondroblastic phenotype that expresses fetal type IIA collagen and type III collagen 78
The differentiation of chondrocytes into these phenotypes probably occurs due to a “panic-
attack” from the insufficient adult chondrocytes, who cannot keep up with the repair
mechanisms of damaged cartilage. As a solution, the very productive phenotypes seen from the
skeletal development are starting to emerge. Evidence of phenotypic change for chondrocytes is
reflected in the presence of collagens not normally found in adult articular cartilage. As an
example, the splice-variant of collagen type II, type IIA (which is normally expressed by
chondroblasts during development) is re-expressed in early and late-stage OA 78 . Other indicators
of a potential reversion of the cells to an earlier developmental phenotype are the hypertrophic
chondrocyte marker, collagen type X 79 as well as the general increased production of proteins,
proteinases, growth factors, cytokines, and other inflammatory mediators by more
active/productive chondrocytes 80,81,82 .
The adult articular chondrocytes normally maintain the cartilage
homeostasis with a low turnover of matrix constituents. In OA,
the chondrocytes attempt to recapitulate phenotypes of early stages
of cartilage development, whose rate of synthesis of matrix
constituents are faster. However, and unfortunately, the original
cartilage cannot be replicated.
27

2.3.2 The extracellular matrix of cartilage in OA
CHAPTER 2: Introduction
The health of normal cartilage is maintained by a balance between matrix synthesis and
degradation. This balance is disturbed in OA. In early OA, there is increased synthesis of the
principal matrix molecules of cartilage 82 , including collagen type II and aggrecan. However, the
increased levels of these matrix molecules may be exported to the synovial fluid rather than
incorporated into the tissue 56,83 . This is seen by an elevated rate of cartilage proteoglycan turnover
in emerging OA and observed in the serum as an increase of aggrecan fragments in the synovial
fluid and sulphate epitopes 84 . Thus, there is a decrease in total proteoglycan concentration, chain
length and aggregation in OA cartilage 85 .
Proteoglycans play an important role in keeping the water in
the matrix. Thus it is important to understand the
degradation processes of aggrecan in OA, to find the right
treatment for the right stages of OA. This thesis investigates
the different degradations processes for aggrecan (PAPER III)
The elevated matrix synthesis in OA is accompanied by increased synthesis of proteases and
other inflammatory factors, which may be related to elevated levels of interleukin-1 (IL-1) and
tumour necrosis factor α (TNF-α) 86,87 . Multiple proteases are involved in the degradation of
cartilage, including cysteine-, aspartic-, serine-, and metalloproteinases. Especially, the expression
of the zinc-dependent matrix metalloproteinase (MMP) family of enzymes is highly up-regulated
in OA cartilage 88,89 . Mainly, MMP-1, -2, -3, -8, -9, -13 and -14 play a major part in protein
degradation. MMP-1, -8 and -13 (collagenases) mainly cleave collagen type II in the triple
helix 81,90,91,92 , whereas MMP-2 and -9 (gelatinases) cleave in the N-teleopeptide end 93 . Most MMPs
are secreted by chondrocytes as latent proMMPs, which can be activated by other proteases
through cleavage. As an example, MMP-2, -3 and membrane bound MMP-14 cleave other
proenzymes (e.g. MMP-13) 94,95 . MMP-3 (stromelysin) also breaks down collagen type II and
proteoglycans 96,97 . The breakdown of proteoglycans is also mediated by the aggrecanases, a
disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS), and this breakdown
is thought to be an early step in the loss of cartilage in OA 98 .
The activity of MMPs and ADAMTS are normally controlled by the endogenous
tissue inhibitors of metalloproteinases (TIMPs) 99 , which are also expressed by chondrocytes. An
imbalance between metalloproteinases and TIMPs, results in cartilage degradation 89,100 . The
expression of these TIMPs is down regulated by TNF-α in OA 101 , which furthermore contributes
to the imbalance between metalloproteinases and their inhibitors.
28

CHAPTER 2: Introduction
Cysteine proteases are also essential players in the proteolytic cleavage of ECM in OA, which are
synthesized both by the chondrocytes and synovial cells in response to cytokines and growth
factors 102 . Cysteine proteases are normally responsible for the breakdown of collagen in bone,
especially collagen type I 89 . However, cysteine proteases, like cathepsins K, B and L, have also
been implicated as major participants in the degradation of collagen type II and
proteoglycans 103,104,105,106 . As an example, expression of cathepsin K is up regulated in articular
chondrocytes in a transgenic mouse model for OA 102 and urine concentrations of the cartilage
breakdown products of collagen type II (CTX-II) are reduced by chronic cathepsin K
inhibition 107 . Furthermore, expression of cathepsin B, H, L, and S are seen in hypertrophic
chondrocytes, which are a hallmark of OA 108 . Besides the above mentioned actions, the main role
of cathepsins is believed to be indirect degradation of matrix proteins through activation of
MMPs 109,110,111 .
2.3.3 The subchondral bone in OA
The subchondral bone turnover increases in OA 5,112 , which is demonstrated by the increase in
alkaline phosphatase, osteocalcin, MMP expression (MMP-2 and -3), cytokines (TNF-α, IL-1, IL-
8, IL-10), transforming growth factor beta (TGF-β) and TIMP-1 113,114 . Furthermore, the
thickening of the subchondral bone with an abnormally low mineralization pattern 115 supports a
greater proportion of osteoid in the diseased tissue with a higher turnover rate 112 . These changes
in mineralization and structural organization, of the subchondral bone compartment, result in
tissue stiffness. This stiffness impairs the ability of the subchondral bone to act as shock absorber
to the overlying cartilage 116 .
As described in section 2.1, the etiology of OA is unknown. An increasing line of
evidence suggests that both bone and cartilage, and not cartilage alone, contribute to the onset
and progression of the disease. Several studies have addressed the importance of subchondral
bone in the pathogenesis of OA. In a study with a primate model of spontaneous OA, the
progression of the disease severity of cartilage correlated with the increased thickness of the
subchondral bone. Moreover, the bone changes preceded the changes in cartilage 117 . Another
study showed a correlation between cartilage loss in OA patients and increased subchondral bone
turnover as measured by uptake of technetium-labelled bisphosphonate 118 . These studies confirm
the importance of subchondral bone, either as an initiator or as a “partner in crime” in the
pathogenesis of OA.
The knowledge of subchondral bone in OA is limited. Thus this area
needs to be investigated more thoroughly, as both tissues are involved in
the pathogenesis of OA. This thesis also focus on the communication
between the cartilage and subchondral bone (PAPER I and II)
29

CHAPTER 2: Introduction
2.3.4 Diagnostics - biochemical markers as a tool
Several methods may be used to detect OA. X-ray detection is the golden standard method, where
narrowing of the joint space indicates cartilage loss. Other ways to detect cartilage damage include
arthroscopy and magnetic resonance imaging (MRI) 54,119 . However, OA is often diagnosed at late
stages when the pain is unbearable for the patients, and the damages are too severe to repair.
Today, many biochemical markers have been developed for the purpose of diagnosing the stage or
degree of a specific disease.
A biochemical marker is a single molecule or fragment of a molecule that is released
into biological fluids (e.g. blood and urine) during tissue pathogenesis 54 . An increase in a specific
biochemical marker indicates alterations of turnover, i.e., increased synthesis or breakdown 120 .
Biochemical markers give a more accurate diagnosis of the severity of a disease for each patient, so
that treatment could be designed specifically for the individual patient. Furthermore, the response
to treatment or slow-down of the disease-progression can also be detected with these biochemical
markers.
OA is a disease affecting the whole joint and potential biochemical markers could
originate from cartilage, bone, synovium, ligaments, tendons and muscles. However, the systemic
levels from muscles, and some of the other tissues may exceed the small alterations seen from the
OA-joint. At present, clinical studies indicate that different biochemical markers of cartilaginous
matrix proteins may be useful to predict the early stages of cartilage damage and progression in
OA. However, detection of OA by biochemical markers alone is not possible yet, as they measure
systemic levels 54 . Nevertheless, the market for developing biochemical markers is increasing as it is
possible to design more target-specific biochemical markers (e.g. neo-epitopes with post-
translational modifications) or combine a specific selection of markers that represent one unique
disease 121,122 .
In OA, a central hallmark is the gradual destruction of articular cartilage and
alterations in the subchondral bone 121 . The turnover of subchondral bone increases in early OA,
which can be measured by increased formation and degradation of collagen type I (fig. 10A) and
osteocalcin 123,124 . The resorption of collagen type I by osteoclasts is mediated by the protease,
cathepsin K, which generates the neo-epitope, C-terminal telopeptide of collagen type I
(CTX-I) 125,126,127 . However, studies have shown that both CTX-I and osteocalcin decrease in the
body fluids of OA patients, indicating that overall bone turnover decreases in OA, whereas the
increased turnover seen in subchondral bone is lost in systemic levels 128,129 .
The turnover of cartilage is normally maintained by a balance between catabolic and
anabolic processes; however, in the case of pathological matrix destruction, the rate of cartilage
degradation exceeds the rate of formation, resulting in a net loss of cartilage matrix. During
degradation of articular cartilage, MMPs and aggrecanases are considered the most important
proteases for degradation of articular cartilage 130,131,132 . The key target proteins in cartilage is
collagen type II and aggrecan (fig. 10B+C). Other important molecules in articular cartilage
30

CHAPTER 2: Introduction
include cartilage oligomeric matrix protein (COMP), link protein and hyaluronic acid, which are
out of the scope of the thesis.
Fig. 10. Schematic illustration of proteolytic cleavage sites on collagen type I, collagen type II, and
aggrecan. A) Collagen type I from bone. Formation of collagen type I is measured by PINP and PICP.
Degradation of collagen type I by osteoclasts through cathepsin K cleavage is measured by CTX-I. B) Collagen
type II from cartilage. Formation of collagen type II is measured by PIINP, PIIANP, and PIICP. Degradation of
collagen type II by MMP cleavage is measured by CTX-II or CIIM. C) Aggrecan from cartilage. Aggrecan
degradation is measured by the MMP-mediated marker, 342-G2, and the aggrecanase-mediated marker, 374-G2.
The sGAGs (red lines) are also measureable as a marker, representing aggrecan turnover. The three figures were
produced by Madsen, S.H.
31

CHAPTER 2: Introduction
The degradation and formation of collagen type II can be measured by several biochemical
markers 133 , but the present thesis mainly focuses on the MMP-mediated biochemical markers; e.g.
C-terminal telopeptide of collagen type II (CTX-II) 129,134 and the novel CIIM 135 (fig. 10B). Several
studies have shown that the CTX-II level is elevated in OA patients and in cartilage degradation
from ex vivo-cultures 134,136,137 . It remains to be resolved whether this increase in CTX-II results from
degradation of pre-existing collagen type II or newly synthesized collagen 138,139 . The formation
markers of collagen type II, PIINP or PIIANP can in combination with the CTX-II or CIIM
markers provide a more complete understanding of the turnover-actions. PIIANP is an attractive
formation marker for OA-cartilage as it is an indicator of expression of the fetal form of collagen
type IIA, which is re-expressed by OA chondrocytes 78,140 . The formation markers, PIINP and
PIIANP, are normally suppressed in OA patients 141,142 . Another study has shown that PIIANP
levels are increased in OA patients 138 , suggesting that PIIANP values are not always consistent.
The degradation of aggrecan has attracted attention as a biochemical marker, as
aggrecan degradation occurs early in OA and prior to collagen degradation 143 . Both MMP and
aggrecanase-cleaved aggrecan fragments have been identified in synovial fluid 144 . The
concentration of aggrecanase-mediated fragments (374-G2) dominated the MMP-mediated
fragments (342-G2) in cartilage degradation ex vivo 144,145 , suggesting that the major mediators of
aggrecan loss in OA are the aggrecanases.
Biochemical markers are key-tools used to
investigate the hypotheses in this thesis.
32

2.4 The effect of Glucocorticoids
CHAPTER 2: Introduction
Glucocorticoids (GCs) belongs to a class of steroid hormones that bind to the glucocorticoid
receptor (GR). This receptor is present in almost every vertebrate animal cell and activation of
GRs initiate an up-regulation of the expression of anti-inflammatory proteins and represses the
expression of pro-inflammatory proteins 146 . Consequently, GCs have potent anti-inflammatory
and immunosuppressive properties and are thus part of the feedback mechanism in the immune
system that decreases the inflammation 147,148 . Therefore, GCs are used as drugs to treat
inflammatory conditions such as arthritis (e.g. rheumatoid arthritis), dermatitis, and diseases that
are caused by an overactive immune system (allergies, asthma, and autoimmune diseases) (Table
1) 146,148 . GCs are also administered as post-transplant immunosuppressants to prevent any acute
transplant rejection and graft-versus-host disease. However, they do not prevent an infection and
they also inhibit later reparative processes 146 .
Treatment with GCs has many diverse effects, including potentially harmful side
effects (Table 1) 146 . Excessive GC levels resulting from administration as a drug affect many
systems, including inhibition of bone formation, suppression of calcium absorption and delayed
wound healing 148 . With respect to OA, intra-articular injections of GCs appear to have a reducing
effect on inflammation and pain 149 . Unfortunately, prolonged GC therapy results in GC-induced
osteoporosis 150 . Thus, to treat OA, GCs without detrimental effects on bone are needed.
The major endogenous GC in humans is cortisol, which was first used
therapeutically for rheumatoid arthritis by Hench and co-workers in 1949 151 . Since then, a large
number of synthetic compounds with GC activity have been developed for therapeutic use. They
differ in pharmacokinetics and pharmacodynamics 152 . The potencies and effects of some of the
synthetic glucocorticoids used for treatment are described in Table 1.
Table 1. Synthetic glucocorticoids and their effects/side-effects. Most of the synthetic glucocorticoids have
the same general effect (anti-inflammatory and immunosuppressive effects), but the potency and specificity for
certain diseases vary 153,152 .
33

CHAPTER 2: Introduction
Hopefully, future studies may better define the molecular mechanisms of GC actions, thus
triggering the discovery of new GC formulations providing a better pharmacokinetic profile
and fewer adverse effects.
As GCs may be a potential treatment for
OA, the effects of prednisolone and
dexamethasone on bone and cartilage cells
are investigated in PAPER II
34

2.5 References for CHAPTER 2
1 WebMD. Better information. Better health.
http://www.webmd.com/rheumatoidarthritis/guide/most-common-arthritis-types.
2011;
2 Gigtforeningen.
www.gigtforeningen.dk/viden+om+gigt/diagnoser/slidgi
gt. 2011;
3 Wu C.W. and Kalunian K.C. New developments in
osteoarthritis. Clin Geriatr Med. 2005;21(3):589-601, vii.
4 Lawrence R.C., Helmick C.G., Arnett F.C., Deyo R.A.,
Felson D.T., Giannini E.H., Heyse S.P. et al. Estimates of
the prevalence of arthritis and selected musculoskeletal
disorders in the United States. Arthritis Rheum.
1998;41(5):778-799.
5 Moskowitz R.W., Altman R.D., Hochberg M.C.,
Buckwalter J.A., and Goldberg V.M. Chapter 3.
Osteoarthritis; Diagnosis and Medical/surgical
Management. 2007;4th(3)
6 Cooper C., Cook P.L., Osmond C., Fisher L., and Cawley
M.I. Osteoarthritis of the hip and osteoporosis of the
proximal femur. Ann Rheum Dis. 1991;50(8):540-542.
7 Foss M.V. and Byers P.D. Bone density, osteoarthrosis of
the hip, and fracture of the upper end of the femur. Ann
Rheum Dis. 1972;31(4):259-264.
8 Hilal G., Martel-Pelletier J., Pelletier J.P., Ranger P., and
Lajeunesse D. Osteoblast-like cells from human
subchondral osteoarthritic bone demonstrate an altered
phenotype in vitro: possible role in subchondral bone
sclerosis. Arthritis Rheum. 1998;41(5):891-899.
9 Dequeker J., Aerssens J., and Luyten F.P. Osteoarthritis
and osteoporosis: clinical and research evidence of inverse
relationship. Aging Clin Exp Res. 2003;15(5):426-439.
10 Brandt K.D., Doherty M., and Lomander L.S.
Osteoarthritis. 2003;
11 Goldring M.B. Update on the biology of the chondrocyte
and new approaches to treating cartilage diseases. Best
Pract Res Clin Rheumatol. 2006;20(5):1003-1025.
12 Felson D.T., Lawrence R.C., Dieppe P.A., Hirsch R.,
Helmick C.G., Jordan J.M., Kington R.S. et al.
Osteoarthritis: new insights. Part 1: the disease and its risk
factors. Ann Intern Med. 2000;133(8):635-646.
13 Worrall J.G. Understanding Arthritis and Rheumatism.
2005;
CHAPTER 2: Introduction
35
14 Rowan A.D. Cartilage catabolism in arthritis: factors that
influence homeostasis. Expert Rev Mol Med. 2001;2001:1-
20.
15 Burr D.B. Targeted and nontargeted remodeling. Bone.
2002;30(1):2-4.
16 Baron R, Olsen B.J, Lian J.B, Canalis E., Monroe D.G.,
Robey P.G., Mundy G.R. et al. Anatomy and Biology of
Bone Matrix and Cellular Elements. Primer on the
Metabolic Bone Diseases and Disorders of Mineral
Metabolism. 2003;5:1-58.
17 Clarke B. Normal bone anatomy and physiology. Clin J
Am Soc Nephrol. 2008;3 Suppl 3:S131-S139.
18 Erlebacher A., Filvaroff E.H., Gitelman S.E., and Derynck
R. Toward a molecular understanding of skeletal
development. Cell. 1995;80(3):371-378.
19 Keraplis A.C. Principles of bone biology. 2002;33-58.
20 Provot S. and Schipani E. Molecular mechanisms of
endochondral bone development. Biochem Biophys Res
Commun. 2005;328(3):658-665.
21 Seeman E. and Delmas P.D. Bone quality--the material
and structural basis of bone strength and fragility. N Engl
J Med. 2006;354(21):2250-2261.
22 Robey P.G. and Bodkey A.L. Extracellular matrix and
biomineralization of bone. Primer on the metabolic bone
diseases and disorders of mineral metabolism.
2003;5(6):38-46.
23 Baron R. General principles of Bone Biology. Primer on
the metabolic bone diseases and disorders of mineral
metabolism. 2003;(1):1-8.
24 Matos M.A., Araujo F.P., and Paixao F.B.
Histomorphometric evaluation of bone healing in rabbit
fibular osteotomy model without fixation. J Orthop Surg
Res. 2008;3:4-
25 Matsuo K. Cross-talk among bone cells. Curr Opin
Nephrol Hypertens. 2009;18(4):292-297.
26 Matsuo K. and Irie N. Osteoclast-osteoblast
communication. Arch Biochem Biophys. 2008;473(2):201-
209.
27 Henriksen K., Neutzsky-Wulff A.V., Bonewald L.F., and
Karsdal M.A. Local communication on and within bone
controls bone remodeling. Bone. 2009;44(6):1026-1033.

28 Iqbal J., Sun L., and Zaidi M. Coupling bone degradation
to formation. Nat Med. 2009;15(7):729-731.
29 Knothe Tate M.L., Adamson J.R., Tami A.E., and Bauer
T.W. The osteocyte. Int J Biochem Cell Biol. 2004;36(1):1-
8.
30 Aubin J.E. Advances in the osteoblast lineage. Biochem
Cell Biol. 1998;76(6):899-910.
31 Gori F., Hofbauer L.C., Dunstan C.R., Spelsberg T.C.,
Khosla S., and Riggs B.L. The expression of
osteoprotegerin and RANK ligand and the support of
osteoclast formation by stromal-osteoblast lineage cells is
developmentally regulated. Endocrinology.
2000;141(12):4768-4776.
32 Lerner U.H. NEW MOLECULES IN THE TUMOR
NECROSIS FACTOR LIGAND AND RECEPTOR
SUPERFAMILIES WITH IMPORTANCE FOR
PHYSIOLOGICAL AND PATHOLOGICAL BONE
RESORPTION. Crit Rev Oral Biol Med. 2004;15(2):64-
81.
33 Thomas G.P., Baker S.U., Eisman J.A., and Gardiner E.M.
Changing RANKL/OPG mRNA expression in
differentiating murine primary osteoblasts. J Endocrinol.
2001;170(2):451-460.
34 Teitelbaum S.L. Bone resorption by osteoclasts. Science.
2000;289(5484):1504-1508.
35 Martin T., Gooi J.H., and Sims N.A. Molecular
mechanisms in coupling of bone formation to resorption.
Crit Rev Eukaryot Gene Expr. 2009;19(1):73-88.
36 Sims N.A. and Gooi J.H. Bone remodeling: Multiple
cellular interactions required for coupling of bone
formation and resorption. Semin Cell Dev Biol.
2008;19(5):444-451.
37 Karsdal M.A., Henriksen K., Sorensen M.G., Gram J.,
Schaller S., Dziegiel M.H., Heegaard A.M. et al.
Acidification of the osteoclastic resorption compartment
provides insight into the coupling of bone formation to
bone resorption. Am J Pathol. 2005;166(2):467-476.
38 Martin T.J. and Sims N.A. Osteoclast-derived activity in
the coupling of bone formation to resorption. Trends Mol
Med. 2005;11(2):76-81.
39 Mundy G.R., Chen D., and Oyajobi B.O. Bone
remodeling. Primer on the metabolic bone diseases and
disorders of mineral metabolism. 2003;5(7):46-58.
CHAPTER 2: Introduction
36
40 Andersen T.L., Sondergaard T.E., Skorzynska K.E., gnaes-
Hansen F., Plesner T.L., Hauge E.M., Plesner T. et al. A
physical mechanism for coupling bone resorption and
formation in adult human bone. Am J Pathol.
2009;174(1):239-247.
41 Eriksen E.F., Eghbali-Fatourechi G.Z., and Khosla S.
Remodeling and vascular spaces in bone. J Bone Miner
Res. 2007;22(1):1-6.
42 Hauge E.M., Qvesel D., Eriksen E.F., Mosekilde L., and
Melsen F. Cancellous bone remodeling occurs in
specialized compartments lined by cells expressing
osteoblastic markers. J Bone Miner Res. 2001;16(9):1575-
1582.
43 Kaunitz J.D. and Yamaguchi D.T. TNAP, TrAP, ecto-
purinergic signaling, and bone remodeling. J Cell Biochem.
2008;105(3):655-662.
44 Karsdal M.A., Martin T.J., Bollerslev J., Christiansen C.,
and Henriksen K. Are nonresorbing osteoclasts sources of
bone anabolic activity? J Bone Miner Res. 2007;22(4):487-
494.
45 Karsdal M.A., Neutzsky-Wulff A.V., Dziegiel M.H.,
Christiansen C., and Henriksen K. Osteoclasts secrete
non-bone derived signals that induce bone formation.
Biochem Biophys Res Commun. 2008;366(2):483-488.
46 Zhao C., Irie N., Takada Y., Shimoda K., Miyamoto T.,
Nishiwaki T., Suda T. et al. Bidirectional ephrinB2-
EphB4 signaling controls bone homeostasis. Cell Metab.
2006;4(2):111-121.
47 Aubin J.E. and Bonnelye E. Osteoprotegerin and its
ligand: a new paradigm for regulation of
osteoclastogenesis and bone resorption. Osteoporos Int.
2000;11(11):905-913.
48 Kwan T.S., Lajeunesse D., Pelletier J.P., and Martel-
Pelletier J. Targeting subchondral bone for treating
osteoarthritis: what is the evidence? Best Pract Res Clin
Rheumatol. 2010;24(1):51-70.
49 Mankin H.J., Brandt K.D., and Shulman L.E. Workshop
on Etiopathogenesis of Osteoarthritis. The Journal of
Rheumatology. 1986;(13):1130-1160.
50 Lefebvre V. and de C.B. Toward understanding SOX9
function in chondrocyte differentiation. Matrix Biol.
1998;16(9):529-540.

51 Lorenz H. and Richter W. Osteoarthritis: cellular and
molecular changes in degenerating cartilage. Prog
Histochem Cytochem. 2006;40(3):135-163.
52 Wollheim F.A. Serum markers of articular cartilage
damage and repair. Rheum Dis Clin North Am.
1999;25(2):417-32, viii.
53 Wotton S.F. and Duance V.C. Type III collagen in normal
human articular cartilage. Histochem J. 1994;26(5):412-
416.
54 Charni-Ben T.N. and Garnero P. Monitoring cartilage
turnover. Curr Rheumatol Rep. 2007;9(1):16-24.
55 Hardingham T.E. and Fosang A.J. Proteoglycans: many
forms and many functions. FASEB J. 1992;6(3):861-870.
56 Matyas J.R., Adams M.E., Huang D., and Sandell L.J.
Discoordinate gene expression of aggrecan and type II
collagen in experimental osteoarthritis. Arthritis Rheum.
1995;38(3):420-425.
57 Darling E.M. and Athanasiou K.A. Rapid phenotypic
changes in passaged articular chondrocyte subpopulations.
J Orthop Res. 2005;23(2):425-432.
58 Fox A.J.S, Bedi A., and Rodeo S.A. The Basic Science of
Articular Cartilage: Structure, Composition, and Function.
Sports Health: A Multidisciplinary Approach.
2009;1(6):461-461.
59 Sah R.L, Wong B.L., Li K.W., Bae W.C, and Price J.H.
Three dimentional histomorphometric analysis of normal
adult human articular cartilage. 48th AMORS. 2001;
60 Cao Y., Vacanti J.P., Paige K.T., Upton J., and Vacanti
C.A. Transplantation of chondrocytes utilizing a polymercell
construct to produce tissue-engineered cartilage in the
shape of a human ear. Plast Reconstr Surg.
1997;100(2):297-302.
61 Poole A.R. An introduction to the pathophysiology of
osteoarthritis. Front Biosci. 1999;4:D662-D670.
62 Hayami T., Pickarski M., Wesolowski G.A., McLane J.,
Bone A., Destefano J., Rodan G.A. et al. The role of
subchondral bone remodeling in osteoarthritis: reduction
of cartilage degeneration and prevention of osteophyte
formation by alendronate in the rat anterior cruciate
ligament transection model. Arthritis Rheum.
2004;50(4):1193-1206.
63 Casden A.M., Jaffe F.F., Kastenbaum D.M., and Bonar
S.F. Osteoarthritis associated with osteopetrosis treated by
total knee arthroplasty. Report of a case. Clin Orthop
Relat Res. 1989;(247):202-207.
CHAPTER 2: Introduction
37
64 www.zimmer.co.uk. 2011;
65 Bay-Jensen A.C., Hoegh-Madsen S., Dam E., Henriksen
K., Sondergaard B.C., Pastoureau P., Qvist P. et al. Which
elements are involved in reversible and irreversible
cartilage degradation in osteoarthritis? Rheumatol Int.
2010;30(4):435-442.
66 Vignon E., Arlot M., Patricot L.M., and Vignon G. The
cell density of human femoral head cartilage. Clin Orthop
Relat Res. 1976;(121):303-308.
67 Horton W.E., Jr., Feng L., and Adams C. Chondrocyte
apoptosis in development, aging and disease. Matrix Biol.
1998;17(2):107-115.
68 Hashimoto S., Takahashi K., Amiel D., Coutts R.D., and
Lotz M. Chondrocyte apoptosis and nitric oxide
production during experimentally induced osteoarthritis.
Arthritis Rheum. 1998;41(7):1266-1274.
69 Guerne P.A., Blanco F., Kaelin A., Desgeorges A., and
Lotz M. Growth factor responsiveness of human articular
chondrocytes in aging and development. Arthritis Rheum.
1995;38(7):960-968.
70 Martin J.A. and Buckwalter J.A. The role of chondrocyte
senescence in the pathogenesis of osteoarthritis and in
limiting cartilage repair. J Bone Joint Surg Am. 2003;85-A
Suppl 2:106-110.
71 Poole C.A., Ayad S., and Schofield J.R. Chondrons from
articular cartilage: I. Immunolocalization of type VI
collagen in the pericellular capsule of isolated canine tibial
chondrons. J Cell Sci. 1988;90 ( Pt 4):635-643.
72 Pfander D., Swoboda B., and Kirsch T. Expression of
early and late differentiation markers (proliferating cell
nuclear antigen, syndecan-3, annexin VI, and alkaline
phosphatase) by human osteoarthritic chondrocytes. Am J
Pathol. 2001;159(5):1777-1783.
73 Aigner T., Kurz B., Fukui N., and Sandell L. Roles of
chondrocytes in the pathogenesis of osteoarthritis. Curr
Opin Rheumatol. 2002;14(5):578-584.
74 Valcourt U., Gouttenoire J., ubert-Foucher E., Herbage
D., and Mallein-Gerin F. Alternative splicing of type II
procollagen pre-mRNA in chondrocytes is oppositely
regulated by BMP-2 and TGF-beta1. FEBS Lett.
2003;545(2-3):115-119.
75 Kielty C.M., Kwan A.P., Holmes D.F., Schor S.L., and
Grant M.E. Type X collagen, a product of hypertrophic
chondrocytes. Biochem J. 1985;227(2):545-554.

76 Benya P.D. and Shaffer J.D. Dedifferentiated
chondrocytes reexpress the differentiated collagen
phenotype when cultured in agarose gels. Cell.
1982;30(1):215-224.
77 Elima K. and Vuorio E. Expression of mRNAs for
collagens and other matrix components in
dedifferentiating and redifferentiating human
chondrocytes in culture. FEBS Lett. 1989;258(2):195-198.
78 Aigner T., Zhu Y., Chansky H.H., Matsen F.A., III,
Maloney W.J., and Sandell L.J. Reexpression of type IIA
procollagen by adult articular chondrocytes in
osteoarthritic cartilage. Arthritis Rheum. 1999;42(7):1443-
1450.
79 von der M.K., Kirsch T., Nerlich A., Kuss A., Weseloh G.,
Gluckert K., and Stoss H. Type X collagen synthesis in
human osteoarthritic cartilage. Indication of chondrocyte
hypertrophy. Arthritis Rheum. 1992;35(7):806-811.
80 Sandell L.J. and Aigner T. Articular cartilage and changes
in arthritis. An introduction: cell biology of osteoarthritis.
Arthritis Res. 2001;3(2):107-113.
81 Tchetina E.V., Squires G., and Poole A.R. Increased type
II collagen degradation and very early focal cartilage
degeneration is associated with upregulation of
chondrocyte differentiation related genes in early human
articular cartilage lesions. J Rheumatol. 2005;32(5):876-
886.
82 Goldring M.B. and Goldring S.R. Osteoarthritis. J Cell
Physiol. 2007;213(3):626-634.
83 Aigner T., Gluckert K., and von der M.K. Activation of
fibrillar collagen synthesis and phenotypic modulation of
chondrocytes in early human osteoarthritic cartilage
lesions. Osteoarthritis Cartilage. 1997;5(3):183-189.
84 Lohmander L.S., Ionescu M., Jugessur H., and Poole A.R.
Changes in joint cartilage aggrecan after knee injury and in
osteoarthritis. Arthritis Rheum. 1999;42(3):534-544.
85 Bollet A.J. and Nance J.L. Biochemical Findings in
Normal and Osteoarthritic Articular Cartilage. II.
Chondroitin Sulfate Concentration and Chain Length,
Water, and Ash Content. J Clin Invest. 1966;45(7):1170-
1177.
86 Goldring M.B. The role of the chondrocyte in
osteoarthritis. Arthritis Rheum. 2000;43(9):1916-1926.
87 Goldring M.B. Osteoarthritis and cartilage: the role of
cytokines. Curr Rheumatol Rep. 2000;2(6):459-465.
CHAPTER 2: Introduction
38
88 Bramono D.S., Richmond J.C., Weitzel P.P., Kaplan D.L.,
and Altman G.H. Matrix metalloproteinases and their
clinical applications in orthopaedics. Clin Orthop Relat
Res. 2004;(428):272-285.
89 Cawston T.E. and Young D.A. Proteinases involved in
matrix turnover during cartilage and bone breakdown. Cell
Tissue Res. 2010;339(1):221-235.
90 Hollander A.P., Pidoux I., Reiner A., Rorabeck C., Bourne
R., and Poole A.R. Damage to type II collagen in aging
and osteoarthritis starts at the articular surface, originates
around chondrocytes, and extends into the cartilage with
progressive degeneration. J Clin Invest. 1995;96(6):2859-
2869.
91 Shibakawa A., Yudoh K., Masuko-Hongo K., Kato T.,
Nishioka K., and Nakamura H. The role of subchondral
bone resorption pits in osteoarthritis: MMP production by
cells derived from bone marrow. Osteoarthritis Cartilage.
2005;13(8):679-687.
92 Bluteau G., Conrozier T., Mathieu P., Vignon E., Herbage
D., and Mallein-Gerin F. Matrix metalloproteinase-1, -3, -
13 and aggrecanase-1 and -2 are differentially expressed in
experimental osteoarthritis. Biochim Biophys Acta.
2001;1526(2):147-158.
93 Freemont A.J., Hampson V., Tilman R., Goupille P.,
Taiwo Y., and Hoyland J.A. Gene expression of matrix
metalloproteinases 1, 3, and 9 by chondrocytes in
osteoarthritic human knee articular cartilage is zone and
grade specific. Ann Rheum Dis. 1997;56(9):542-549.
94 Hanemaaijer R., Sorsa T., Konttinen Y.T., Ding Y.,
Sutinen M., Visser H., van H., V et al. Matrix
metalloproteinase-8 is expressed in rheumatoid synovial
fibroblasts and endothelial cells. Regulation by tumor
necrosis factor-alpha and doxycycline. J Biol Chem.
1997;272(50):31504-31509.
95 Salminen H.J., Saamanen A.M., Vankemmelbeke M.N.,
Auho P.K., Perala M.P., and Vuorio E.I. Differential
expression patterns of matrix metalloproteinases and their
inhibitors during development of osteoarthritis in a
transgenic mouse model. Ann Rheum Dis. 2002;61(7):591-
597.
96 Cole A.A. and Kuettner K.E. MMP-8 (neutrophil
collagenase) mRNA and aggrecanase cleavage products
are present in normal and osteoarthritic human articular
cartilage. Acta Orthop Scand Suppl. 1995;266:98-102.

97 Nakamura H., Tanaka M., Masuko-Hongo K., Yudoh K.,
Kato T., Beppu M., and Nishioka K. Enhanced
production of MMP-1, MMP-3, MMP-13, and RANTES
by interaction of chondrocytes with autologous T cells.
Rheumatol Int. 2006;26(11):984-990.
98 Neame P.J. and Sandy J.D. Cartilage aggrecan.
Biosynthesis, degradation and osteoarthritis. J Fla Med
Assoc. 1994;81(3):191-193.
99 Smith R.L. Degradative enzymes in osteoarthritis. Front
Biosci. 1999;4:D704-D712.
100 Blanco F.J., Guitian R., Moreno J., de Toro F.J., and
Galdo F. Effect of antiinflammatory drugs on COX-1 and
COX-2 activity in human articular chondrocytes. J
Rheumatol. 1999;26(6):1366-1373.
101 Hui W., Rowan A.D., and Cawston T. Modulation of the
expression of matrix metalloproteinase and tissue
inhibitors of metalloproteinases by TGF-beta1 and IGF-1
in primary human articular and bovine nasal chondrocytes
stimulated with TNF-alpha. Cytokine. 2001;16(1):31-35.
102 Morko J.P., Soderstrom M., Saamanen A.M., Salminen
H.J., and Vuorio E.I. Up regulation of cathepsin K
expression in articular chondrocytes in a transgenic mouse
model for osteoarthritis. Ann Rheum Dis. 2004;63(6):649-
655.
103 Inaoka T., Bilbe G., Ishibashi O., Tezuka K., Kumegawa
M., and Kokubo T. Molecular cloning of human cDNA
for cathepsin K: novel cysteine proteinase predominantly
expressed in bone. Biochem Biophys Res Commun.
1995;206(1):89-96.
104 Dodds R.A., Connor J.R., Drake F., Feild J., and Gowen
M. Cathepsin K mRNA detection is restricted to
osteoclasts during fetal mouse development. J Bone Miner
Res. 1998;13(4):673-682.
105 Kafienah W., Bromme D., Buttle D.J., Croucher L.J., and
Hollander A.P. Human cathepsin K cleaves native type I
and II collagens at the N-terminal end of the triple helix.
Biochem J. 1998;331 ( Pt 3):727-732.
106 Maciewicz R.A., Wotton S.F., Etherington D.J., and
Duance V.C. Susceptibility of the cartilage collagens types
II, IX and XI to degradation by the cysteine proteinases,
cathepsins B and L. FEBS Lett. 1990;269(1):189-193.
107 McDougall J.J., Schuelert N., and Bowyer J. Cathepsin K
inhibition reduces CTXII levels and joint pain in the
guinea pig model of spontaneous osteoarthritis.
Osteoarthritis Cartilage. 2010;18(10):1355-1357.
CHAPTER 2: Introduction
39
108 Uusitalo H., Hiltunen A., Soderstrom M., Aro H.T., and
Vuorio E. Expression of cathepsins B, H, K, L, and S and
matrix metalloproteinases 9 and 13 during chondrocyte
hypertrophy and endochondral ossification in mouse
fracture callus. Calcif Tissue Int. 2000;67(5):382-390.
109 Knauper V., Will H., Lopez-Otin C., Smith B., Atkinson
S.J., Stanton H., Hembry R.M. et al. Cellular mechanisms
for human procollagenase-3 (MMP-13) activation.
Evidence that MT1-MMP (MMP-14) and gelatinase a
(MMP-2) are able to generate active enzyme. J Biol Chem.
1996;271(29):17124-17131.
110 Knauper V., Bailey L., Worley J.R., Soloway P., Patterson
M.L., and Murphy G. Cellular activation of proMMP-13
by MT1-MMP depends on the C-terminal domain of
MMP-13. FEBS Lett. 2002;532(1-2):127-130.
111 Knauper V., Murphy G., and Tschesche H. Activation of
human neutrophil procollagenase by stromelysin 2. Eur J
Biochem. 1996;235(1-2):187-191.
112 Mansell J.P. and Bailey A.J. Abnormal cancellous bone
collagen metabolism in osteoarthritis. J Clin Invest.
1998;101(8):1596-1603.
113 Mansell J.P., Tarlton J.F., and Bailey A.J. Biochemical
evidence for altered subchondral bone collagen
metabolism in osteoarthritis of the hip. Br J Rheumatol.
1997;36(1):16-19.
114 Hulejova H., Baresova V., Klezl Z., Polanska M., Adam
M., and Senolt L. Increased level of cytokines and matrix
metalloproteinases in osteoarthritic subchondral bone.
Cytokine. 2007;38(3):151-156.
115 Grynpas M.D., Alpert B., Katz I., Lieberman I., and
Pritzker K.P. Subchondral bone in osteoarthritis. Calcif
Tissue Int. 1991;49(1):20-26.
116 Radin E.L. and Rose R.M. Role of subchondral bone in
the initiation and progression of cartilage damage. Clin
Orthop Relat Res. 1986;(213):34-40.
117 Carlson C.S., Loeser R.F., Purser C.B., Gardin J.F., and
Jerome C.P. Osteoarthritis in cynomolgus macaques. III:
Effects of age, gender, and subchondral bone thickness on
the severity of disease. J Bone Miner Res. 1996;11(9):1209-
1217.
118 Dieppe P., Cushnaghan J., Young P., and Kirwan J.
Prediction of the progression of joint space narrowing in
osteoarthritis of the knee by bone scintigraphy. Ann
Rheum Dis. 1993;52(8):557-563.

119 Menkes C.J. Radiographic criteria for classification of
osteoarthritis. J Rheumatol Suppl. 1991;27:13-15.
120 Rousseau J.C. and Delmas P.D. Biological markers in
osteoarthritis. Nat Clin Pract Rheumatol. 2007;3(6):346-
356.
121 Fosang A.J., Stanton H., Little C.B., and Atley L.M.
Neoepitopes as biomarkers of cartilage catabolism.
Inflamm Res. 2003;52(7):277-282.
122 Karsdal M.A., Henriksen K., Leeming D.J., Woodworth
T., Vassiliadis E., and Bay-Jensen A.C. Novel
combinations of Post-Translational Modification (PTM)
neo-epitopes provide tissue-specific biochemical markers--
are they the cause or the consequence of the disease? Clin
Biochem. 2010;43(10-11):793-804.
123 Price P.A., Parthemore J.G., and Deftos L.J. New
biochemical marker for bone metabolism. Measurement
by radioimmunoassay of bone GLA protein in the plasma
of normal subjects and patients with bone disease. J Clin
Invest. 1980;66(5):878-883.
124 Mouritzen U., Christgau S., Lehmann H.J., Tanko L.B.,
and Christiansen C. Cartilage turnover assessed with a
newly developed assay measuring collagen type II
degradation products: influence of age, sex, menopause,
hormone replacement therapy, and body mass index. Ann
Rheum Dis. 2003;62(4):332-336.
125 Bonde M., Qvist P., Fledelius C., Riis B.J., and
Christiansen C. Immunoassay for quantifying type I
collagen degradation products in urine evaluated. Clin
Chem. 1994;40(11 Pt 1):2022-2025.
126 Rosenquist C., Fledelius C., Christgau S., Pedersen B.J.,
Bonde M., Qvist P., and Christiansen C. Serum CrossLaps
One Step ELISA. First application of monoclonal
antibodies for measurement in serum of bone-related
degradation products from C-terminal telopeptides of type
I collagen. Clin Chem. 1998;44(11):2281-2289.
127 Schaller S., Henriksen K., Hoegh-Andersen P.,
Sondergaard B.C., Sumer E.U., Tanko L.B., Qvist P. et al.
In vitro, ex vivo, and in vivo methodological approaches
for studying therapeutic targets of osteoporosis and
degenerative joint diseases: how biomarkers can assist?
Assay Drug Dev Technol. 2005;3(5):553-580.
128 Garnero P., Rousseau J.C., and Delmas P.D. Molecular
basis and clinical use of biochemical markers of bone,
cartilage, and synovium in joint diseases. Arthritis Rheum.
2000;43(5):953-968.
CHAPTER 2: Introduction
40
129 Garnero P., Piperno M., Gineyts E., Christgau S., Delmas
P.D., and Vignon E. Cross sectional evaluation of
biochemical markers of bone, cartilage, and synovial tissue
metabolism in patients with knee osteoarthritis: relations
with disease activity and joint damage. Ann Rheum Dis.
2001;60(6):619-626.
130 Bondeson J., Wainwright S., Hughes C., and Caterson B.
The regulation of the ADAMTS4 and ADAMTS5
aggrecanases in osteoarthritis: a review. Clin Exp
Rheumatol. 2008;26(1):139-145.
131 Zhen E.Y., Brittain I.J., Laska D.A., Mitchell P.G., Sumer
E.U., Karsdal M.A., and Duffin K.L. Characterization of
metalloprotease cleavage products of human articular
cartilage. Arthritis Rheum. 2008;58(8):2420-2431.
132 Bay-Jensen A.C., Sondergaard B.C., Christiansen C.,
Karsdal M.A., Madsen S.H., and Qvist P. Biochemical
markers of joint tissue turnover. Assay Drug Dev
Technol. 2010;8(1):118-124.
133 Qvist P., Christiansen C., Karsdal M.A., Madsen S.H.,
Sondergaard B.C., and Bay-Jensen A.C. Application of
biochemical markers in development of drugs for
treatment of osteoarthritis. Biomarkers. 2010;15(1):1-19.
134 Christgau S., Garnero P., Fledelius C., Moniz C., Ensig M.,
Gineyts E., Rosenquist C. et al. Collagen type II C-
telopeptide fragments as an index of cartilage degradation.
Bone. 2001;29(3):209-215.
135 Bay-Jensen A.C., Liu Q., Byrjalsen I., Li Y., Wang J.,
Pedersen C., Leeming D.J. et al. Enzyme-linked
immunosorbent assay (ELISAs) for metalloproteinase
derived type II collagen neoepitope, CIIM--increased
serum CIIM in subjects with severe radiographic
osteoarthritis. Clin Biochem. 2011;44(5-6):423-429.
136 Garnero P., Charni N., Juillet F., Conrozier T., and
Vignon E. Increased urinary type II collagen helical and C
telopeptide levels are independently associated with a
rapidly destructive hip osteoarthritis. Ann Rheum Dis.
2006;65(12):1639-1644.
137 Sondergaard B.C., Henriksen K., Wulf H., Oestergaard S.,
Schurigt U., Brauer R., Danielsen I. et al. Relative
contribution of matrix metalloprotease and cysteine
protease activities to cytokine-stimulated articular cartilage
degradation. Osteoarthritis Cartilage. 2006;14(8):738-748.
138 Sharif M., Kirwan J., Charni N., Sandell L.J., Whittles C.,
and Garnero P. A 5-yr longitudinal study of type IIA
collagen synthesis and total type II collagen degradation in
patients with knee osteoarthritis--association with disease
progression. Rheumatology (Oxford). 2007;46(6):938-943.

139 Garnero P., Ayral X., Rousseau J.C., Christgau S., Sandell
L.J., Dougados M., and Delmas P.D. Uncoupling of type
II collagen synthesis and degradation predicts progression
of joint damage in patients with knee osteoarthritis.
Arthritis Rheum. 2002;46(10):2613-2624.
140 Aigner T., Stoss H., Weseloh G., Zeiler G., and von der
M.K. Activation of collagen type II expression in
osteoarthritic and rheumatoid cartilage. Virchows Arch B
Cell Pathol Incl Mol Pathol. 1992;62(6):337-345.
141 Nemirovskiy O.V., Sunyer T., Aggarwal P., Abrams M.,
Hellio Le Graverand M.P., and Mathews W.R. Discovery
and development of the N-terminal procollagen type II
(NPII) biomarker: a tool for measuring collagen type II
synthesis. Osteoarthritis Cartilage. 2008;16(12):1494-1500.
142 Rousseau J.C., Zhu Y., Miossec P., Vignon E., Sandell L.J.,
Garnero P., and Delmas P.D. Serum levels of type IIA
procollagen amino terminal propeptide (PIIANP) are
decreased in patients with knee osteoarthritis and
rheumatoid arthritis. Osteoarthritis Cartilage.
2004;12(6):440-447.
143 Caterson B., Flannery C.R., Hughes C.E., and Little C.B.
Mechanisms involved in cartilage proteoglycan catabolism.
Matrix Biol. 2000;19(4):333-344.
144 Sumer E.U., Qvist P., and Tanko L.B. Matrix
Metalloproteinase and Aggrecanase Generated Aggrecan
Fragments: Implications for the Diagnostics and
Therapeutics of Destructive Joint Diseases. Drug
development research. 2007;68(1):1-13.
145 Sumer E.U., Sondergaard B.C., Rousseau J.C., Delmas
P.D., Fosang A.J., Karsdal M.A., Christiansen C. et al.
MMP and non-MMP-mediated release of aggrecan and its
fragments from articular cartilage: a comparative study of
three different aggrecan and glycosaminoglycan assays.
Osteoarthritis Cartilage. 2007;15(2):212-221.
CHAPTER 2: Introduction
41
146 Glucocorticoid - wikipedia.
http://en.wikipedia.org/wiki/Glucocorticoid. 2011;
147 Rhen T. and Cidlowski J.A. Antiinflammatory action of
glucocorticoids--new mechanisms for old drugs. N Engl J
Med. 2005;353(16):1711-1723.
148 www.vivo.colostate.edu. 2011;
149 Hepper C.T., Halvorson J.J., Duncan S.T., Gregory A.J.,
Dunn W.R., and Spindler K.P. The efficacy and duration
of intra-articular corticosteroid injection for knee
osteoarthritis: a systematic review of level I studies. J Am
Acad Orthop Surg. 2009;17(10):638-646.
150 Eastell R., Reid D.M., Compston J., Cooper C., Fogelman
I., Francis R.M., Hosking D.J. et al. A UK Consensus
Group on management of glucocorticoid-induced
osteoporosis: an update. J Intern Med. 1998;244(4):271-
292.
151 HENCH P.S., SLOCUMB C.H., BARNES A.R., SMITH
H.C., POLLEY H.F., and KENDALL E.C. The effects of
the adrenal cortical hormone 17-hydroxy-11dehydrocorticosterone
(Compound E) on the acute phase
of rheumatic fever; preliminary report. Mayo Clin Proc.
1949;24(11):277-297
152 www.endotext.org/adrenal/adrenal14/adrenal14.htm.
2011;
153 Leung D.Y., Hanifin J.M., Charlesworth E.N., Li J.T.,
Bernstein I.L., Berger W.E., Blessing-Moore J. et al.
Disease management of atopic dermatitis: a practice
parameter. Joint Task Force on Practice Parameters,
representing the American Academy of Allergy, Asthma
and Immunology, the American College of Allergy,
Asthma and Immunology, and the Joint Council of
Allergy, Asthma and Immunology. Work Group on
Atopic Dermatitis. Ann Allergy Asthma Immunol.
1997;79(3):197-211.

Suzi Høgh Madsen, PhD thesis
42

CHAPTER 3
Overview of OA models
CHAPTER 3: Overview of OA models
This Overview aims to introduce some of the different pre-clinical models used today to
investigate the pathogenesis of OA with emphasis on the ex vivo model. In brief, the complexity
and advantages/disadvantages of the different models and the motivation for developing the
femur head model (PAPER I) are described. Further, the controls used in previous ex vivo models
are evaluated in order to understand the selection of these controls for the present femur head
model (PAPER I).
3.1 Pre-clinical models
Clinical studies of OA in humans are expensive and very time consuming. Furthermore, clinical
studies do not provide information about processes and mechanisms involved in pathogenic OA-
tissues, which could be key information to prevent or stop the ongoing development of the
disease. For these and ethical reasons, pre-clinical models are commonly used to investigate the
metabolism of healthy/diseased bone or cartilage, which can be classified according to their
complexity (fig. 11):
In vitro; monolayer or pellet systems of isolated cells
Ex vivo explants cultures, in which the cells are held in their natural matrix
In vivo animal models; such as spontaneous, surgically induced, or transgenic, which resemble
the human disease (apart from the species/origin)
Each of these models has advantages and disadvantages depending on the purpose of the
investigation. In the following section, the methodology behind the three models will be described
in brief in relation to OA with main focus on the ex vivo model.
Fig. 11. The different types of pre-clinical models. In an in vitro model, cells are isolated from their natural
environments, which may cause them to change phenotype during culture. In an ex vivo model, the cells are
maintained in their natural environments, but lack systemic influence (pressure, blood supply etc.). In in vivo
models, the cells are maintained in their natural environment and host. The figure was produced by Madsen, S.H.
43

CHAPTER 3: Overview of OA models
3.2 Advantages/disadvantages of pre-clinical models
3.2.1 In vitro
Cultures of isolated chondrocytes are the easiest and most convenient way to investigate
chondrogenesis in vitro; inexpensive, easy to setup and handle. However, several limitations must
be considered. Two-dimensional monolayer cultures support the proliferation of articular
chondrocytes, but change their phenotypic appearance to de-differentiated fibroblast-like cells,
which synthesize collagen type I (a marker of osteoblasts) 1,2 . Consequently, cultures of isolated
chondrocytes are not suitable for investigating cartilage degradation or for studying mechanisms
involved in cartilage matrix assembly, but are rather used for investigating proteoglycan and
collagen biosynthesis in response to various stimuli (Table 2) 3 . Pellet culture systems were
originally described as a method for preventing the phenotypic de-differentiation of chondrocytes
in vitro 4 . This culture system allows three-dimensional cell-cell interaction and is investigated by
the same techniques as for isolated chondrocytes in monolayer, and additionally by
histology/immunohistochemistry (Table 1) 5 .
When investigating bone cells (osteoclasts and osteoblasts), the cells are not
isolated from the bone matrix, but differentiated from their progenitor cells (e.g. isolated from
blood) and cultured on e.g. bone or plastic 6 . Differently from isolated chondrocytes, osteoclasts
and osteoblasts are able to resorb and deposit bone, respectively, which is their natural task in
vivo. Thus, in vitro models for bone cells are very close to imitating in vivo conditions 6 . The
osteoclast and osteoblast cell cultures have been used to evaluate the direct effect of
glucocorticoids in this thesis (PAPER II).
3.2.2 Ex vivo
The more complicated and technically advanced ex vivo explants model offers the ability to
investigate the biology of the tissue, including both ECM and cells. This model also offers the
ability to investigate how the tissue is affected by different compounds 7 . The cells are embedded
in their natural matrix where the immediate environment is preserved, including macromolecules
and cell-binding proteins. Additionally, the explants model allows preservation of the cell
phenotype; e.g. the chondrocytes in cartilage explants maintain their spherical appearance (Table
2) 8 . Several studies have used the cartilage explants model to resemble the pathological events in
OA cartilage 9,10,11 . The pioneering work on porcine cartilage organ cultures were done by Fell et
al. in 1973 12,13 . Since then, the cartilage explants model is widely used as a degradation assay that
allows investigation of cartilage metabolism in response to different stimuli and agents 14,15,16 . A
prominent example is that of Sondergaard et al., who showed that calcitonin abrogated the CTX-
II release from catabolically induced cartilage explants 9 . Other studies investigated the proteolytic
pathways in other cartilage ex vivo studies 10,17 . The advantage of studying cartilage metabolism in
the cartilage explants model is the unique in vivo likeness. In combination with biochemical
markers, this cartilage explants model is considered to be the best non-in vivo model in the study
44

CHAPTER 3: Overview of OA models
of OA 7,18 . The biochemical markers can be used in combination with histology/
immunohistochemistry to further characterize the pathogenesis of the cartilage 14 . The articular
cartilage explants model has been used in PAPER II, III, and IV to evaluate ECM turnover and
enzymatic activity.
There is an urgent need for the discovery and development of new treatments for
OA. A fundamental step in discovering new therapies for OA is to critically evaluate our
understanding of the etiology of the disease, including the complex relationship between cartilage
and bone. Thus, the cartilage explants model is not complete as it only comprises cartilage. When
the present PhD-project was initiated there were neither biological models nor tools that allowed
the investigation of the interaction between the bone and cartilage cells. Thus, there was a strong
need for an ex vivo model system comprising both bone and cartilage, which could provide a
unique opportunity to investigate cell-cell interactions and test novel drug candidates for OA.
This resulted in testing murine femur heads as potential candidates to constitute a novel explants
model since the femur heads comprise both bone and cartilage (fig. 12). Importantly, the femur
head is a closed system, which would not be damaged during isolation procedures (except on the
shaft), keeping the homogeneity high between the samples, whereas the cartilage explants have a
low homogeneity between the samples (Table 1). Isolation as a whole unit reduces the stress on
the cells as the ECM is still intact and uncontrollable repair mechanisms are avoided.
Fig. 12. The femur head comprises both bone and cartilage. The femur head is isolated from the femur
by cutting with a scissor. The femur head is a whole unit, which is only damaged at the cutting site (at the
femur neck). The femur head comprises both bone and cartilage, but also the transition zone of
bone/cartilage (e.g. growth plate during development). The interactions between the cells from the different
compartments, primarily the chondrocytes, osteoclasts, and osteoblasts, can be investigated in this femur
head explants model. The figure was produced by Madsen, S.H.
The pathology of OA involves the whole joint. The
development of a novel ex vivo model comprising both
bone and cartilage is characterized in PAPER I
45

3.2.3 In vivo
CHAPTER 3: Overview of OA models
Animal models are commonly used for investigating pathology that resembles the human disease
of interest. These in vivo animal models can often follow preliminary in vitro or ex vivo studies (fig.
11), which need their results to be confirmed in an in situ setting 7,19 . Furthermore, the complex
changes in the joint-tissue of OA patients may take decades to develop and the need to
investigate the early changes in robust and valid in vivo models has necessitated the use of living
animals in experiments. A range of animal models of OA have been developed, such as
spontaneous, surgically induced, and transgenic animal models. These models are not perfect, as
species are not alike, but the models offer the opportunity to display many of the pathologic
features that characterize the human disease. For example, transgenic mice with mutations in
collagen type II develops OA-like changes with age 20,21 , whereas surgically removed menisci can
be observed to induce osteophyte formation, cartilage fibrillation, loss of proteoglycans, and
cellular clustering within few weeks 22,23 . Moreover, several species spontaneously develop OA,
which highly resembles human OA, although these models are relatively slow to develop the
disease 24,25 .
To summarize the advantages, animal models can be used to study pathologic
events representing early OA in human patients, and they are applicable for investigation of the
effects of a variety of agents on the progression of the disease. Furthermore, only animal models
can show altered mental behaviour or catch potential side-effect from a drug (other tissues). One
obvious disadvantage of the in vivo animal models is the potential differences in tissue response
between animals and humans. Eventually, phase-trials of a potential drug would demonstrate the
effect or lack of effect in humans. Another important disadvantage is the ethical aspect of using
living animals, which will not be discussed further in this thesis. Below is a table with the
different pre-clinical models of cartilage.
Table 2. Comparison of different cartilage pre-clinical models. The advantages and disadvantages of the in
vitro, ex vivo and in vivo models are outlined in the table, which makes it easy to compare to each other. Adapted
with modifications from Sondergaard 26 .
46

3.3 Ex vivo controls
CHAPTER 3: Overview of OA models
In order to design an ex vivo model for experimentation, it is critical to acquire appropriate
controls, which induce an anabolic or catabolic system that represents regenerative or damaged
tissue. The acknowledged cartilage explants model has established the best controls for cartilage:
1) insulin-like growth factor-I (IGF-I) as an anabolic control, and 2) a combination of tumour
necrosis factor-α (TNF-α) and oncostatin M (OSM) as a catabolic control 9,10,18 . In the early stages
of OA, an increased amount of degrading enzymes is produced by the chondrocytes 27 . When
articular cartilage explants are stimulated with TNF-α and OSM, the chondrocytes also produce
these degrading enzymes 9 . OSM only induces a mild catabolic response in chondrocytes, but it
act synergistically with TNF-α 10,28,29 . Thus, articular cartilage explants cultured in the presence of
OSM and TNF-α are shown to be a useful ex vivo model of OA 29 . Next is a short description of
the three key factors used in the cartilage explants model:
IGF-I is important in skeletal development 30 , although this factor is primarily synthesized in the
liver under the influence of growth hormones. IGF-I is essential for bone formation and it
accounts for most of the chondrocyte-stimulating activity found in serum 30 . Several reports have
described IGF-I to be the major anabolic factor of cartilage 31 . The ability of IGF-I to enhance
matrix synthesis in normal cartilage is well established in vivo and in vitro 28,32 . For example, in
cultures of human articular chondrocytes, IGF-I induces the production of GAGs and collagen
type II 31 . Besides the chondroanabolic effects, it has been shown that IGF-I opposes the
activities of catabolic cytokines in articular cartilage. Tyler showed in 1989 32 that IGF-I reduces
proteoglycan degradation, probably by inhibiting the production and release of proteinases 28,33 .
TNF-α is mainly produced by macrophages, but also by a broad variety of other cell types.
TNF-α induces apoptosis, cellular proliferation, differentiation and inflammation. TNF-α has
been shown to be an important mediator of cartilage destruction in vivo 28,32 . Besides mediating
destruction of cartilage, TNF-α also inhibits synthesis of proteoglycan and collagen type II 27 .
OSM is a member of the IL-6 family and is synthesized by stimulated T-cells and macrophages.
Besides inhibition of proteoglycan synthesis, OSM also stimulates chondrocytes to produce
proteinases, which induce proteoglycan degradation 27 . However, Wahl & Wallace indicated in
2001 that OSM is anabolic; promoting wound healing, bone formation and anti-inflammatory
effects 34 .
The controls from the cartilage explants model
were evaluated and used in the femur head
explants model (PAPER I)
47

3.4 References for CHAPTER 3
1 von der M.K., Gauss V., von der M.H., and Muller P.
Relationship between cell shape and type of collagen
synthesised as chondrocytes lose their cartilage phenotype
in culture. Nature. 1977;267(5611):531-532.
2 Lee D.A., Reisler T., and Bader D.L. Expansion of
chondrocytes for tissue engineering in alginate beads
enhances chondrocytic phenotype compared to
conventional monolayer techniques. Acta Orthop Scand.
2003;74(1):6-15.
3 Takigawa M., Okawa T., Pan H., Aoki C., Takahashi K.,
Zue J., Suzuki F. et al. Insulin-like growth factors I and II
are autocrine factors in stimulating proteoglycan synthesis,
a marker of differentiated chondrocytes, acting through
their respective receptors on a clonal human
chondrosarcoma-derived chondrocyte cell line, HCS-2/8.
Endocrinology. 1997;138(10):4390-4400.
4 Manning W.K. and Bonner W.M., Jr. Isolation and culture
of chondrocytes from human adult articular cartilage.
Arthritis Rheum. 1967;10(3):235-239.
5 Mackay A.M., Beck S.C., Murphy J.M., Barry F.P.,
Chichester C.O., and Pittenger M.F. Chondrogenic
differentiation of cultured human mesenchymal stem cells
from marrow. Tissue Eng. 1998;4(4):415-428.
6 Neutzsky-Wulff A.V., Karsdal M.A., and Henriksen K.
Characterization of the bone phenotype in ClC-7-deficient
mice. Calcif Tissue Int. 2008;83(6):425-437.
7 Schaller S., Henriksen K., Hoegh-Andersen P.,
Sondergaard B.C., Sumer E.U., Tanko L.B., Qvist P. et al.
In vitro, ex vivo, and in vivo methodological approaches
for studying therapeutic targets of osteoporosis and
degenerative joint diseases: how biomarkers can assist?
Assay Drug Dev Technol. 2005;3(5):553-580.
8 Hascall V.C., Morales T.I., Hascall G.K., Handley C.J.,
and McQuillan D.J. Biosynthesis and turnover of
proteoglycans in organ culture of bovine articular cartilage.
J Rheumatol Suppl. 1983;11:45-52.
9 Sondergaard B.C., Wulf H., Henriksen K., Schaller S.,
Oestergaard S., Qvist P., Tanko L.B. et al. Calcitonin
directly attenuates collagen type II degradation by
inhibition of matrix metalloproteinase expression and
activity in articular chondrocytes. Osteoarthritis Cartilage.
2006;14(8):759-768.
CHAPTER 3: Overview of OA models
48
10 Sondergaard B.C., Henriksen K., Wulf H., Oestergaard S.,
Schurigt U., Brauer R., Danielsen I. et al. Relative
contribution of matrix metalloprotease and cysteine
protease activities to cytokine-stimulated articular cartilage
degradation. Osteoarthritis Cartilage. 2006;14(8):738-748.
11 Charni-Ben T.N., Desmarais S., Bay-Jensen A.C., Delaisse
J.M., Percival M.D., and Garnero P. The type II collagen
fragments Helix-II and CTX-II reveal different enzymatic
pathways of human cartilage collagen degradation.
Osteoarthritis Cartilage. 2008;16(10):1183-1191.
12 Dingle J.T., Horsfield P., Fell H.B., and Barratt M.E.
Breakdown of proteoglycan and collagen induced in pig
articular cartilage in organ culture. Ann Rheum Dis.
1975;34(4):303-311.
13 Fell H.B. and Barratt M.E. The role of soft connective
tissue in the breakdown of pig articular cartilage cultivated
in the presence of complement-sufficient antiserum to pig
erythrocytes. I. Histological changes. Int Arch Allergy
Appl Immunol. 1973;44(3):441-468.
14 Roy-Beaudry M., Martel-Pelletier J., Pelletier J.P., M'Barek
K.N., Christgau S., Shipkolye F., and Moldovan F.
Endothelin 1 promotes osteoarthritic cartilage degradation
via matrix metalloprotease 1 and matrix metalloprotease
13 induction. Arthritis Rheum. 2003;48(10):2855-2864.
15 Saklatvala J. Tumour necrosis factor alpha stimulates
resorption and inhibits synthesis of proteoglycan in
cartilage. Nature. 1986;322(6079):547-549.
16 Cawston T.E., Curry V.A., Summers C.A., Clark I.M.,
Riley G.P., Life P.F., Spaull J.R. et al. The role of
oncostatin M in animal and human connective tissue
collagen turnover and its localization within the
rheumatoid joint. Arthritis Rheum. 1998;41(10):1760-
1771.
17 Sondergaard B.C., Schultz N., Madsen S.H., Bay-Jensen
A.C., Kassem M., and Karsdal M.A. MAPKs are essential
upstream signaling pathways in proteolytic cartilage
degradation--divergence in pathways leading to
aggrecanase and MMP-mediated articular cartilage
degradation. Osteoarthritis Cartilage. 2010;18(3):279-288.
18 Madsen S.H., Sondergaard B.C., Bay-Jensen A.C., and
Karsdal M.A. Cartilage formation measured by a novel
PIINP assay suggests that IGF-I does not stimulate but
maintains cartilage formation ex vivo. Scand J Rheumatol.
2009;38(3):222-226.

19 Brandt K.D. Animal models of osteoarthritis.
Biorheology. 2002;39(1-2):221-235.
20 Helminen H.J., Saamanen A.M., Salminen H., and
Hyttinen M.M. Transgenic mouse models for studying the
role of cartilage macromolecules in osteoarthritis.
Rheumatology (Oxford). 2002;41(8):848-856.
21 Glasson S.S. In vivo osteoarthritis target validation
utilizing genetically-modified mice. Curr Drug Targets.
2007;8(2):367-376.
22 Manicourt D.H., Altman R.D., Williams J.M., Devogelaer
J.P., Druetz-Van E.A., Lenz M.E., Pietryla D. et al.
Treatment with calcitonin suppresses the responses of
bone, cartilage, and synovium in the early stages of canine
experimental osteoarthritis and significantly reduces the
severity of the cartilage lesions. Arthritis Rheum.
1999;42(6):1159-1167.
23 Behets C., Williams J.M., Chappard D., Devogelaer J.P.,
and Manicourt D.H. Effects of calcitonin on subchondral
trabecular bone changes and on osteoarthritic cartilage
lesions after acute anterior cruciate ligament deficiency. J
Bone Miner Res. 2004;19(11):1821-1826.
24 Bendele A.M., White S.L., and Hulman J.F. Osteoarthrosis
in guinea pigs: histopathologic and scanning electron
microscopic features. Lab Anim Sci. 1989;39(2):115-121.
25 Jimenez P.A., Glasson S.S., Trubetskoy O.V., and Haimes
H.B. Spontaneous osteoarthritis in Dunkin Hartley guinea
pigs: histologic, radiologic, and biochemical changes. Lab
Anim Sci. 1997;47(6):598-601.
26 Sondergaard B.C. The Pharmacological Role of Calcitonin
in Osteoarthritis. PhD thesis. 2010;41-41.
CHAPTER 3: Overview of OA models
49
27 Goldring S.R. and Goldring M.B. The role of cytokines in
cartilage matrix degeneration in osteoarthritis. Clin Orthop
Relat Res. 2004;(427 Suppl):S27-S36.
28 Goldring M.B. Update on the biology of the chondrocyte
and new approaches to treating cartilage diseases. Best
Pract Res Clin Rheumatol. 2006;20(5):1003-1025.
29 Hui W., Rowan A.D., Richards C.D., and Cawston T.E.
Oncostatin M in combination with tumor necrosis factor
alpha induces cartilage damage and matrix
metalloproteinase expression in vitro and in vivo. Arthritis
Rheum. 2003;48(12):3404-3418.
30 Kronenberg H.M. Developmental regulation of the
growth plate. Nature. 2003;423(6937):332-336.
31 van Osch G.J., Mandl E.W., Marijnissen W.J., van d., V,
Verwoerd-Verhoef H.L., and Verhaar J.A. Growth factors
in cartilage tissue engineering. Biorheology. 2002;39(1-
2):215-220.
32 Tyler J.A. Insulin-like growth factor 1 can decrease
degradation and promote synthesis of proteoglycan in
cartilage exposed to cytokines. Biochem J.
1989;260(2):543-548.
33 Verschure P.J., Van Noorden C.J., Van M.J., and Van den
Berg W.B. Articular cartilage destruction in experimental
inflammatory arthritis: insulin-like growth factor-1
regulation of proteoglycan metabolism in chondrocytes.
Histochem J. 1996;28(12):835-857.
34 Wahl A.F. and Wallace P.M. Oncostatin M in the antiinflammatory
response. Ann Rheum Dis. 2001;60 Suppl
3:iii75-iii80.

Suzi Høgh Madsen, PhD thesis
50

CHAPTER 4
CHAPTER 4: PAPER I
PAPER I
___________________________________________________________________________
Introductory remarks
PAPER I validates a novel ex vivo model that allows investigation of cartilage and bone in one
closed system. The development and validation of the novel ex vivo model, included selection and
optimization processes that have not been included in PAPER I:
Selection of femur head as the explant sample (the embryonic tibia was tested first)
Sex of the mice
Time of culture period
Selection of conditioned medium
51

Original Article
Characterization of an Ex vivo Femoral
Head Model Assessed by Markers of
Bone and Cartilage Turnover
Suzi Hoegh Madsen 1 , Anne Sofie Goettrup 1 , Gedske Thomsen 1 ,
Søren Tvorup Christensen 2 , Nikolaj Schultz 1 , Kim Henriksen 3 ,
Anne-Christine Bay-Jensen 1 , and Morten Asser Karsdal 1
Abstract
Introduction
Osteoarthritis (OA) is a degenerative joint disease leading
to cartilage degradation and subchondral bone changes.
Many factors contribute to the onset of OA, including both
metabolic and biomechanical mechanisms, but it is still
debated in which tissue the disease initiates. 1 In 1986, Dr.
Radin and Dr. Rose 2 were the first to suggest that the
subchondral bone plays an important role in the initiation
and progression of OA. Physical characterizations of OA
include cartilage degradation, subchondral bone sclerosis
and thickening, and osteophyte formation. 3-5 In addition,
trabecular bone in the subchondral region is thinned, and its
elasticity is lost. 6-8
An increasing line of evidence suggests that there is a
strong interrelationship between subchondral bone and
articular cartilage. Subchondral bone is separated from the
articular cartilage by a thin layer of calcified cartilage. 9
Cartilage
2(3) 265 –278
© The Author(s) 2011
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1947603510383855
http://cart.sagepub.com
Objective: The pathophysiology of osteoarthritis involves the whole joint and is characterized by cartilage degradation
and altered subchondral bone turnover. At present, there is a need for biological models that allow investigation of the
interactions between the key cellular players in bone/cartilage: osteoblasts, osteoclasts, and chondrocytes. Methods: Femoral
heads from 3-, 6-, 9-, and 12-week-old female mice were isolated and cultured for 10 days in serum-free media in the absence
or presence of IGF-I (100 nM) (anabolic stimulation) or OSM (10 ng/mL) + TNF-α (20 ng/mL) (catabolic stimulation).
Histology on femoral heads before and after culture was performed, and the growth plate size was examined to evaluate the effects
on cell metabolism. The conditioned medium was examined for biochemical markers of bone and cartilage degradation/
formation. Results: Each age group represented a unique system regarding the interest of bone or cartilage metabolism.
Stimulation over 10 days with OSM + TNF-α resulted in depletion of proteoglycans from the cartilage surface in all ages.
Furthermore, OSM + TNF-α decreased growth plate size, whereas IGF-I increased the size. Measurements from the
conditioned media showed that OSM + TNF-α increased the number of osteoclasts by approximately 80% and induced
bone and cartilage degradation by approximately 1200% and approximately 2600%, respectively. Stimulation with IGF-I
decreased the osteoclast number and increased cartilage formation by approximately 30%. Conclusion: Biochemical markers
and histology together showed that the catabolic stimulation induced degradation and the anabolic stimulation induced
formation in the femoral heads. We propose that we have established an explant whole-tissue model for investigating cellcell
interactions, reflecting parts of the processes in the pathogenesis of joint degenerative diseases.
Keywords
osteoarthritis, growth plate, ex vivo, femoral head, cytokines, IGF-I
52
How much do bone and its interaction with cartilage contribute
to the causality of OA? It has been suggested that
bone turnover increases in patients with OA. 3,4 In a canine
model, subchondral bone loss was seen in the early OA and
bone sclerosis in the late OA. 10,11 A murine OA model, an
anterior cruciate ligament transaction (ACLT) model,
showed that bisphosphonates, approved as antiresorptive
1
Cartilage Biology and Biomarkers R&D, Nordic Bioscience A/S, Herlev,
Denmark
2
Department of Molecular Biology, University of Copenhagen,
Copenhagen, Denmark
3
Bone Biology R&D, Nordic Bioscience A/S, Herlev, Denmark
Corresponding Author:
Suzi Hoegh Madsen, Nordic Bioscience A/S, Herlev Hovedgade 207, 2730
Herlev, Denmark
Email: shm@nordicbioscience.com

266 Cartilage 2(3)
therapy for osteoporosis, effectively retarded the progression
of cartilage degeneration and osteophyte formation,
indicating the importance of bone remodeling in the pathogenesis
of OA. 12 This and other studies suggest the existence
of important interactions between bone and cartilage,
which might be key events in OA pathogenesis. 13-17 However,
human cartilage explants cultured in the presence of isolated
osteoblasts from OA patients have shown increased
proteoglycan degradation compared to cartilage cultured
with osteoblasts from healthy patients. 13 Another group
suggested that osteoblasts from sclerotic subchondral bone
could initiate chondrocyte hypertrophy. 14 The communication
between osteoblasts and osteoclasts via PTH/RANKL
in normal bone turnover, where PTH stimulates osteoblasts
to release RANKL that in turn stimulates bone resorption
via osteoclasts, is well acknowledged. 16 However, this
communication via PTH/RANKL is altered in joint tissues
from OA patients. 17 This lack of correlation suggests that
the communication between the cells is very important for
maintaining a healthy joint, and more attention should be
directed toward investigating this communication.
Bone and cartilage degradation can be measured by biochemical
markers, such as fragments of C-terminal telopeptide
of type I collagen (CTX-I), reflecting bone resorption, 18 and
fragments from C-terminal telopeptide of type II collagen
(CTX-II), reflecting cartilage degradation. 19,20 The turnover of
aggrecan, the predominating proteoglycan in cartilage, can be
measured by the release of sulfated glycosaminoglycans
(sGAGs) from the extracellular matrix (ECM). The formation
of cartilage can be measured by the biomarker, PIINP, which
is the collagen type II propeptide at the N-terminal. 21
Biochemical markers for measuring biological processes,
especially in OA, are valuable descriptive tools.
A model that not only identifies the different changes in
cartilage degradation but simultaneously changes adjacent
bone may be useful for investigating interactions between
these 2 tissues. An in vitro model like this would allow us
to investigate and evaluate potential effects of a given compound
before in vivo experiments. Thus, such a model may
be an important tool for finding potential joint disease–
modifying drugs with “dualaction” targeting both bone and
cartilage. Femoral heads from mice are small, enclosed
compartments, which comprise both bone and cartilage.
Thus, the interaction between osteoblasts, osteoclasts, and
chondrocytes remains intact in the closed system when
isolated from mice. The femoral heads are an open opportunity
for investigating the signaling and communication
between cells of bone and cartilage in near-normal conditions
and under stimulation from a compound of interest.
In this study, we developed and characterized a murine
femoral head ex vivo model in which we were able to
induce a catabolic or anabolic response. As controls, we
chose oncostatin M (OSM) + tumor necrosis factor (TNF)-α
and insulin-like growth factor (IGF)-I, which we have
53
previously established in other ex vivo models as catabolic
and anabolic controls, respectively. 21,22 The idea for this
model has emerged from a previously reported murine
femoral head model, where only the cartilage compartment
was cultured as part of investigating the enzymatic proteo-
glycan degradation. 23
Methods
The Murine Femoral Head Explant Model
Femoral heads from 3-, 6-, 9-, and 12-week-old NMRI
(inbred) female mice (Charles River, Sulzfeld, Germany)
were isolated by cutting at the femoral neck. The femoral
heads were cultured in the media DMEM:F12 (Invitrogen,
Taastrup, Denmark) + penicillin and streptomycin (P/S)
(Lonza, Vallensbaek Strand, Denmark) + 10% serum
(Invitrogen) for 2 hours to adjust to the isolation. The
femoral heads were washed in serum-free DMEM:F12 +
P/S 5 times and randomly placed in doublets in each well
in a 24-well plate. Some femoral heads were saved directly
in 4% formaldehyde (Merck, Hellerup, Denmark) to be
able to evaluate the femoral heads at the point before culturing
(T = 0). The rest of the femoral heads were cultured
for 10 days in DMEM:F12 + P/S in the absence (nonstimulated
control named W/O) or presence of the anabolic factor,
IGF-I (100 nM) (Sigma-Aldrich, Copenhagen,
Denmark), or catabolic cytokines, OSM (10 ng/mL)
(Sigma-Aldrich) + TNF-α (20 ng/mL) (R&D Systems,
Abingdon, UK). Each treatment was repeated 5 times. The
conditioned media were changed and saved in –20 °C at
days 4, 7, and 10, except for experiments with 12-week-old
mice, which only were saved at day 10 (the conditioned
media at day 10 comprises the total release from the whole
culture period). Overall cell viability was monitored using
the dye Alamar Blue (Invitrogen) at day 10. The femoral
heads were washed 3 times in PBS (Lonza) and saved in
4% formaldehyde. All culturing of femoral heads was at
37°C and 5% CO 2 . For experiments with 3-, 6-, and 9-weekold
mice, the conditioned media from days 4, 7, and 10
were assessed by biochemical markers of bone and cartilage.
All 3 days were accumulated into one graph, to represent
the total release of the respective biochemical marker
from the femoral heads, during the 10-day culture period.
For the experiments with 12-week-old mice, the total
releases were measured at day 10. The microscopic changes
were assessed by histology.
Biochemical Markers
The biochemical marker CTX-I was used to detect type I collagen
degradation fragments in the conditioned medium. The
competitive ELISA used, RatLaps (IDS Ltd., Herlev,
Denmark), quantified cathepsin K–mediated degradation

Madsen et al. 267
products of type I collagen. The primary antibody was
directed toward the neoepitope 1196 EKSQDGGR. 24 Serum
Pre-clinical CartiLaps (IDS Ltd.) was used to quantify MMPmediated
degradation products from type II collagen. The
primary antibody was directed toward the neoepitope
1230 EKGPDP. 24 PIINP (IDS Ltd.) was a biochemical marker for
cartilage formation. The primary antibody was directed against
an internal epitope of the N-terminal propeptides of collagen
type II: 122 GPQGPAGEQGPRGDR 136 . 25 All 3 ELISAs were
preformed accordingly to the manufacturer’s protocols.
Measurement of Collagen Turnover
The amino acid hydroxyproline was used as a measurement
of total collagen release since it is a major component of
collagen. The 20 uL sample or standard (prepared from
1-hydroxyproline diluted in 1 mM HCl) was diluted 5 times
in 7.5 M HCl and hydrolyzed overnight (20 hours) at 110 °C.
Samples were centrifuged for 2 minutes at 3000 g/min and
then evaporated at 50 °C until samples were completely
dry. Isopropanol/H 2 O (2:1) of 25 uL (Merck) was used to
dissolve samples, and 10 uL of dissolved sample was transferred
to a new plate, to which 20 uL isopropanol was
added. This was oxidized by the addition of 10 uL of solution
I (1 part of 24 M chloramines [Sigma-Aldrich] to 4
parts of 1 M Na-acetate [Sigma-Aldrich], 0.33 M Na 3
citrate [Merck] and 0.07 M citric acid [Merck] dissolved in
isopropanol) incubated for 4 ± 1 minutes. There was 130 uL of
solution II (3 parts of 4.5 M 4-dimethylamino-benzaldehyde
[Sigma-Aldrich] dissolved in 60% perchloric acid [Merck]
to 13 parts of isopropanol) added and incubated at 60 °C for
25 ± 5 minutes. The absorbance was measured at 558 nm
on an ELISA reader.
Quantification of Osteoclast Number
TRAP was a quantitative measurement of osteoclast
number. The reaction buffer (1.0 M sodium-acetate, 0.5%
Triton X-100, 1 M NaCl, 10 mM EDTA [pH 5]) (Merck),
50 mM ascorbic acid (Sigma-Aldrich), 0.2 M disodium
tartrate (Merck), 82 mM 4-nitrophenylphosphate (Sigma-
Aldrich), and milli = Q were mixed 2:1:1:1:3 into a final
TRAP solution buffer. Sample and final TRAP solution
buffer was mixed (1:4) and incubated for 1 hour at 37 °C.
There was 0.3 M NaOH (Merck) added to stop the color
reaction, and the absorbance was measured at 405 nm with
650 nm as reference.
Quantification of Proteoglycans by Measurement of
Sulfated Glycoaminoglycans
Proteoglycans were used to quantify cartilage turnover by
measuring sulfated glycosaminoglycans. An assay using the
dye 1.9-dimethylmehylene blue chloride (Sigma-Aldrich)
54
in a solution of sodium formate (Sigma-Aldrich) and formic
acid (Merck) was used. A standard curve was prepared
with chondroitin sulfate (Sigma-Aldrich) ranging from 0 to
100 ng/mL. The 40 uL standard or sample was transferred
to a microtiter plate, followed by 250 uL 38.5 uM
1.9-dimethylmethylene blue chloride. The absorbance was
immediately read at 650 nm on an ELISA reader to avoid
precipitation.
Histological Analysis
Tissue preparation. After the culture period, femoral heads
were decalcified in 15% EDTA at room temperature for
approximately 1 week. Femoral heads were embedded separately
in paraffin blocks and cut into 5-µm-thick sections.
Excessive paraffin was removed by placing sections at 60 °C
for 1 hour and dried at 37 °C overnight before staining.
Safranin O and fast green staining. To visualize bone and
cartilage, safranin O (Sigma-Aldrich) stained proteoglycans
red, and fast green (Sigma-Aldrich) stained collagens
green. Cartilage was stained red due to the excess of proteoglycans
compared to collagens. Sections (5 µm) were
stained according to Bay-Jensen et al. 26 Digital histographs
were taken using an Olympus BX60F-3 microscope and an
Olympus DP71 camera (Tokyo, Japan).
Annotations of the femoral head to measure size of the
growth plate. Digital histographs of femoral heads from
12-week-old mice were evaluated by standardized annotations,
ensuring anatomical correspondence. A line connects
the black dots where the cartilage starts on the femoral
neck. The center of this line is perpendicularly connected to
the cartilage surface, and 2 lines are placed at an angle of
30° from this perpendicular line. The lengths of the section
of the 3 lines where they cross the growth plate are calculated
to find the average length of the growth plate. The
growth plate was evaluated in 4 different femoral heads
(from 4 different experiments) from each treatment.
Statistical Analysis
Results are shown as mean ± standard error of the mean
(SEM). Differences between mean values were compared by
the Student t test for unpaired observations, assuming normal
distribution where 5 replicates were used. Differences were
considered statistically significant if P < 0.05.
Results
Femoral Heads Exhibit Different Bone and Cartilage
Morphology during Development
Each individual age has a unique bone and cartilage morphology.
In 3-week-old mice, the cartilage compartment is
dominating (red stain) compared to the bone compartment

268 Cartilage 2(3)
Figure 1. Characterization of the development in murine femoral heads. Sections from femoral heads from 3-, 6-, 9-, and 12-week-old
mice were stained with safranin O and fast green, which color the proteoglycans red and the collagens green. The black bars represent
250 µM (A-D) and 100 µM (E-L). The numbers represent growth plate (1), ligamentum teres femoris (2), endochondral ossification (3),
immature bone embedded with calcified cartilage (4), prehypertrophic and hypertrophic chondrocytes (5), calcified cartilage (6), woven
bone (unmineralized) (7), lamellar bone (mineralized) (8), secondary ossification center (9), and articular cartilage (10). The black squares
in the whole femoral heads (A-D) represent the magnification of either the cartilage compartment (E-H) or the bone compartment
(I-L), which is divided in columns after age.
(green stain); however, the cartilage is not fully materialized
(Fig. 1A). The cartilage-bone ratio changes over time
due to the development of the secondary ossification center
at the age of 9 weeks (Fig. 1C). The growth plate is visible
as a thick line of dense red staining separating bone from
cartilage in the 3-week-old mouse, which decreases concurrently
with increasing age (Fig. 1A-D). In femoral heads
of 3-week-old mice, the cartilage predominantly consists of
proliferating, prehypertrophic, and hypertrophic chondrocytes
(Fig. 1E), whereas the 6- and 9-week-old mice have
less proliferating chondrocytes, which are located in the
reduced growth plate (Fig. 1F and G). Additionally, the
extracellular matrix of the cartilage compartment in 3- to
9-week-old mice is calcified (Fig. 1E-G), preparing itself
for the secondary ossification center, whereas the cartilage
compartment is mainly articular cartilage (noncalcified) at
12 weeks of age (Fig. 1H). The bone compartment in the
3-week-old mice consists of immature bone embedded with
calcified cartilage (Fig. 1I). As the mice get older, the calcified
cartilage is replaced by unmineralized bone (normal
endochondral ossification process), resulting in woven
bone with a high proportion of osteocytes (Fig. 1J and K).
The mechanically weak woven bone is replaced by more
55
resilient lamellar bone (mechanically strong) with a low
proportion of osteocytes when the mice reach the age of 12
weeks (Fig. 1L). Overall, these results suggest that the different
ages represent different developmental stages, as
expected, especially with regards to ratio and quality of
bone and cartilage, and thus, these can individually be used
in our ex vivo model for specific purposes.
Anabolic and Catabolic Stimulations Induce
Morphological Changes
The viability of the cells of the cultured femoral heads is
not impaired by the 10 days of anabolic or catabolic stimulation
(data not shown). Femoral heads stimulated with
OSM + TNF-α show a major loss of proteoglycans from the
cartilage compartment at the age of 3 weeks (Fig. 2F) compared
to the nonstimulated control (W/O) (Fig. 2A). In
9-week-old femoral heads, OSM + TNF-α stimulation
increases the loss of proteoglycans from the cartilage surfaces
(Fig. 2H) compared to the nonstimulated controls
(W/O) (Fig. 2C). However, in femoral heads from 6-weekold
mice, both nonstimulated (Fig. 2B) and OSM + TNFα–stimulated
(Fig. 2G) femoral heads have lost proteoglycans

Madsen et al. 269
Figure 2. Microscopic changes after catabolic and anabolic stimulation. Sections from femoral heads from 3-, 6-, 9-, and 12-week-old
mice were stained with safranin O (proteoglycans) and fast green (collagens) after culture for 10 days in the absence (W/O) or presence
of IGF-I (anabolic) and OSM + TNF-α (catabolic) stimulation. The black bars represent 250 µM (A-D, F-I, K-N) and 20 µM (E, J, O). The
black squares in the 12-week-old femoral heads (D, I, N) represent the magnification of the cartilage surface (E, J, O), which is divided
in rows after treatment.
from the cartilage surface. In 12-week-old femoral heads,
the proteoglycans seem to be almost depleted from the cartilage
compartment in the nonstimulated (Fig. 2D) and the
OSM + TNF-α–stimulated (Fig. 2I) explants. Anabolic
stimulation with IGF-I protects against the proteoglycan
loss seen in W/O (Fig. 2A-D and K-N). Additionally, in
12-week-old mice, anabolic stimulation shows proteoglycan
staining around the chondrocytes (Fig. 2O) compared to the
W/O (Fig. 2E). Next, we measured the size of the growth
plate to investigate whether morphological changes could
be induced (Fig. 3A). The size of the growth plate in IGF-
I–stimulated 12-week-old femoral heads is increased by
approximately 48% (P < 0.001) compared to W/O, whereas
OSM + TNF-α stimulation decreases the growth plate size
by approximately 34% (P < 0.001) compared to W/O (Fig.
3B). This indicates that we are able to induce alteration to
the femoral heads ex vivo with catabolic (degenerative) or
anabolic (generative) factors.
Catabolic Stimulation Increases Collagen and
Proteoglycan Turnover
Levels of hydroxyproline, an indicator of total collagen
turnover, in the conditioned medium indicate that stimulation
by OSM + TNF-α increases the total release of collagens by
approximately 49% (P < 0.05) to approximately 190%
(P < 0.001) in femoral heads aged 6, 9, and 12 weeks, but not in
femoral heads from 3-week-old mice (Fig. 4A). The longitudinal
effects from the different stimulations are shown in
Table 1. OSM + TNF-α increases the release of collagens
56
Figure 3. Quantification of growth plate zone in femoral heads.
Sections from femoral heads from 12-week-old mice were stained
with safranin O (proteoglycans) and fast green (collagens) after
culture for 10 days in the absence (W/O) or presence of IGF-I
(anabolic) and OSM + TNF-α (catabolic) stimulation. The growth
plate size was measured in one femoral head per treatment from 4
different experiments using the annotation system (see Methods).
The growth plate thickness (µm) is calculated by the average of
the black lines in the growth plate (A). The catabolic stimulation
decreased the growth plate size, whereas the anabolic stimulation
increased the size of the growth plate (B). ***P < 0.001.
over time in 6- and 9-week-old mice. In femoral heads from
12-week-old mice, the anabolic stimulation also increases
total collagen release by approximately 49% (P < 0.01)
compared to W/O (Fig. 4A). OSM + TNF-α stimulation
increases the total release of sulfated glycosaminoglycans
(sGAG) to the conditioned media by approximately 22%
(P < 0.05) to approximately 40% (P < 0.01), except from
the experiment with 6-week-old mice (Fig. 4B); however,
similar trends were observed.

270 Cartilage 2(3)
Figure 4. The catabolic control induces a release of collagens and proteoglycans. The experiments were assessed by the collagen
degradation marker, hydroxyproline (A), and the proteoglycan turnover marker, sGAG (B). Each graph represents the accumulated
measurements from days 4, 7, and 10 except from 12-week-old mice (only day 10). Cytokine stimulation increases the collagen release
in experiments with 6-, 9-, and 12-week-old mice (A). Cytokine stimulation also increases proteoglycan release in experiments with 3-,
9-, and 12-week-old mice (B). Asterisks indicate statistical significance: *P < 0.05; **P < 0.01; ***P < 0.001.
57

Madsen et al. 271
Table 1. Biochemical Markers Measured from the Conditioned Medium at Days 4, 7, and 10 in the Different Age Groups
Age of Culture Stimulated
Mice, wk Day with HP, µg/mL sGAG, µg/mL TRAP CTX-I, ng/mL PIINP, ng/mL CTX-II, pg/mL
W/O 5.22 ± 0.81 113.12 ± 3.46 0.47 ± 0.06 2.97 ± 0.82 93.61 ± 4.65 48.69 ± 3.73
4 O + T 6.21 ± 1.22 143.84 ± 8.45** 0.81 ± 0.12* 3.23 ± 0.27 54.92 ± 9.58** 74.96 ± 19.19
IGF 5.16 ± 0.59 92.61 ± 10.11* 0.63 ± 0.06 3.79 ± 0.84 100 ± 0.00 26.10 ± 3.87**
W/O 2.59 ± 0.46 49.51 ± 8.51 0.27 ± 0.02 4.61 ± 0.54 68.78 ± 5.55 6.30 ± 1.40
3 7 O + T 2.41 ± 0.35 87.18 ± 5.04** 0.33 ± 0.01* 11.72 ± 3.54 40.83 ± 3.02** 17.44 ± 3.83*
IGF 4.38 ± 0.46* 58.96 ± 4.34 0.26 ± 0.01 4.49 ± 0.39 74.38 ± 1.76 15.28 ± 1.86**
W/O 3.56 ± 0.53 70.74 ± 4.85 0.22 ± 0.001 2.55 ± 0.56 46.01 ± 3.05 0.48 ± 0.38
10 O + T 3.09 ± 0.29 83.18 ± 9.70** 0.25 ± 0.01** 3.62 ± 0.47 28.95 ± 2.11** 7.90 ± 1.93**
IGF 4.53 ± 0.55 89.71 ± 3.88* 0.21 ± 0.01 2.26 ± 0.17 69.24 ± 1.38*** 8.51 ± 1.50**
W/O 1.34 ± 0.16 40.52 ± 5.05 0.29 ± 0.03 3.21 ± 0.39 54.55 ± 6.48 10.70 ± 0.86
4 O + T 1.69 ± 0.11 50.53 ± 0.68 0.36 ± 0.03 5.29 ± 0.76* 64.01 ± 4.38 29.66 ± 4.14**
IGF 1.27 ± 0.09 24.16 ± 3.10 0.21 ± 0.01* 4.26 ± 0.16* 82.18 ± 6.23* 8.38 ± 2.30
W/O 1.48 ± 0.28 24.22 ± 4.24 0.24 ± 0.01 3.11 ± 0.65 74.48 ± 2.48 16.26 ± 1.68
6 7 O + T 1.57 ± 0.13 20.86 ± 1.01 0.46 ± 0.05** 13.77 ± 4.18* 59.43 ± 3.58** 139.18 ± 27.05**
IGF 1.77 ± 0.06 24.86 ± 2.79 0.18 ± 0.002*** 4.21 ± 0.47 64.86 ± 5.31 14.81 ± 2.50
W/O 1.53 ± 1.02 6.42 ± 1.32 0.19 ± 0.01 2.08 ± 0.14 56.22 ± 1.54 12.94 ± 1.70
10 O + T 2.37 ± 0.38 10.81 ± 1.00 0.45 ± 0.06** 6.02 ± 1.07** 48.12 ± 4.02 160.84 ± 45.55*
IGF 2.02 ± 0.45 10.35 ± 0.51 0.19 ± 0.02 4.58 ± 0.29*** 44.61 ± 5.05 14.47 ± 2.19
W/O 2.26 ± 0.40 25.56 ± 1.35 0.33 ± 0.03 10.82 ± 0.84 65.45 ± 3.70 68.75 ± 11.89
4 O + T 1.61 ± 0.38 28.26 ± 2.72 0.36 ± 0.04 14.46 ± 2.72 54.89 ± 3.22 59.61 ± 9.14
IGF 2.05 ± 0.37 22.08 ± 2.46 0.24 ± 0.004** 4.80 ± 0.29*** 78.57 ± 2.69* 36.53 ± 8.24
W/O 1.49 ± 0.35 9.70 ± 1.93 0.36 ± 0.04 14.86 ± 2.50 53.16 ± 4.19 32.18 ± 3.13
9 7 O + T 3.01 ± 0.88 15.95 ± 1.56* 0.48 ± 0.06 18.71 ± 1.98 46.87 ± 2.82 309.83 ± 34.52***
IGF 1.49 ± 0.27 11.39 ± 2.18 0.23 ± 0.01** 14.87 ± 2.68 66.78 ± 2.62* 18.63 ± 5.09
12 a
W/O 2.34 ± 0.41 14.12 ± 2.34 0.30 ± 0.02 9.09 ± 1.07 29.98 ± 3.36 30.86 ± 6.86
10 O + T 6.76 ± 0.13*** 20.10 ± 0.91* 0.55 ± 0.07** 24.13 ± 3.33** 33.73 ± 2.81 361.52 ± 2.09***
IGF 3.05 ± 0.81 13.32 ± 1.35 0.23 ± 0.03 6.25 ± 0.49* 33.09 ± 1.78 37.77 ± 13.16
W/O 4.79 ± 0.53 19.74 ± 1.31 0.19 ± 0.008 27.84 ± 8.69 20.31 ± 3.61 62.88 ± 28.45
10 O + T 13.89 ± 1.59*** 24.00 ± 1.39* 0.30 ± 0.037* 341.39 ± 78.25*** 0.80 ± 0.80* 1714.80 ± 379.31**
IGF 7.15 ± 0.43** 23.67 ± 1.11 0.17 ± 0.011 42.60 ± 16.09 67.78 ± 0.73*** 54.07 ± 22.04
Note: Values are significantly different from W/O.
a The biochemical markers released from 12-week-old mice at day 10 represent total release from the entire culture period.
*P < 0.05. **P < 0.01. ***P < 0.001.
OSM + TNF-α Increases the Osteoclast Number
and Bone Resorption
Measurements of the conditioned media show that stimulation
with OSM + TNF-α increases the tartrate-resistant
acid phosphatase (TRAP) activity for all the age groups in
a range from approximately 41% (P < 0.05) to approximately
78% (P < 0.01) compared with their respective
W/O, indicating increased osteoclast numbers (Fig. 5A).
OSM + TNF-α increases the osteoclast number over time
in 6- and 9-week-old mice. However, the osteoclast number
decreases over time in 3-week-old mice (Table 1).
Conversely to OSM + TNF-α stimulation, IGF-I stimulation
shows approximately 17% (P < 0.05) to approximately
29% (P < 0.01) decrease in the osteoclast number
58
in 6- and 9-week-old mice (Fig. 5A), which results from
constant low values measured at all 3 days (Table 1). The
collagen type I resorption measurements reflect the
respective osteoclast number during catabolic stimulation.
The resorption increases approximately 65% (P <
0.05) to approximately 1200% (P < 0.001) in the presence
of OSM + TNF-α stimulation compared to W/O (Fig. 5B).
However, in 6-week-old mice, the collagen type I resorption
measurements do not reflect the TRAP activity when
stimulated with IGF-I. In 3- and 6-week-old mice, the
resorption of collagen type I peaks at day 7, whereas it
peaks at or after day 10 in 9-week-old mice (Table 1).
These measurements show that it is possible to measure
osteoclast number and activity via biochemical markers in
our ex vivo model.

272 Cartilage 2(3)
Figure 5. Cytokines increase the osteoclast number and resorption of bone. The conditioned media from the 4 experiments with 4
individual age groups were assessed by the osteoclast number measured by TRAP activity (A) and the collagen type I resorption marker,
CTX-I (B). Each graph represents the accumulated measurements from days 4, 7, and 10 except from 12-week-old mice (only day 10).
Catabolic stimulation increases the osteoclast number in all 4 experiments, whereas the anabolic stimulation decreases the osteoclast
number in the experiments with 6- and 9-week-old mice (A). Cytokines (O + T) also increase the total collagen type I resorption in the
experiments with 6-, 9-, and 12-week-old mice (B). *P < 0.05. **P < 0.01. ***P < 0.001.
59

Madsen et al. 273
IGF-I and OSM + TNF-α Stimulation Mediate
Collagen Type II Formation and Degradation,
Respectively
Collagen type II formation, measured by the PIINP released
into the conditioned medium, decreases over time (Table 1).
However, the total collagen type II formation increases by
approximately 22% (P < 0.05), approximately 21% (P <
0.05), and approximately 230% (P < 0.001) in anabolic
stimulated (IGF-I) femoral heads from 3-, 9-, and 12-weekold
mice, respectively, when compared to W/O (Fig. 6A).
OSM + TNF-α decreases collagen type II formation by
approximately 38% (P < 0.01) and approximately 96%
(P < 0.05) compared to W/O in 3- and 12-week-old mice,
respectively (Fig. 6A).
Stimulation with OSM + TNF-α increases the degradation
of collagen type II by approximately 450% (P < 0.001)
to approximately 2600% (P < 0.01) in femoral heads from
mice aged 6, 9, and 12 weeks, while femoral heads from
3-week-old mice show only a nonsignificant increase
(~80%) compared with W/O (Fig. 6B). The longitudinal
release pattern of CTX-II shows that OSM + TNF-α
increases the collagen type II degradation over time in 6-
and 9-week-old mice (Table 1). In 3-week-old mice, the
release of CTX-II decreases over time (Table 1). These
results indicate that we are able to induce cartilage formation
and degradation in our ex vivo model.
Discussion
OA is a complex disease of the entire joint affecting both
bone and cartilage. 12,27 Currently, there are no effective
disease-modifying OA drug (DMOAD) treatments. 28 We
have developed an ex vivo murine femoral head model,
comprising both bone and cartilage in one closed compartment
system that allows us to identify changes in bone and
cartilage simultaneously. This model allows interactions
between cartilage and bone cells—chondrocytes, osteoblasts,
and osteoclasts—in a manner resembling their in situ
conditions and microenvironment. However, the availability
for humoral communication, due to the direct contact
between the culture media and the different tissues, far
exceeds what is possible in vivo. Furthermore, the ex vivo
model does not mimic the biomechanical communication
between bone and cartilage that is part of both normal
physiology and pathophysiology. Nevertheless, the model
is an important fundamental tool, allowing us to find preliminary
results for potential therapies targeting one or both
tissues. These preliminary results could aid the selection of
which future in vivo experiments, concerning joint diseases,
should be investigated.
We were able to induce catabolic and anabolic responses by
stimulation with catabolic and anabolic factors, respectively.
60
Catabolic stimulation resulted in increased bone resorption
and degradation of cartilage, whereas anabolic stimulation
had a protective effect against proteoglycan loss and
resulted in increased cartilage formation.
Even though the different developmental stages of
endochondral ossification, including that of the femoral
head, are known from the literature already, 29 it was important
to determine the individual stages in our particular
murine model for comparison with the available literature.
The microscopic evaluation of femoral heads from mice
aged 3 to 12 weeks revealed unique bone and cartilage
morphologies.
In 3-week-old mice, the cartilage had a high metabolic
rate due to the ongoing endochondral ossification process.
The cartilage compartment was dominating compared to
the bone compartment and mainly consisted of proliferating
and hypertrophic chondrocytes, which included a large
and active growth plate. Hypertrophic chondrocytes are a
hallmark of early OA, 30 so this murine age group is highly
relevant for preliminary studies, for finding potential
DMOADs, by investigating the complicated nature of this
phenotype. This age group is also relevant for investigating
the metabolism of very active cartilage in vitro.
In mice aged 6 and 9 weeks, the ratio of bone to cartilage
was more even, and an emerging secondary ossification
center had started to penetrate the cartilage compartment
in 9-week-old mice. The bone compartment was deposited
as woven bone with a high proportion of osteocytes. Woven
bone forms quickly when osteoblasts produce osteoid,
which occurs initially in all fetal bones but is weak due to
a disorganized collagen structure. The growth plate was
reduced by more than 70%, separating the woven bone
from the indefinable mix of prehypertrophic and hypertrophic
chondrocytes (in calcified cartilage) in an evenlooking
arc. These 2 age groups are unique for investigating
the metabolism of hypertrophic chondrocytes in the presence
of woven bone (comprised of osteoclasts, osteoblasts,
and osteocytes) or for studying the initial development of
the secondary ossification center.
In 12-week-old mice, the bone compartment had become
larger than the cartilage compartment, in direct contrast to
the situation with 3-week-old mice. The calcified cartilage
had been replaced by woven bone during the secondary
ossification, which was enclosed by noncalcified cartilage,
resembling mature articular cartilage. The woven bone
from the endochondral ossification had been replaced by
the stronger and more resilient lamellar bone, which was
highly organized in concentric collagen sheets with a low
proportion of osteocytes. A small growth plate was still
present and separated the endochondral ossification (lamellar
bone) from the secondary ossification (woven bone).
This older age group is unique for investigating mature
bone and subchondral bone metabolism in the presence of

274 Cartilage 2(3)
Figure 6. The anabolic and catabolic controls induce cartilage formation and degradation, respectively. The experiments were assessed
by the collagen type II formation marker, PIINP (A), and the degradation marker, CTX-II (B). Each graph represents the accumulated
measurements from days 4, 7, and 10 except from 12-week-old mice (only day 10). IGF-I induces formation in the experiments with 3-,
9-, and 12-week-old mice (A). Cytokines (O + T) increased the total cartilage degradation in experiments with 6-, 9-, and 12-week-old
mice (B). *P < 0.05. **P < 0.01. ***P < 0.001.
61

Madsen et al. 275
a reduced quantity of mature articular cartilage. Importantly,
12-week-old mice resemble a human joint compared to the
other age groups. Furthermore, the model system enables
us to induce anabolic or catabolic responses, which mimic
parts of the processes from different bone and cartilage
diseases in an adult individual.
The microscopic evaluation of femoral heads cultured
for 10 days in the absence or presence of IGF-I (anabolic)
or OSM + TNF-α (catabolic) stimulation showed that
femoral heads stimulated with OSM + TNF-α lost more
proteoglycans from the cartilage compartment than the
W/O at almost all age groups. Nonstimulated femoral
heads from 6- and 12-week-old mice also lost proteoglycans
from the cartilage, as seen in the OSM + TNF-α
stimulation. This suggests that the proteoglycan loss may
be a spontaneous effect from the culturing rather than a
catabolic effect from the OSM + TNF-α stimulation. The
femoral heads stimulated with the anabolic factor (IGF-I)
showed that this loss and depletion of proteoglycans during
culture were protected compared to the W/O for all 4 age
groups. Furthermore, in 12-week-old mice, anabolic stimulation
showed far more proteoglycan staining around the
chondrocytes compared to the W/O, indicating a regenerative
effect of IGF-I, as previously observed in bovine cartilage
explants. 31 We also showed that IGF-I stimulation in
12-week-old mice increased the growth plate size significantly,
whereas the OSM + TNF-α stimulation significantly
decreased the size. These findings corroborate previous
findings in the literature; IGF-I is very important for the
growth plate during bone growth as it increases proliferation
of resting and proliferative chondrocytes, and OSM +
TNF-α has, on the contrary, shown to decrease the rate of
endochondral bone growth. 32-34
Biochemical markers are valuable, longitudinal, and
dynamic tools, which in combination with histology can
assess the systematic effects of specific compounds. The
histology shows the end result of the effect of a specific
compound on the femoral head, whereas biochemical
markers reveal which protein processes are up-regulated
and down-regulated and in which order they are released.
We have chosen bar graphs to present the total release of
the biochemical markers to the conditioned media as accumulated
data from all media changes. Table 1 presents all
the details for the individual days.
Hydroxyproline measurements showed that the release
of collagen fragments increased in the presence of OSM +
TNF-α stimulation, except from 3 weeks of age. However,
IGF-I also increased the release of collagens in 12-weekold
mice. If this collagen release results from increased
degradation, increased turnover or lack of incorporation
into the extracellular matrix after synthesis may be an irrelevant
discussion for our use. We can use the results of “collagen
release” in combination with other biochemical
62
markers of collagens in our model. We can furthermore not
exclude the fact that some portion of the hydroxyproline
measured is released from the ligamentum teres.
The release of proteoglycans from the femoral heads
decreased proportionally with the increasing age of the
mice, indicating that the release of proteoglycans mainly
comes from the cartilage compartment and not the bone.
However, a limitation of the current study is that we were
not able to analyze the specific proteolytic mechanisms
responsible for sGAG release. Future studies should thus
aim to determine the longitudinal patterns of MMP- and
ADAMTS-mediated aggrecan degradation in the femur
head model and compare to other in vitro and in vivo models
of bone and cartilage diseases. To evaluate the results
further, we then measured more protein-specific biochemical
markers.
We found that OSM + TNF-α stimulation for 10 days
increased the number of osteoclasts and increased the
resorption of collagen type I. Even though the total release
of TRAP increased in the presence of OSM + TNF-α (in
3-week-old mice), the osteoclast number decreased over
time for both nonstimulated and stimulated treatments.
However, the TRAP concentrations at day 4 were higher
than the concentration at day 10 for the other age groups.
This incoherence might be explained by a peak at day 4 for
the 3-week-old mice, in which we did not find a large
amount of bone. The increased resorption seen in this
model is consistent with the literature, which reports that
subchondral bone was lost in the early stage of OA,
whereas bone sclerosis was present at the late stage of
OA. 10,11 Furthermore, TNF-α can induce osteoclast differentiation
in vitro through a mechanism independent of
RANKL. 35,36 Additionally, in murine osteoblastic MC3T3-E1
cells, M-CSF expression is constitutive and can be further
increased by TNF-α. 37-39 OSM can, in vitro, induce the
formation of osteoclasts, leading to bone resorption.
However, OSM can also influence differentiation and proliferation
of osteoblasts, leading to bone formation. 40-42
The anabolic stimulation with IGF-I generally showed,
in all the age groups, a tendency to decrease the number of
osteoclasts, indicating a protective effect on the bone by
reducing bone resorption. However, the collagen type I
resorption did not decrease but seemed to be more or less
unaffected by the anabolic stimulation. Interestingly, collagen
type I resorption data did not reflect the anabolic
TRAP data from mice aged 6 weeks. The femoral heads
have decreased osteoclast number but increased bone
resorption when treated with IGF-I. This ambiguity between
the different age groups should be investigated further but
may primarily result from the different developmental
stages of the femoral head, suggesting an increased activity
of osteoclasts in 6-week-old mice, where the secondary
ossification center might initiate. However, the literature

276 Cartilage 2(3)
supports the increased resorption in 6-week-old mice. One
study indicates that IGF-I regulates osteoclastogenesis,
which subsequently results in increased bone resorption,
and that IGF-I is required for maintaining the normal interaction
between osteoclasts and osteoblasts through RANKL
and RANK. 43 Another study indicates that IGF-I stimulates
both resorption and formation of bone. 44
The concentration of CTX-I, as a function of OSM +
TNF-α stimulation, increased with increasing age, which is
consistent with the increasing size of the bone compartment
shown by the histology. We expect the majority of CTX-I
to come from bone resorption because it is mediated by
cathepsin K, which is produced by the osteoclasts in this
model. Correspondingly, the cartilage compartment
decreased proportionally with the increasing age of the
mice. However, our study showed that it was still possible
to measure the anabolic and catabolic pattern of total collagen
type II formation (PIINP) in 9- and 12-week-old
mice, as we saw in 3-week-old mice. The formation of collagen
type II decreased over time for all 3 treatments.
However, IGF-I slowed the drop of the formation. This
indicates that the total increase in collagen type II formation,
seen by IGF-I in the bar graph (Fig. 6A), was mediated
by a protective effect instead of an anabolic effect. The
same protective effect of IGF-I stimulation has been published
for bovine cartilage explants. 21 Another possibility is
that the formation of collagen type II peaks at day 4; thus,
we only detect a decrease over time. OSM + TNF-α stimulation
increased the degradation of collagen type II (CTX-II)
in femoral heads of all 4 age groups, which correlates with
earlier ex vivo experiments stimulated with OSM + TNF-α,
which showed that collagen type II degraded. 22,45 The concentration
of CTX-II fragments released to the conditioned
medium from the femoral heads decreased over time in
3-week-old mice, probably due to a peak at day 4 since the
OSM + TNF-α stimulation was very high at that particular
day. Conversely, the concentrations of CTX-II in the other
age groups increased over time and proportionally with the
age, even though the ratio of cartilage to bone became
smaller. This suggests that the peaks for CTX-II release
come later in mice older than 3 weeks. Furthermore, this
indicates that the secondary ossification center plays an
important part in the release of CTX-II. Collagen type II is
the predominant collagen of cartilage. Thus, we expect the
biochemical markers, PIINP and CTX-II, to primarily
come from the cartilage compartment and not from the
bone or ligamentum teres.
In the present study, the biochemical markers in combination
with histology have shown that we now have a
simple ex vivo model that henceforth can be used for
screening compounds of interest. Furthermore, this model
could be essential as a preliminary study to in vivo animal
studies to narrow down expenses and time spent on big
animal studies with futile outcomes.
63
In conclusion, the murine femoral head model we developed
comprises both bone and cartilage in one unit and may
be of particular relevance for investigating the communication
between the different cell types in these 2 tissues. This
might aid the understanding of diseases dealing with miscommunication
or improve the efficacy for finding new
therapies targeting either cartilage or bone or both. Our
model enables observations not only of the cells in in situlike
conditions but also monitoring of changes in bone and
cartilage as mice age and in response to catabolic and anabolic
stimulation. As the morphology of the femoral head
changes with age, our study indicates it is critical to select
the appropriate age of mice to investigate a specific question.
Acknowledgments and Funding
The authors gratefully acknowledge the funding from the Danish
Research Foundation (Den Danske Forskningsfond) for supporting
this work. The salary of the corresponding author is supported
in part by The Ministry of Science Technology and Innovation of
Denmark.
Declaration of Conflicting Interests
All authors declare that the affiliation reveals full disclosure. In
addition, Dr. Karsdal owns stock in Nordic Bioscience.
References
1. Felson DT, Neogi T. Osteoarthritis: is it a disease of cartilage
or of bone? Arthritis Rheum. 2004;50(2):341-4.
2. Radin EL, Rose RM. Role of subchondral bone in the initiation
and progression of cartilage damage. Clin Orthop Relat
Res. 1986;213:34-40.
3. Dieppe P, Cushnaghan J, Young P, Kirwan J. Prediction of the
progression of joint space narrowing in osteoarthritis of the knee
by bone scintigraphy. Ann Rheum Dis. 1993;52(8):557-63.
4. Ratcliffe A, Seibel MJ. Biochemical markers of osteoarthritis.
Curr Opin Rheumatol. 1990;2(5):770-6.
5. Rogers J, Shepstone L, Dieppe P. Is osteoarthritis a systemic
disorder of bone? Arthritis Rheum. 2004;50(2):452-7.
6. Wohl GR, Shymkiw RC, Matyas JR, Kloiber R, Zernicke RF.
Periarticular cancellous bone changes following anterior cruciate
ligament injury. J Appl Physiol. 2001;91(1):336-42.
7. Boyd SK, Matyas JR, Wohl GR, Kantzas A, Zernicke RF.
Early regional adaptation of periarticular bone mineral density
after anterior cruciate ligament injury. J Appl Physiol.
2000;89(6):2359-64.
8. Boyd SK, Muller R, Zernicke RF. Mechanical and architectural
bone adaptation in early stage experimental osteoarthritis.
J Bone Miner Res. 2002;17(4):687-94.
9. Burr DB, Radin EL. Microfractures and microcracks in subchondral
bone: are they relevant to osteoarthrosis? Rheum Dis
Clin North Am. 2003;29(4):675-85.
10. Dedrick DK, Goulet R, Huston L, Goldstein SA, Bole GG.
Early bone changes in experimental osteoarthritis using microscopic
computed tomography. J Rheumatol Suppl. 1991;27:44-5.

Madsen et al. 277
11. Dedrick DK, Goldstein SA, Brandt KD, O’Connor BL, Goulet
RW, Albrecht M. A longitudinal study of subchondral
plate and trabecular bone in cruciate-deficient dogs with
osteoarthritis followed up for 54 months. Arthritis Rheum.
1993;36(10):1460-7.
12. Hayami T, Pickarski M, Wesolowski GA, McLane J, Bone
A, Destefano J, et al. The role of subchondral bone remodeling
in osteoarthritis: reduction of cartilage degeneration and
prevention of osteophyte formation by alendronate in the rat
anterior cruciate ligament transection model. Arthritis Rheum.
2004;50(4):1193-206.
13. Westacott CI, Webb GR, Warnock MG, Sims JV, Elson CJ.
Alteration of cartilage metabolism by cells from osteoarthritic
bone. Arthritis Rheum. 1997;40(7):1282-91.
14. Sanchez C, Deberg MA, Piccardi N, Msika P, Reginster JY,
Henrotin YE. Subchondral bone osteoblasts induce phenotypic
changes in human osteoarthritic chondrocytes. Osteoarthritis
Cartilage. 2005;13(11):988-97.
15. Sanchez C, Deberg MA, Piccardi N, Msika P, Reginster JY,
Henrotin YE. Osteoblasts from the sclerotic subchondral bone
downregulate aggrecan but upregulate metalloproteinases expression
by chondrocytes: this effect is mimicked by interleukin-6,
-1beta and oncostatin M pre-treated non-sclerotic osteoblasts.
Osteoarthritis Cartilage. 2005;13(11):979-87.
16. Takahashi N, Udagawa N, Takami M, Suda T. Cells of bone:
osteoclast generation. In: Bilezikian JP, Raisz LG, Rodan GA,
editors. Principles of bone biology. Vol. 1. 2nd ed. San Diego:
Academic Press; 2002. p. 109-26.
17. Fazzalari NL, Kuliwaba JS, Atkins GJ, Forwood MR, Findlay
DM. The ratio of messenger RNA levels of receptor
activator of nuclear factor kappaB ligand to osteoprotegerin
correlates with bone remodeling indices in normal human
cancellous bone but not in osteoarthritis. J Bone Miner Res.
2001;16(6):1015-27.
18. Bonde M, Qvist P, Fledelius C, Riis BJ, Christiansen C. Immunoassay
for quantifying type I collagen degradation products
in urine evaluated. Clin Chem. 1994;40(11 Pt 1):2022-5.
19. Christgau S, Garnero P, Fledelius C, Moniz C, Ensig M,
Gineyts E, et al. Collagen type II C-telopeptide fragments as
an index of cartilage degradation. Bone. 2001;29(3):209-15.
20. Schaller S, Henriksen K, Hoegh-Andersen P, Sondergaard
BC, Sumer EU, Tanko LB, et al. In vitro, ex vivo, and in vivo
methodological approaches for studying therapeutic targets of
osteoporosis and degenerative joint diseases: how biomarkers
can assist? Assay Drug Dev Technol. 2005;3(5):553-80.
21. Madsen SH, Sondergaard BC, Bay-Jensen AC, Karsdal MA.
Cartilage formation measured by a novel PIINP assay suggests
that IGF-I does not stimulate but maintains cartilage formation
ex vivo. Scand J Rheumatol. 2009;38(3):222-6.
22. Sondergaard BC, Henriksen K, Wulf H, Oestergaard S,
Schurigt U, Brauer R, et al. Relative contribution of matrix
metalloprotease and cysteine protease activities to cytokinestimulated
articular cartilage degradation. Osteoarthritis Cartilage.
2006;14(8):738-48.
64
23. Glasson SS, Askew R, Sheppard B, Carito BA, Blanchet T,
Ma HL, et al. Characterization of and osteoarthritis susceptibility
in ADAMTS-4-knockout mice. Arthritis Rheum.
2004;50(8):2547-58.
24. Hoegh-Andersen P, Tanko LB, Andersen TL, Lundberg CV,
Mo JA, Heegaard AM, et al. Ovariectomized rats as a model
of postmenopausal osteoarthritis: validation and application.
Arthritis Res Ther. 2004;6(2):R169-80.
25. Olsen AK, Sondergaard BC, Byrjalsen I, Tanko LB, Christiansen
C, Muller A, et al. Anabolic and catabolic function of
chondrocyte ex vivo is reflected by the metabolic processing of
type II collagen. Osteoarthritis Cartilage. 2007;15(3):335-42.
26. Bay-Jensen AC, Andersen TL, Charni-Ben TN, Kristensen
PW, Kjaersgaard-Andersen P, Sandell L, et al. Biochemical
markers of type II collagen breakdown and synthesis are positioned
at specific sites in human osteoarthritic knee cartilage.
Osteoarthritis Cartilage. 2008;16(5):615-23.
27. Oettmeier R, Abendroth K. Osteoarthritis and bone: osteologic
types of osteoarthritis of the hip. Skeletal Radiol.
1989;18(3):165-74.
28. Qvist P, Bay-Jensen AC, Christiansen C, Dam EB, Pastoureau
P, Karsdal MA. The disease modifying osteoarthritis drug
(DMOAD): is it in the horizon? Pharmacol Res. 2008;58(1):1-7.
29. Keraplis AC. Embryonic development of bone and the molecular
regulation of intramembranous and endochondral bone
formation. In: Bilezikian JP, Raisz LG, Rodan GA, editors.
Principles of bone biology. Vol. 1. 2nd ed. San Diego: Academic
Press; 2002. p. 33-58.
30. von der MK, Kirsch T, Nerlich A, Kuss A, Weseloh G, Gluckert
K, et al. Type X collagen synthesis in human osteoarthritic
cartilage: indication of chondrocyte hypertrophy. Arthritis
Rheum. 1992;35(7):806-11.
31. Karsdal MA, Madsen SH, Christiansen C, Henriksen K,
Fosang AJ, Sondergaard BC. Cartilage degradation is fully
reversible in the presence of aggrecanase but not matrix metalloproteinase
activity. Arthritis Res Ther. 2008;10(3):R63.
32. Nilsson O, Marino R, De LF, Phillip M, Baron J. Endocrine
regulation of the growth plate. Horm Res. 2005;64(4):157-65.
33. Hutchison MR, Bassett MH, White PC. Insulin-like growth
factor-I and fibroblast growth factor, but not growth hormone,
affect growth plate chondrocyte proliferation. Endocrinology.
2007;148(7):3122-30.
34. Pass C, Macrae VE, Ahmed SF, Farquharson C. Inflammatory
cytokines and the GH/IGF-I axis: novel actions on bone
growth. Cell Biochem Funct. 2009;27(3):119-27.
35. Thomson BM, Mundy GR, Chambers TJ. Tumor necrosis factors
alpha and beta induce osteoblastic cells to stimulate osteoclastic
bone resorption. J Immunol. 1987;138(3):775-9.
36. Kobayashi K, Takahashi N, Jimi E, Udagawa N, Takami M,
Kotake S, et al. Tumor necrosis factor alpha stimulates osteoclast
differentiation by a mechanism independent of the ODF/
RANKL-RANK interaction. J Exp Med. 2000;191(2):275-86.
37. Weir EC, Horowitz MC, Baron R, Centrella M, Kacinski
BM, Insogna KL. Macrophage colony-stimulating factor

278 Cartilage 2(3)
release and receptor expression in bone cells. J Bone Miner Res.
1993;8(12):1507-18.
38. Felix R, Fleisch H, Elford PR. Bone-resorbing cytokines enhance
release of macrophage colony-stimulating activity by the osteoblastic
cell MC3T3-E1. Calcif Tissue Int. 1989;44(5):356-60.
39. Sato K, Kasono K, Fujii Y, Kawakami M, Tsushima T,
Shizume K. Tumor necrosis factor type alpha (cachectin)
stimulates mouse osteoblast-like cells (MC3T3-E1) to produce
macrophage-colony stimulating activity and prostaglandin
E2. Biochem Biophys Res Commun. 1987;145(1):323-9.
40. Heymann D, Rousselle AV. gp130 cytokine family and bone
cells. Cytokine. 2000;12(10):1455-68.
41. Tamura T, Udagawa N, Takahashi N, Miyaura C, Tanaka S,
Yamada Y, et al. Soluble interleukin-6 receptor triggers osteoclast
formation by interleukin 6. Proc Natl Acad Sci U S A.
1993;90(24):11924-8.
42. Richards CD, Langdon C, Deschamps P, Pennica D, Shaughnessy
SG. Stimulation of osteoclast differentiation in vitro by
mouse oncostatin M, leukaemia inhibitory factor, cardiotrophin-1
and interleukin 6: synergy with dexamethasone. Cytokine. 2000;
12(6):613-21.
43. Wang Y, Nishida S, Elalieh HZ, Long RK, Halloran BP, Bikle
DD. Role of IGF-I signaling in regulating osteoclastogenesis.
J Bone Miner Res. 2006;21(9):1350-8.
44. Chihara K, Sugimoto T. The action of GH/IGF-I/IGFBP in
osteoblasts and osteoclasts. Horm Res. 1997;48(Suppl 5):
45-9.
45. Karsdal MA, Sumer EU, Wulf H, Madsen SH, Christiansen C,
Fosang AJ, et al. Induction of increased cAMP levels in articular
chondrocytes blocks matrix metalloproteinase-mediated
cartilage degradation, but not aggrecanase-mediated cartilage
degradation. Arthritis Rheum. 2007;56(5):1549-58.
65

CHAPTER 4: PAPER I
66

CHAPTER 5
CHAPTER 5: PAPER II
PAPER II
___________________________________________________________________________
Introductory remarks
This manuscript shows the effect of glucocorticoids on bone and/or cartilage, performed in pre-
clinical models. The novel ex vivo femoral head model (PAPER I) was used to investigate the
effect of glucocorticoids on a bone/cartilage unit, whereas the articular cartilage ex vivo model and
the osteoclast and osteoblast cell cultures were used to investigate the effects on either bone or
cartilage. This manuscript highlights the importance of the individual cell type and verifies the
novel ex vivo femoral head model as a useful screening model for potential drugs.
67

Glucocorticoids exert context-dependent effects on cells of the joint in vitro
Suzi H. Madsen a,⇑ , Kim V. Andreassen b , Søren T. Christensen c , Morten A. Karsdal a ,
Francis M. Sverdrup d , Anne-Christine Bay-Jensen a , Kim Henriksen b
a Cartilage Biology and Biomarkers, Nordic Bioscience A/S, Herlev Hovedgade 207, DK-2730 Herlev, Denmark
b Bone Biology, Nordic Bioscience A/S, Herlev Hovedgade 207, DK-2730 Herlev, Denmark
c Department of Biology, The August Krogh Building, University of Copenhagen, Universitetsparken 13, DK-2100 Copenhagen OE, Denmark
d Center for World Health and Medicine, Saint Louis University, Doisy Research Rm 355, 1205 Carr Ln, Saint Louis, MO 63104, USA
article info
Article history:
Received 21 June 2011
Received in revised form 27 July 2011
Accepted 28 July 2011
Available online 10 August 2011
Keywords:
Prednisolone
Dexamethasone
Chondrocytes
Osteoclasts
Osteoblasts
Ex vivo model
1. Introduction
abstract
Glucocorticoids (GCs) are used to overcome inflammatory conditions,
such as inflammatory bowel diseases and rheumatoid
Abbreviations: IGF-I, insulin-like growth factor 1; BMP-2, bone morphogenetic
protein 2; OSM, oncostatin M; TNF-a, tumor necrosis factor-a; GC, glucocorticoid;
OA, osteoarthritis; RANKL, receptor activator of nuclear factor kappa-B ligand; OPG,
osteoprotegerin; PRED, prednisolone; DEX, dexamethasone; MMP, matrix metalloproteinase;
PBS, phosphate buffered saline; M-CSF, macrophage colony-stimulating
factor; a-MEM, a-modified eagle medium; ELISA, enzyme-linked immunosorbent
assay; TRAP, tartrate-resistant acid phosphatase; sGAG, sulfated glycosaminoglycans.
⇑ Corresponding author. Tel.: +45 44 52 52 52; fax: +45 44 52 52 51.
E-mail addresses: shm@nordicbioscience.com (S.H. Madsen), kan@nordicbioscience.com
(K.V. Andreassen), stchristensen@bio.ku.dk (S.T. Christensen), mk@nordicbioscience.com
(M.A. Karsdal), fsverdrup@charter.net (F.M. Sverdrup),
acbj@nordicbioscience.com (A.-C. Bay-Jensen), kh@nordicbioscience.com
(K. Henriksen).
0039-128X/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2011.07.018
Steroids 76 (2011) 1474–1482
Contents lists available at SciVerse ScienceDirect
Steroids
journal homepage: www.elsevier.com/locate/steroids
Introduction: Glucocorticoids are known to attenuate bone formation in vivo leading to decreased bone
volume and increased risk of fractures, whereas effects on the joint tissue are less characterized. However,
glucocorticoids appear to have a reducing effect on inflammation and pain in osteoarthritis. This
study aimed at characterizing the effect of glucocorticoids on chondrocytes, osteoclasts, and osteoblasts.
Experimental: We used four model systems to investigate how glucocorticoids affect the cells of the joint;
two intact tissues (femoral head- and cartilage-explants), and two separate cell cultures of osteoblasts
(2T3-pre-osteoblasts) and osteoclasts (CD14 + -monocytes). The model systems were cultured in the presence
of two glucocorticoids; prednisolone or dexamethasone. To induce anabolic and catabolic conditions,
cultures were activated by insulin-like growth factor I/bone morphogenetic protein 2 and
oncostatin M/tumor necrosis factor-a, respectively. Histology and markers of bone- and cartilage-turnover
were used to evaluate effects of glucocorticoid treatment.
Results: Prednisolone treatment decreased collagen type-II degradation in immature cartilage, whereas
glucocorticoids did not affect collagen type-II in mature cartilage. Glucocorticoids had an anti-catabolic
effect on catabolic-activated cartilage from a bovine stifle joint and murine femoral heads. Glucocorticoids
decreased viability of all bone cells, leading to a reduction in osteoclastogenesis and bone resorption;
however, bone morphogenetic protein 2-stimulated osteoblasts increased bone formation, as
opposed to non-stimulated osteoblasts.
Conclusions: Using highly robust in vitro models of bone and cartilage turnover, we suggest that effects of
glucocorticoids highly depend on the activation and differential stage of the cell targeted in the joint.
Present data indicated that glucocorticoid treatment may be beneficial for articular cartilage, although
detrimental effects on bone should be taken into account.
Ó 2011 Elsevier Inc. All rights reserved.
68
arthritis [1]. In addition, intra-articular injections of GCs have been
tested for their ability to reduce inflammation and pain in
osteoarthritis (OA) and appear to have a reducing effect on both
parameters [2]. Furthermore, OA chondrocytes express fewer glucocorticoid
receptors (GR) than normal chondrocytes, indicating
the importance of glucocorticoid stimulation of normal cartilage
[3]. However, GC treatment in children and adolescents is associated
with severe growth retardation caused by apoptosis of the
growth plate cartilage, especially in the proliferative zones of the
cartilage [4,5]. With respect to the effects on articular chondrocytes,
in vitro and ex vivo studies of chondrocytes are contradicting
and highly context dependent [6–9].
Unfortunately, prolonged GC therapy has also been associated
with bone loss and fractures, resulting in severe osteoporosis
[10,11]. In this scenario, GCs cause significant reductions in bone
formation, as well as a transient acceleration of bone resorption,

leading to fragility fractures [12], which especially occur in highturnover
bone compartments, such as vertebrae [13]. In vivo, this
effect is mediated via induction of osteoblast and osteocytes
apoptosis, as well as a less well-understood activation of osteoclasts
[1,14–17]. Interestingly, a mouse study showed that the
detrimental effect of GCs on bone formation was mediated
through the osteoclasts; whether it was dependent on bone
resorption was not clear [17]. In vitro studies of the effects of
GCs on osteoclasts have shown highly context dependent effects,
i.e., both inhibitory and activating effects were observed [17–19].
One possible explanation for these contradicting results may lay
within the heterogeneity of the model systems used, which comprise
both mouse and human cells in either purified or mixed cell
populations [17–19]. In osteoblasts and their precursors, GCs are
known to stimulate Receptor Activator of Nuclear Factor Kappa-B
ligand (RANKL), while reducing osteoprotegerin (OPG) [18],
potentially explaining the transient activation of bone resorption
observed in vivo. GCs also possess the ability to stimulate in vitro
osteoblastogenesis in cell cultures [20,21], although separate
studies show that GCs may arrest osteoblastogenesis and introduce
apoptosis [22,23], illustrating the complex and context
dependent nature of GCs.
Finally, GC signaling in osteoblasts and osteocytes was shown to
be involved in the pathogenesis of inflammatory arthritis, demonstrating
the importance of cellular cross-talk in the joint in relation
to effects of GCs [24]. So, whether GCs administered in vivo affect
both bone and cartilage, or just bone, is presently not clear
[25,26]. Especially in OA, a disease of both bone and cartilage
[27,28], the effects of GCs in the context of a whole joint need to
be studied in much further detail.
In this study, we used four highly robust cell systems, namely
murine femoral head explants [29], bovine articular cartilage explants
[30], CD14 + derived human osteoclasts [31] and the 2T3
osteoblast cell line [32], to characterize the effects of the GCs –
prednisolone (PRED) and dexamethasone (DEX) – on cells of the
joint. Furthermore, we cultured the cells under different conditions
to shed light on the context dependency of the effects of
GCs.
2. Experimental
2.1. Ex vivo models
2.1.1. Murine femoral heads
The experiment was performed in accordance with relevant
guidelines and regulations from the Ethical Committee, as mice
were euthanized before the experiment started. Femoral heads
from 12-week-old outbred Naval Medical Research Institute
(NMRI) female mice (Charles River, Germany) were isolated and
cultured accordingly to Madsen et al. [29]. Femoral heads were cultured
for 3 weeks, in the presence or absence of different factors:
the anabolic factor Insulin-like Growth Factor I (IGF-I) [100 ng/
ml] (Sigma, DK), the catabolic cytokines Oncostatin M (OSM)
[10 ng/ml] (Sigma, DK) + Tumor Necrosis Factor-a (TNF-a)
[20 ng/ml] (R&D Systems, UK) and PRED [100 nM] (Sigma, DK).
2.1.2. Bovine articular cartilage
Bovine knees were bought at the local slaughter house
(Slangerup, DK) just after slaughtering. Articular cartilage explants
were isolated and cultured accordingly to our previous
studies [30,33]. The explants were cultured for 3 weeks, in the
presence or absence of different factors: OSM [10 ng/ml] + TNFa
[20 ng/ml], IGF-I [100 ng/ml], PRED [100 nM] and DEX
[100 nM] (Sigma, DK).
S.H. Madsen et al. / Steroids 76 (2011) 1474–1482 1475
69
2.2. In vitro models
2.2.1. Human osteoclast preparation
Human blood was donated from the Danish University Hospital
of Copenhagen (has an approval from the Ethical Committee to donate
blood to us), and CD14 + monocytes were isolated and differentiated
into osteoclasts accordingly to Sorensen et al. [31].
2.2.2. Differentiating osteoclasts
The cells were seeded on plastic or bovine cortical bone slices
(Nordic Bioscience Diagnostics, DK) at a density of 100,000 cells/
well in a 96-well plate and cultured in phenol red free a-MEM containing
10% serum, P/S, M-CSF [25 ng/ml] (R&D System, DK),
RANKL [25 ng/ml] (R&D Systems, DK), either in the presence or absence
of DEX [100 nM] (Sigma, DK). Incubation was performed at
37 °C and 5% CO2. The media were changed every 2nd or 3rd day
and saved at 20 °C.
2.2.3. Mature osteoclasts
For experiments with mature human osteoclasts, cells were
grown in 75 cm 2 culture flasks for 10 days in phenol red free a-
MEM containing 10% serum, P/S, M-CSF [25 ng/ml], and RANKL
[25 ng/ml] to induce osteoclastogenesis until the cells had 3–5 nuclei
on average. The cells were washed with PBS twice, trypsin was
added, and the cells were incubated at 37 °C for approximately
20 min. The cells were scraped off, and re-seeded on plastic or bovine
cortical bone slices at a density of 40,000 cells/well in a 96well
plate in the presence of phenol red free a-MEM containing
10% serum, P/S, M-CSF [25 ng/ml], RANKL [25 ng/ml] in the presence
or absence of DEX [100 nM] at 37 °C and 5% CO2. The media
were changed every 2nd or 3rd day and saved at 20 °C.
2.2.4. Murine osteoblasts
2T3 pre-osteoblast cells were purchased from Dr. Steven E. Harris
at the University of Texas Health Science Center at San Antonio.
The murine mesenchymal stem cell line, 2T3, is able to differentiate
into osteoblasts under proper predetermined conditions. Osteoblasts
were prepared accordingly to Ghosh-Choudhury et al. [32].
To induce bone nodule formation, the 2T3 cells were lifted by trypsin
and reseeded in 24-well plates at a density of 10,000 cells/well,
and then cultured in the presence of ascorbic acid [100 lg/ml] and
b-glycerol-phosphate [4 mM], in the presence or absence of BMP-2
[30 ng/ml] or DEX [100 nM] for 10 days. Incubation was performed
at 37 °C, 5% CO2. The media were replaced every 2nd or 3rd day
and saved at 20 °C.
2.3. Concentrations
The concentrations of IGF-I, BMP-2 and OSM + TNF-a were selected
accordingly to previous studies [29–31,33,34]. The concentrations
of DEX and PRED were selected accordingly to previous
studies [5,6,17].
2.4. Cell-survival after culture-period
The viability of the cells (in explants or isolated cells) were measured
at the end of the culture period, using the colorimetric dye,
alamar blue (Invitrogen, DK), as a 10% solution in the conditioned
media, accordingly to the manufacturer’s instructions [35].
2.5. Biomarkers
The resorption marker CTX-I (RatLaps Ò ), the bone formation
marker Rat/Mouse PINP EIA and the cartilage formation marker
PIINP (IDS Ltd., UK) were measured accordingly to the manufacturer’s
instructions [33]. The competitive ELISA CIIM (Nordic

1476 S.H. Madsen et al. / Steroids 76 (2011) 1474–1482
Bioscience, DK) was used to quantify MMP-mediated degradation
products of type II collagen, accordingly to Bay-Jensen et al. [36].
342-G2 measures MMP-mediated aggrecan degradation accordingly
to Sumer et al. [37].
2.6. Quantification of osteoclast number
The TRAP assay was performed as previously described [31].
2.7. Measurements of alkaline phosphatase activity
The culture supernatants were added to a substrate mixture of
95% AMP buffer (pH 10), 10.1 mM phosphatase substrate, 18 mM
MgCl2, and MilliQ in a ratio of 1:4 and incubated in the dark for
20 min. After incubation, the reaction was stopped suing 0.5 M
NaOH, and finally the absorbance at 405 nm was measured using
an ELISA reader.
2.8. Resorption assays
The concentration of total calcium was measured in culture
supernatants after resorption, by using a colorimetric assay and a
Hitachi 912 Automatic Analyzer (Roche Diagnostics) following
the assay method validated and warranted by Roche Diagnostics
[38].
2.9. Alizarin Red-S staining of mineralized bone nodules
and extraction of dye
After 10 days of culture, the osteoblast cultures were stopped
by fixation in 4% formaldehyde for 20 min. The cells were stained
in 40 mM Alizarin Red-S, pH 4.2, for 10 min, and washed thoroughly.
The dye was extracted from the cells by incubation in
10% cetylpyridinium chloride for 15 min while shaking in the dark,
and finally the absorbance at 561 nm was measured using an ELISA
reader.
2.10. Histologic analysis
The bovine articular cartilage explants were fixed in 4% formaldehyde,
processed, paraffin embedded, and subsequently sectioned
(5 lm). Proteoglycans were stained after a standard toluidine blue
protocol [39] after deparaffinization and rehydration.
2.10.1. Image analysis
Digital images of stained sections were displayed on a computer
monitor and captured using an Olympus BX-60 microscope
equipped with a 60 objective and digital camera.
2.11. Statistic analysis
Results are shown as mean + standard error of mean (SEM). Statistic
analysis was performed by the Student’s two-tailed unpaired
t-test assuming normal distribution and equal variance. Statistic
significance is indicated by the number of asterisks, ⁄ p < 0.05,
⁄⁄ p < 0.01, and ⁄⁄⁄ p < 0.001.
3. Results
3.1. Prednisolone decreases cartilage degradation and bone
turnover in femoral heads ex vivo
Prednisolone (PRED) treatment for 2 weeks decreased overall
cell viability in femoral heads by 40% (p < 0.01) compared to
the vehicle control (Fig. 1A). Cell viability was not affected by
70
OSM + TNF-a stimulation (catabolic control) (Fig. 1A). PRED in
combination with OSM + TNF-a showed a tendency towards decreased
viability compared to the catabolic control (Fig. 1A).
After 1 week, PRED caused a reduction in the number of osteoclasts
in both resting and catabolic activated femoral heads, compared
to the respective vehicle and catabolic controls (Fig. 1B).
After 2 weeks in culture, osteoclast number was decreased to a
low level in the vehicle control and no differences were detected
with any of the treatment (Fig. 1B). After 1 week, PRED treatment
decreased bone resorption by 75% (p < 0.01) (Fig. 1C), consistent
with the reduced number of osteoclasts (Fig. 1B). Bone resorption
increased by 5.5-fold (p < 0.01) after 1 week when stimulated
with OSM + TNF-a, whereas this increase in bone resorption decreased
by 90% (p < 0.05) in the presence of PRED (Fig. 1C). The
increase in bone resorption mediated by OSM + TNF-a after 1
week, was not seen after 2 weeks (Fig. 1C). PRED treatment decreased
bone formation by 50% (p < 0.01) after 1 week, and
90% (p < 0.01) after 2 weeks, compared to the respective vehicles
(Fig. 1D). Furthermore, OSM + TNF-a decreased bone formation by
75% (p < 0.01) after 1 week, compared to the vehicle (Fig. 1D).
Treatment with PRED in combination with OSM + TNF-a did not
significantly alter bone formation compared to the catabolic
control; however, the combination seemed to have a tendency to
decrease bone formation slightly (Fig. 1D). After 2 weeks,
OSM + TNF-a stimulation inhibited bone formation, compared to
the vehicle (Fig. 1D). With respect to cartilage metabolism, PRED
treatment showed at tendency to decrease cartilage degradation
in non-stimulated femoral heads after 1 week (by 75%) and significantly
by 60% (p < 0.05) after 2 weeks (Fig. 1E). OSM + TNF-a
stimulation increased the cartilage degradation by 33-fold
(p < 0.01) after 1 week and 30-fold (p < 0.01) after 2 weeks, compared
to respective vehicles (Fig. 1E). These catabolic-mediated increases
in cartilage degradation were both inhibited by the
presence of PRED ( 95% (p < 0.05)) after 1 and 2 weeks (Fig. 1E).
The formation of cartilage did not significantly vary between vehicle
and stimulated explants after 1 week (Fig. 1F). However,
OSM + TNF-a stimulation inhibited cartilage formation by 95%
(p < 0.05) after 2 weeks of culture (Fig. 1F). While PRED itself did
not affect cartilage formation at one or 2 weeks, it failed to prevent
the inhibition caused by 2 weeks of catabolic stimulation (Fig. 1F).
3.2. Prednisolone does not decrease the formation of bone and
cartilage in anabolic activated femoral heads
Stimulation with IGF-I increased viability of the femoral heads
after 2 weeks of culture by 1.6-fold (p < 0.01) (Fig. 2A). PRED in
combination with IGF-I did not affect the viability compared to
the anabolic control (IGF-I alone) (Fig. 2A). Due to the low level
of TRAP, we were not able to detect any differences in the number
of osteoclasts by IGF-I or IGF-I + PRED treatment compared to
respective vehicles (Fig. 2B). However, IGF-I stimulation decreased
bone resorption by 70% (p < 0.05) after 1 week, which was further
decreased by addition of PRED (Fig. 2C). PRED also decreased the
IGF-I-reduced resorption by 75% (p < 0.05) after 2 weeks
(Fig. 2C). The anabolic-stimulated femoral heads showed increased
bone formation by 2.1-fold (p < 0.01) after 1 week and 13.4-fold
(p < 0.01) after 2 weeks (Fig. 2D). PRED did not significantly decrease
anabolic activated bone formation, although a downward
trend was observed after 2 weeks (Fig. 2D). IGF-I-stimulated femoral
heads showed very low levels of cartilage degradation after 1
and 2 weeks, due to the overall low level of degradation in this setup.
However, after 1 week, IGF-I stimulation decreased by 72%
(p < 0.05) compared to vehicle. Addition of PRED to the anabolic
control showed the same low levels of cartilage degradation compared
to the anabolic controls (Fig. 2E). Cartilage formation, on the
other hand, showed a tendency to increase with IGF-I stimulation

after 1 week (Fig. 2F). The addition of PRED to IGF-I stimulation
was however not different from the vehicle. On the contrary, after
2 weeks, PRED in combination with IGF-I showed a tendency to increase
cartilage formation compared to the anabolic control
(Fig. 2F).
3.3. Cytokine-induced cartilage turnover is decreased
by glucocorticoid treatment
The effect of PRED was tested on bovine cartilage explants.
However, since the effects of PRED were weak, we also tested the
more potent dexamethasone (DEX) [8] in a separate experiment.
The effects of PRED and DEX were measured at day 21.
After 21 days of culture, the viability of the bovine cartilage explants
increased with OSM + TNF-a stimulation by 1.6-fold
(p < 0.05) (Fig. 3A) and 1.7-fold (p < 0.01) (Fig. 3E), compared to
respective vehicles (Fig. 3A and E). However, this increase seemed
inhibited by DEX (p < 0.05) (Fig. 3E). Aggrecan degradation was induced
by OSM + TNF-a by more than 150-fold (p < 0.001) for both
experiments, and reduced by addition of DEX by 70% (p < 0.01);
however not to the level of the vehicle or DEX alone (Fig. 3B and
F). The same pattern was observed for collagen type II degradation,
where OSM + TNF-a stimulation increased the degradation by
more than 40-fold (p < 0.01), which was reduced by addition of
PRED and significantly by addition of DEX by 75% (p < 0.01)
(Fig. 3C and G). Aggrecan and collagen degradation were also measured
at days 5 and 12, but did not show any differences between
the vehicle and activated explants (data not shown). OSM + TNF-a
stimulation increased cartilage formation by more than 4-fold
(p < 0.01) in both experiments (Fig. 3D and H). Cartilage formation
S.H. Madsen et al. / Steroids 76 (2011) 1474–1482 1477
Fig. 1. The effects of prednisolone in resting and catabolic-induced bone and cartilage from femoral heads. Femoral heads were cultured in the presence of 100nM
prednisolone (PRED) and/or cytokines (O + T) [10 ng/ml] + [20 ng/ml] for 2 weeks. Cell viability was measured with alamar blue after 2 weeks (A). Biomarkers were measured
after 1 week (black bars) and 2 weeks (gray bars) (B–F). The number of osteoclasts was measured with TRAP activity (B). The bone resorption and formation were measured
by the collagen type I markers, CTX-I (C) and PINP (D), respectively. Cartilage degradation and formation were measured by the collagen type II markers, CIIM (E) and PIINP
(F), respectively. PRED [100 nM] treatment decreased viability of the cells in the femoral heads, osteoclast number, bone resorption, bone formation and cartilage degradation.
Each treatment n P 4.
71
was also induced by IGF-I stimulation by more than 8.5-fold
(p < 0.01) in both experiments (Fig. 3D and H). DEX inhibited formation
induced by OSM + TNF-a by 50% (p < 0.05) (Fig. 3H). PRED
or DEX by themselves could not induce cartilage formation
although there was a tendency towards increased formation when
combined with IGF-I (Fig. 3D and H). Generally, we saw the same
pattern for both GCs. However, PRED was less potent than DEX
as observed in the literature [8].
The biomarker data for aggrecan degradation were partly supported
by histology of the explants, using toluidine blue to detect
sulfated glycosaminoglycans (sGAGs) (Fig. 3I–T). The controls
worked; as shown in previous studies [30], stimulation with
OSM + TNF-a depleted the cartilage explants of sGAGs (Fig. 3J and
P), whereas IGF-I stimulation increased the content of sGAGs
(Fig. 3K and Q), compared to respective vehicles (Fig. 3I and O).
The cartilage explants treated with either PRED or DEX seemed to
have less sGAGs retained in the matrixes (Fig. 3L and R) than the
vehicle explants (Fig. 3I and O). On the other hand, when adding
DEX to OSM + TNF-a-stimulated cartilage explant, a small portion
of the sGAGs seemed to be retain in the matrix (Fig. 3S) compared
to the OSM + TNF-a-stimulated explant (Fig. 3P), although the visible
difference is almost lost on the digital images. Furthermore,
addition of PRED or DEX to the IGF-I-stimulated cartilage explants
seemed to increase the content of sGAGs (Fig. 3N and T).
3.4. The effect of dexamethasone depends on the activation
stage of osteoblasts
To investigate the effects of GCs on bone cells, the more potent
DEX [8] was used for further experiments/studies (Fig. 3). When

1478 S.H. Madsen et al. / Steroids 76 (2011) 1474–1482
Fig. 2. The effects of prednisolone in anabolic-induced bone and cartilage from femoral heads. Femoral heads were cultured in the presence of 100 nM prednisolone (PRED)
and/or 100 ng/ml IGF-I for 2 weeks. Cell viability was measured with alamar blue after 2 weeks (A). Biomarkers were measured after 1 week (black bars) and 2 weeks (gray
bars) (B–F). The number of osteoclasts was measured with TRAP activity (B). The bone resorption and formation were measured by the collagen type I markers, CTX-I (C) and
PINP (D), respectively. Cartilage degradation and formation were measured by the collagen type II markers, CIIM (E) and PIINP (F), respectively. PRED [100 nM] + IGF-I
[100 ng/ml] treatment decreased bone resorption, but not bone formation. Furthermore, PRED [100 nM] + IGF-I [100 ng/ml] treatment increased cartilage formation after 2
weeks. Each treatment n P 4.
osteoblasts were cultured in the presence of DEX, cell viability was
reduced by 55% (p < 0.05), compared to the vehicle (Fig. 4A). The
same pattern was seen in the presence of both DEX and BMP-2,
where viability decreased by 35% (p < 0.05), compared to BMP-
2 stimulation alone (Fig. 4A). DEX also decreased the nodule formation
by 65% (p < 0.05), compared to vehicle (Fig. 4B). Nodule
formation was induced by addition of the positive control [34],
BMP-2, by 2.4-fold (p < 0.05) compared to vehicle (Fig. 4B). Furthermore,
addition of DEX to BMP-2-stimulated osteoblasts, led
to a further increase in nodule formation by 1.8-fold (p < 0.05)
(Fig. 4B). ALP production by the osteoblasts was reduced when cultured
with DEX, both in the presence (by 45%) or absence of BMP-
2 (by 85%) (p < 0.05) (Fig. 4C), correlating with the cell viability
measurements (Fig. 4A).
3.5. Dexamethasone decreases viability and function of both
mature osteoclasts and their precursors
When pre-osteoclasts were cultured either on bone or plastic,
DEX reduced the viability of the cells by more than 95%
(p < 0.01), as well as the osteoclast marker TRAP by more than
98% (p < 0.01) (Fig. 5A and B). Furthermore, the release of calcium
during resorption of bone, decreased by 85% (p < 0.01) when treated
with DEX (Fig. 5C). To investigate whether this effect was specific
to RANKL induction of osteoclastogenesis, an experiment
omitting RANKL was conducted [31]. In these cells, we again observed
a reduction in cell viability in the presence of DEX (data
not shown), thus showing that the reduction in viability is related
to the macrophages rather than specifically to the osteoclasts. In
mature osteoclasts, DEX decreased cell viability by 80%
72
(p < 0.01) on both bone and plastic (Fig. 5D). The number of osteoclasts
was decreased by 70% (p < 0.01) on bone and 75%
(p < 0.01) on plastic (Fig. 5E). Bone resorption decreased by 40%
(p < 0.05) (Fig. 5F). Interestingly, the overall level of suppression induced
by the DEX on mature osteoclasts was lower than the level
of suppression induced in the pre-osteoclasts.
4. Discussion
Intra-articular corticosteroid injections have been tested for
beneficial effects on inflammation and joint destruction, and appear
to have some applicability by reducing pain in OA patients
[2,40,41]. However, how GCs affect the individual cell types in
the joint is not known. Thus, this study focused on elucidating
whether GCs have any effects on cartilage turnover. Furthermore,
since subchondral bone has been shown to play an essential role
in the development of OA, and since the effects of GCs on bone cells
are still discussed [27,42], we extended the study by performing a
thorough characterization of the effect of GCs on the two major
bone cell types, namely osteoblasts and osteoclasts. This study
used four model systems to investigate how GCs affect the cells
of the joint both in the setting of intact tissues, as femoral heads
or cartilage explants (resting, catabolically active or anabolically
active), and as separate cell cultures of osteoblasts (resting or differentiating)
and osteoclasts (differentiating or mature and on
plastic or bone).
Several studies of GC-mediated effects on chondrocyte function
have been published; however no clear-cut conclusion has been
presented. These studies have indicated both beneficial and detrimental
effects of GCs on cartilage, [6,7] however, the studies also

use a range of assays ranging from isolated chondrocytes through
ex vivo explants to mesenchymal stem cell differentiation and alginate
beads. In the present study, we used a murine ex vivo femoral
head model [29], and found that PRED decreased bone turnover
S.H. Madsen et al. / Steroids 76 (2011) 1474–1482 1479
Fig. 3. The effects of prednisolone and dexamethasone in resting and activated cartilage. Bovine articular cartilage explants were cultured in the presence of 100 nM
prednisolone (PRED), 100 nM dexamethasone (DEX), and/or cytokines (O + T) [10 ng/ml] + [20 ng/ml] (gray bars) and/or IGF-I [100 ng/ml] (light gray bars) for 3 weeks. Cell
viability was measured with alamar blue after 3 weeks (A and E) and biomarkers were measured at day 19 (B–D and F–H). Explants cultured for 3 weeks were stained with
toluidine blue to evaluate proteoglycan content (I–N and O–T). MMP-mediated aggrecan degradation was measured with 342-G2 (B and F). Collagen type II degradation and
formation were measured by the markers, CIIM (E) and PIINP (F), respectively. PRED/DEX [100 nM] treatment did not affect the cartilage. However, cytokine-induced
degradation were attenuated by 100 nM PRED/DEX. PRED [100 nM] was less potent than DEX [100 nM]. Each treatment n P 5.
Fig. 4. The effect of dexamethasone in non-stimulated and anabolic activated osteoblasts. Osteoblasts were cultured in the presence or absence of 100 nM dexamethasone
(DEX), in non-stimulated (black bars) or 30 ng/ml BMP-2 stimulated (gray bars) cells for 10 days. Cell viability was measured with alamar blue after 10 days (A) and bone
formation was measured by extraction of Alizarin Red-S from stained mineralized bone nodules at day 10 (B). Osteoblastogenesis was measured by alkaline phosphatase
activity in the supernatants accumulated over 10 days (C). DEX [100 nM] treatment decreased viability, bone formation and osteoblastogenesis. However, DEX [100 nM]
increased bone formation in BMP-2 [30 ng/ml] stimulated osteoblasts (B). Each treatment n P 4.
73
while mediating anti-catabolic effects on cartilage. Whether osteoclasts,
osteoblasts and chondrocytes were affected directly or indirectly
by PRED, were investigated by separate cell or explants
cultures. To evaluate the effect of GCs directly on chondrocytes,

1480 S.H. Madsen et al. / Steroids 76 (2011) 1474–1482
Fig. 5. The effect of dexamethasone in pre-osteoclasts and mature osteoclasts. Osteoclasts were cultured in on either bone (black bars) or plastic (gray bars) in the presence or
absence of 100 nM dexamethasone (DEX). Pre-osteoclasts were cultured for 21 days (A–C) and mature osteoclasts for 14 days (D–F). Cell viability was measured with alamar
blue after 14 or 21 days (A and D). The number of osteoclasts was measured with TRAP activity (B and E) and bone resorption was measured by calcium release from bone (C
and F). DEX [100 nM] treatment decreased cell viability, osteoclast number and bone resorption in both pre-osteoclasts and mature osteoclasts. However, the effects of DEX
[100 nM] on pre-osteoclasts were higher than the effects on mature osteoclasts. Each treatment n P 5.
we used the bovine articular cartilage explants model that is presently
the preferred model for studying cartilage turnover ex vivo
[30,33]. The strong indications that OSM + TNF-a reduced the
ability to induce cartilage degradation in the presence of DEX or
PRED, evaluated by biomarkers and histology, suggest a chondroprotective
effect by GCs independent of bone cells. Furthermore,
histology showed that GC treatment had a positive effect on proteoglycans
under anabolic conditions, which is in line with a study
by Van der Kraan et al. that showed that PRED stimulates proteoglycan
synthesis in patellar cartilage, validated by [ 35 S] incorporation
[43]. However, the catabolic stimulated cartilage explants in
the presence or absence of DEX, were both close to be depleted
for proteoglycans, measured by the intensity of color (sGAG),
which was not as convincing as the 70% reduction of MMP-mediated
aggrecan degradation shown with the biomarker, 342-G2.
However, depletion of color may results from an early loss of aggrecanase-mediated
highly-sGAG-saturated aggrecan degradation,
leaving low-sGAG-saturated aggrecan to be degraded by MMPs at
day 19 [44], as we did not see any MMP-mediated degradation at
days 5 and 12. This suggests that the anti-catabolic effects of GC
on cartilage may be limited to the degradation mediated by MMPs.
Interestingly, GC treatment in resting conditions resulted in a small
loss of proteoglycans from the articular cartilage explants that we
were unable to detect using the MMP-mediated aggrecan degradation
marker, 342-G2. Thus, the two opposite effects of GCs treatment
(protecting in activated cartilage and degrading in resting
cartilage) indicate that the activation stage of the chondrocytes is
important in relation to the GC response. If the GC-mediated protective
effect results from an alteration of the chondrocyte phenotype
(preventing the induction/release of proteases) or that the
74
GCs inhibit the cytokine pathway, still remains unclear. Yet, our
data indicate that a reduction of MMP activity is involved, as the
MMP-dependent biomarkers, 342-G2 and CIIM, are reduced. This
is in line with a recent study by Garvican et al. suggesting that
GCs reduce MMP-mediated cartilage degradation [9], although
other enzymatic pathways, such as the u-PA/plasminogen/plasmin-pathway,
also have been indicated [45]. Thus, in the search
for new disease modifying osteoarthritic drugs (DMOADs), we suggest
that the inhibition of protease pathways mediated by GCs
could be of interest.
We did not see any protective effects of GCs on the non-stimulated
bovine cartilage explants, as opposed to the results with cartilage
from the femoral head explants. The bovine cartilage
explants represent ‘‘healthy’’ cartilage, whereas the femoral heads
consist of cartilage in development, which contains more proliferating,
pre-hypertrophic and hypertrophic chondrocytes (which are
observed in OA cartilage) than the bovine cartilage explants [29].
This could explain why we have to induce a catabolic milieu with
cytokines in the bovine cartilage explants, before we see a protective
effect of GCs, and suggest that GCs mainly affect chondrocytes
which are activated, whereas resting chondrocytes appear to respond
less. We do acknowledge that we deal with cross-species
comparisons, which may cause the different effects observed in
this study. However, the anabolic and catabolic controls have the
same effects on the different species, suggesting the effects of
GC, or lack of same, are due to action from the GC and not a species-dependent
effect.
The present study clearly shows that GCs have demonstrably
large effects on cells of bone. While it for long has been known that
GCs induce apoptosis of osteoblasts and osteocytes in vivo, there is

still some controversy regarding the mechanism of action. Some
data indicate a role for osteoclasts and others data indicate direct
effects on osteoblasts [17,46]. Controversially, a recent study
showed that the endogenous GC-induced signaling in osteoblasts
is needed for proper bone formation in vivo in mice [47]. In vitro
data regarding bone formation are also confusing, as data showing
that GCs support stimulation of RANKL production, increase nodule
formation, as well as inhibiting bone formation [18,20,21,48]. Our
data shows that GC-mediated effects on osteoblasts in cell cultures
are dependent on the ‘‘activation state’’ of the osteoblasts, with a
toxic effect on non-activated cells, and a pro-osteoblastic effect
on BMP-2-induced cells. This is in line with the results from the
femoral heads, which decreased bone formation with PRED treatment,
but increased bone formation in combination with IGF-I.
The conflicting in vitro and in vivo data could indicate that we need
to look more closely into GCs-mediated effects in the different
functional subgroups of osteoblasts, and more importantly, how
to modulate these with respect to preventing GC-induced bone
damage [48,49].
Several studies have indicated that GCs stimulate osteoclastogenesis
in vitro [50,51]; however, most of these studies were conducted
in the presence of stromal/osteoblastic cells, which are
known to produce RANKL in response to GCs [14,18]. We found
that in pure osteoclast precursor cell cultures, GCs inhibited osteoclastogenesis
by reducing the viability of the cells, which is in line
with the decreased viability and osteoclast number found in PREDstimulated
femoral heads. These data correlate to data showing decreased
osteoclastogenesis in vivo in mice, and strongly indicate
that the primary effect of GCs on osteoclastogenesis is inhibitory
[15,16]. However, an interesting observation in this regard is that
the effect of GCs on the mature osteoclasts, on cell viability, osteoclast
number and resorption, was smaller than the effect on the differentiating
osteoclasts. This could indicate that osteoclasts and
their precursors have different levels of sensitivity to GC treatment.
As all osteoclast cultures are a mix of mature multinucleated
osteoclasts and their precursor cells, and the ratio of mature to
pre-osteoclastic cells is dependent on several different aspects,
such as density, donor, serum, etc. [31] we speculate that the reason
for the predominantly toxic effect could be that we have more
osteoclast precursors present in the cultures than Soe and Delaisse,
who described a prolongation of the resorption cycle, when treating
mature osteoclasts derived from CD14 + cells with PRED [52].
A limitation to the study is that the effects observed with DEX
and PRED have not been tested for steroids specificity with a
non-GC control (e.g. estrogen or cholesterol). However, Kim et al.
[17] showed that GR°C / mice were spared the impact of DEX
on osteoclasts and their precursors, indicating a true GR-mediated
action of DEX. This steroid specificity is furthermore supported by
Battista et al. [3], which showed that the effect from DEX was
dependent on the percentage of the GR in the chondrocytes. Nonetheless,
future studies should include a non-GC control to ensure
that the effects observed are indeed specific to the GC class of
steroids.
In conclusions, we found that GCs did not appear to affect resting
chondrocyte lifespan and function. Furthermore, GCs showed a
protective effect on catabolic-induced cartilage and showed a tendency
to increase cartilage formation under anabolic conditions.
These results are encouraging for the use of intra-articular GC
injections in OA knees. However, considering the well-established
consequences on bone it is important to remember the tight functional
relationship between bone and cartilage under both physiologically
and pathologically conditions [27]. Thus, further studies
into the functional relationship between these compartments are
needed, especially with the focus on the possibility to rescue bone
damage by the addition of anabolic therapy, such as parathyroid
hormone [48].
S.H. Madsen et al. / Steroids 76 (2011) 1474–1482 1481
75
Acknowledgment
This work was partly funded by a grant from the Danish Research
Foundation, The Ministry of Science Technology and
Innovation, Denmark (08-036109) with a stipend for the corresponding
author.
References
[1] Engvall IL, Svensson B, Tengstrand B, Brismar K, Hafstrom I. Impact of low-dose
prednisolone on bone synthesis and resorption in early rheumatoid arthritis:
experiences from a two-year randomized study. Arthritis Res Ther
2008;10(6):R128.
[2] Hepper CT, Halvorson JJ, Duncan ST, Gregory AJ, Dunn WR, Spindler KP. The
efficacy and duration of intra-articular corticosteroid injection for knee
osteoarthritis: a systematic review of level I studies. J Am Acad Orthop Surg
2009;17(10):638–46.
[3] DiBattista JA, Martel-Pelletier J, Antakly T, Tardif G, Cloutier JM, Pelletier JP.
Reduced expression of glucocorticoid receptor levels in human osteoarthritic
chondrocytes. Role in the suppression of metalloprotease synthesis. J Clin
Endocrinol Metab 1993;76(5):1128–34.
[4] Chrysis D, Ritzen EM, Savendahl L. Growth retardation induced by
dexamethasone is associated with increased apoptosis of the growth plate
chondrocytes. J Endocrinol 2003;176(3):331–7.
[5] Chrysis D, Zaman F, Chagin AS, Takigawa M, Savendahl L. Dexamethasone
induces apoptosis in proliferative chondrocytes through activation of caspases
and suppression of the Akt–phosphatidylinositol 3 0 -kinase signaling pathway.
Endocrinology 2005;146(3):1391–7.
[6] Sadowski T, Steinmeyer J. Effects of non-steroidal antiinflammatory drugs and
dexamethasone on the activity and expression of matrix metalloproteinase-1,
matrix metalloproteinase-3 and tissue inhibitor of metalloproteinases-1 by
bovine articular chondrocytes. Osteoarthritis Cartilage 2001;9(5):407–15.
[7] Nakazawa F, Matsuno H, Yudoh K, Watanabe Y, Katayama R, Kimura T.
Corticosteroid treatment induces chondrocyte apoptosis in an experimental
arthritis model and in chondrocyte cultures. Clin Exp Rheumatol
2002;20(6):773–81.
[8] Richardson DW, Dodge GR. Dose-dependent effects of corticosteroids on the
expression of matrix-related genes in normal and cytokine-treated articular
chondrocytes. Inflamm Res 2003;52(1):39–49.
[9] Garvican ER, Vaughan-Thomas A, Redmond C, Gabriel N, Clegg PD. MMPmediated
collagen breakdown induced by activated protein C in equine
cartilage is reduced by corticosteroids. J Orthop Res 2010;28(3):370–8.
[10] Eastell R, Reid DM, Compston J, Cooper C, Fogelman I, Francis RM, et al. A UK
Consensus Group on management of glucocorticoid-induced osteoporosis: an
update. J Intern Med 1998;244(4):271–92.
[11] Hooyman JR, Melton III LJ, Nelson AM, O’Fallon WM, Riggs BL. Fractures after
rheumatoid arthritis. A population-based study. Arthritis Rheum
1984;27(12):1353–61.
[12] Caplan L, Saag KG. Glucocorticoids and the risk of osteoporosis. Expert Opin
Drug Saf 2009;8(1):33–47.
[13] Canalis E, Mazziotti G, Giustina A, Bilezikian JP. Glucocorticoid-induced
osteoporosis: pathophysiology and therapy. Osteoporos Int
2007;18(10):1319–28.
[14] Hofbauer LC, Zeitz U, Schoppet M, Skalicky M, Schuler C, Stolina M, et al.
Prevention of glucocorticoid-induced bone loss in mice by inhibition of RANKL.
Arthritis Rheum 2009;60(5):1427–37.
[15] Jia D, O’Brien CA, Stewart SA, Manolagas SC, Weinstein RS. Glucocorticoids act
directly on osteoclasts to increase their life span and reduce bone density.
Endocrinology 2006;147(12):5592–9.
[16] Weinstein RS, Chen JR, Powers CC, Stewart SA, Landes RD, Bellido T, et al.
Promotion of osteoclast survival and antagonism of bisphosphonate-induced
osteoclast apoptosis by glucocorticoids. J Clin Invest 2002;109(8):1041–8.
[17] Kim HJ, Zhao H, Kitaura H, Bhattacharyya S, Brewer JA, Muglia LJ, et al.
Glucocorticoids suppress bone formation via the osteoclast. J Clin Invest
2006;116(8):2152–60.
[18] Hofbauer LC, Gori F, Riggs BL, Lacey DL, Dunstan CR, Spelsberg TC, et al.
Stimulation of osteoprotegerin ligand and inhibition of osteoprotegerin
production by glucocorticoids in human osteoblastic lineage cells: potential
paracrine mechanisms of glucocorticoid-induced osteoporosis. Endocrinology
1999;140(10):4382–9.
[19] Sivagurunathan S, Muir MM, Brennan TC, Seale JP, Mason RS. Influence of
glucocorticoids on human osteoclast generation and activity. J Bone Miner Res
2005;20(3):390–8.
[20] Oshina H, Sotome S, Yoshii T, Torigoe I, Sugata Y, Maehara H, et al. Effects of
continuous dexamethasone treatment on differentiation capabilities of bone
marrow-derived mesenchymal cells. Bone 2007;41(4):575–83.
[21] Jager M, Fischer J, Dohrn W, Li X, Ayers DC, Czibere A, et al. Dexamethasone
modulates BMP-2 effects on mesenchymal stem cells in vitro. J Orthop Res
2008;26(11):1440–8.
[22] Espina B, Liang M, Russell RG, Hulley PA. Regulation of bim in glucocorticoidmediated
osteoblast apoptosis. J Cell Physiol 2008;215(2):488–96.

1482 S.H. Madsen et al. / Steroids 76 (2011) 1474–1482
[23] Mak W, Shao X, Dunstan CR, Seibel MJ, Zhou H. Biphasic glucocorticoiddependent
regulation of Wnt expression and its inhibitors in mature
osteoblastic cells. Calcif Tissue Int 2009;85(6):538–45.
[24] Buttgereit F, Zhou H, Kalak R, Gaber T, Spies CM, Huscher D, et al. Transgenic
disruption of glucocorticoid signaling in mature osteoblasts and osteocytes
attenuates K/BxN mouse serum-induced arthritis in vivo. Arthritis Rheum
2009;60(7):1998–2007.
[25] Zalavras C, Shah S, Birnbaum MJ, Frenkel B. Role of apoptosis in glucocorticoidinduced
osteoporosis and osteonecrosis. Crit Rev Eukaryot Gene Expr
2003;13(2–4):221–35.
[26] Kerachian MA, Seguin C, Harvey EJ. Glucocorticoids in osteonecrosis of the
femoral head: a new understanding of the mechanisms of action. J Steroid
Biochem Mol Biol 2009;114(3–5):121–8.
[27] Karsdal MA, Leeming DJ, Dam EB, Henriksen K, Alexandersen P, Pastoureau P,
et al. Should subchondral bone turnover be targeted when treating
osteoarthritis? Osteoarthritis Cartilage 2008;16(6):638–46.
[28] Kwan TS, Lajeunesse D, Pelletier JP, Martel-Pelletier J. Targeting subchondral
bone for treating osteoarthritis: what is the evidence? Best Pract Res Clin
Rheumatol 2010;24(1):51–70.
[29] Madsen SH, Goettrup AS, Thomsen G, Christensen ST, Schultz N, Henriksen K,
et al. Characterization of an ex vivo femoral head model assessed by markers
of bone and cartilage turnover. Cartilage 2011;2(3):265–78.
[30] Karsdal MA, Madsen SH, Christiansen C, Henriksen K, Fosang AJ, Sondergaard
BC. Cartilage degradation is fully reversible in the presence of aggrecanase but
not matrix metalloproteinase activity. Arthritis Res Ther 2008;10(3):R63.
[31] Sorensen MG, Henriksen K, Schaller S, Henriksen DB, Nielsen FC, Dziegiel MH,
et al. Characterization of osteoclasts derived from CD14 + monocytes isolated
from peripheral blood. J Bone Miner Metab 2007;25(1):36–45.
[32] Ghosh-Choudhury N, Windle JJ, Koop BA, Harris MA, Guerrero DL, Wozney JM,
et al. Immortalized murine osteoblasts derived from BMP 2-T-antigen
expressing transgenic mice. Endocrinology 1996;137(1):331–9.
[33] Olsen AK, Sondergaard BC, Byrjalsen I, Tanko LB, Christiansen C, Muller A, et al.
Anabolic and catabolic function of chondrocyte ex vivo is reflected by the
metabolic processing of type II collagen. Osteoarthritis Cartilage
2007;15(3):335–42.
[34] Karsdal MA, Neutzsky-Wulff AV, Dziegiel MH, Christiansen C, Henriksen K.
Osteoclasts secrete non-bone derived signals that induce bone formation.
Biochem Biophys Res Commun 2008;366(2):483–8.
[35] Jonsson KB, Frost A, Larsson R, Ljunghall S, Ljunggren O. A new fluorometric
assay for determination of osteoblastic proliferation: effects of glucocorticoids
and insulin-like growth factor-I. Calcif Tissue Int 1997;60(1):30–6.
[36] Bay-Jensen AC, Liu Q, Byrjalsen I, Li Y, Wang J, Pedersen C, et al. Enzyme-linked
immunosorbent assay (ELISAs) for metalloproteinase derived type II collagen
neoepitope, CIIM-increased serum CIIM in subjects with severe radiographic
osteoarthritis. Clin Biochem 2011;44(5-6):423–9.
[37] Sumer EU, Sondergaard BC, Rousseau JC, Delmas PD, Fosang AJ, Karsdal MA,
et al. MMP and non-MMP-mediated release of aggrecan and its fragments
from articular cartilage: a comparative study of three different aggrecan and
glycosaminoglycan assays. Osteoarthritis Cartilage 2007;15(2):212–21.
[38] Henriksen K, Gram J, Neutzsky-Wulff AV, Jensen VK, Dziegiel MH, Bollerslev J,
et al. Characterization of acid flux in osteoclasts from patients harboring a
G215R mutation in ClC-7. Biochem Biophys Res Commun 2009;378(4):804–9.
76
[39] Camplejohn KL, Allard SA. Limitations of safranin ‘O’ staining in proteoglycandepleted
cartilage demonstrated with monoclonal antibodies. Histochemistry
1988;89(2):185–8.
[40] Habib GS. Systemic effects of intra-articular corticosteroids. Clin Rheumatol
2009;28(7):749–56.
[41] Skwara A, Peterlein CD, Tibesku CO, Rosenbaum D, Fuchs-Winkelmann S.
Changes of gait patterns and muscle activity after intraarticular treatment of
patients with osteoarthritis of the knee: a prospective, randomised,
doubleblind study. Knee 2009;16(6):466–72.
[42] Mansell JP, Collins C, Bailey AJ. Bone, not cartilage, should be the major focus in
osteoarthritis. Nat Clin Pract Rheumatol 2007;3(6):306–7.
[43] van der Kraan PM, Vitters EL, Postma NS, Verbunt J, Van den Berg WB.
Maintenance of the synthesis of large proteoglycans in anatomically intact
murine articular cartilage by steroids and insulin-like growth factor I. Ann
Rheum Dis 1993;52(10):734–41.
[44] Madsen SH, Sumer EU, Bay-Jensen AC, Sondergaard BC, Qvist P, Karsdal MA.
Aggrecanase- and matrix metalloproteinase-mediated aggrecan degradation is
associated with different molecular characteristics of aggrecan and separated
in time ex vivo. Biomarkers 2010;15(3):266–76.
[45] Augustine AJ, Oleksyszyn J. Glucocorticosteroids inhibit degradation in bovine
cartilage explants stimulated with concomitant plasminogen and interleukin-
1 alpha. Inflamm Res 1997;46(2):60–4.
[46] Rauch A, Seitz S, Baschant U, Schilling AF, Illing A, Stride B, et al.
Glucocorticoids suppress bone formation by attenuating osteoblast
differentiation via the monomeric glucocorticoid receptor. Cell Metab
2010;11(6):517–31.
[47] Kalak R, Zhou H, Street J, Day RE, Modzelewski JR, Spies CM, et al. Endogenous
glucocorticoid signalling in osteoblasts is necessary to maintain normal bone
structure in mice. Bone 2009;45(1):61–7.
[48] Weinstein RS, Jilka RL, Almeida M, Roberson PK, Manolagas SC. Intermittent
parathyroid hormone administration counteracts the adverse effects of
glucocorticoids on osteoblast and osteocyte viability, bone formation, and
strength in mice. Endocrinology 2010;151(6):2641–9.
[49] Henriksen K, Neutzsky-Wulff AV, Bonewald LF, Karsdal MA. Local
communication on and within bone controls bone remodeling. Bone
2009;44(6):1026–33.
[50] Jevon M, Hirayama T, Brown MA, Wass JA, Sabokbar A, Ostelere S, et al.
Osteoclast formation from circulating precursors in osteoporosis. Scand J
Rheumatol 2003;32(2):95–100.
[51] Takuma A, Kaneda T, Sato T, Ninomiya S, Kumegawa M, Hakeda Y.
Dexamethasone enhances osteoclast formation synergistically with
transforming growth factor-beta by stimulating the priming of osteoclast
progenitors for differentiation into osteoclasts. J Biol Chem
2003;278(45):44667–74.
[52] Soe K, Delaisse JM. Glucocorticoids maintain human osteoclasts in the active
mode of their resorption cycle. J Bone Miner Res 2010 [Epub ahead of print].
Ref Type: Generic.

CHAPTER 6
CHAPTER 6: PAPER III
PAPER III
___________________________________________________________________________
Introductory remarks
This manuscript focuses on the pathogenesis of cartilage, and evaluates the different proteolytic
degradation processes for aggrecan, which vary between the OA-stages. It is important to
understand the different degrading processes of aggrecan if we want to target the right proteases
at a certain stage of disease.
77

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
Biomarkers, 2010; 15(3): 266–276
original article
Aggrecanase- and matrix metalloproteinase- mediated
aggrecan degradation is associated with different
molecular characteristics of aggrecan and separated in
time ex vivo
Suzi Hoegh Madsen, Eren Ufuk Sumer, Anne-Christine Bay-Jensen, Bodil-Cecilie Sondergaard,
Per Qvist, and Morten Asser Karsdal
Nordic Bioscience A/S, Herlev, Denmark
Introduction
abstract
Aggrecan is one of the first proteins to be depleted from articular cartilage in early osteoarthritis. We investigated
the molecular differences between matrix metalloproteinase (MMP)- and aggrecanase-mediated
aggrecan degradation, as a consequence of their distinct time-dependent degradation profiles. Cartilage
degradation was induced by cytokine stimulation in bovine articular cartilage explants and quantified by
a dye-binding assay and immunoassays. The size of degradation fragments was analysed by Western blot.
Cytokine stimulation resulted in the early release of aggrecanase-mediated aggrecan degradation fragments.
In contrast, MMP-mediated aggrecan degradation began only at day 16 and continued to day 21.
Western blot analysis showed that glycosylated high-molecular-weight 374 ARGSVI fragments appeared at
day 7, in contrast to deglycosylated low-molecular-weight 342 FFGVG fragments which were detected at day
21. Aggrecan degradation may be divided into two different pools, a high-molecular-weight aggrecanasemediated
pool, and a low-molecular-weight MMP-mediated pool. This may have implications for the development
of intervention strategies for OA.
Keywords: Proteoglycans; cartilage; osteoarthritis; ELISA; biomarkers; aggrecanases; MMPs
Cartilage is normally maintained by a balance between
catabolic and anabolic processes. However, in joint
degenerative diseases, such as osteoarthritis (OA), the
rate of cartilage degradation exceeds the rate of formation,
resulting in a net loss of cartilage matrix (Behrens
et al. 1989, Felson & Neogi 2004). The extracellular matrix
(ECM) consists mainly of collagen type II, which forms
a fibrous network that entraps proteoglycans of which
aggrecan is the most predominant (Hardingham &
Fosang 1995, Kiani et al. 2002).
Aggrecan is one of the ECM proteins to be degraded
in OA (Caterson et al. 2000), primarily by ADAMTSs
(a disintegrin and metalloproteinase domain with
thrombospondin motifs) (Glasson et al. 2005, Struglics et al.
2006) and matrix metalloproteinases (MMPs) (Fosang et al.
1996, Lark et al. 1997, Struglics et al. 2006). Aggrecanase
activity leads to the release of glycosaminoglycans (GAGs),
although it is not known if MMPs also play a role in aggrecan
GAG loss (Sandy 2006, Struglics et al. 2006). As a consequence
of aggrecan loss, the cartilage exhibits altered
matrix properties such as loss of fixed negatively charged
density of the tissue and thereby lower water content and
elasticity (Roughley & Lee 1994). A deeper understanding
of the molecular mechanisms involved in proteoglycan
depletion may aid the development of efficient diseasemodifying
osteoarthritic drugs (DMOADs).
The core protein of aggrecan has three globular
domains: G1, G2 and G3. G1 and G2 are separated by
Address for Correspondence: Suzi Hoegh Madsen, Nordic Bioscience A/S, Herlev Hovedgade 207, DK-2730 Herlev, Denmark. Tel: + 45 44 52 52 52. Fax:
+ 45 44 52 52 51. E-mail: shm@nordicbioscience.com
(Received 03 August 2009; revised 01 December 2009; accepted 01 December 2009)
ISSN 1354-750X print/ISSN 1366-5804 online © 2010 Informa UK Ltd
DOI: 10.3109/13547500903521810 http://www.informahealthcare.com/bmk
78

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
the interglobular domain (IGD), which is highly sensitive
to proteolysis (Flannery et al. 1992, Hardingham
& Fosang 1995). G2 and G3 are heavily saturated with
highly negatively charged GAGs (Flannery et al. 1992,
Hardingham & Fosang 1995) (Figure 1). The G2 domain
is unique to aggrecan, and sandwich enzyme-linked
immunosorbent assays (ELISAs) of aggrecan degradation
have been developed against peptides with the G2
and the neoepitopes 342 FFGVG or 374 ARGSV (Karsdal
et al. 2007, 2008) generated by MMPs (Flannery et al.
1992) and ADAMTS (Tortorella et al. 1999), respectively.
The ADAMTS-specific 374 ARGSV site is located between
the 342 FFGVG site and the G2 domain. The 342 FFGVG-G2
fragments assayed in the present study represent only
MMP-mediated degradation, which has not already
been processed by ADAMTSs. The 374 ARGSV-G2 represents
fragments, which theoretically may have been
pre-processed by MMPs at the 342 FFGVG site (Figure 1),
although the available evidence suggests that this is very
unlikely (Mercuri et al. 2000).
Studies with ex vivo cultures have shown that cytokinestimulated
cartilage releases collagen and proteoglycan
fragments to the conditioned medium (Sondergaard
et al. 2006a). These fragments can be detected by novel
ELISA assays, which measure neoepitopes cleaved by
specific proteases. These degradation patterns resemble
the molecular mechanisms seen in the pathogenesis
of OA (Caterson et al. 2000). Our group has previously
shown that cytokine-stimulated cartilage releases products
with specific neoepitopes from collagen type II and
aggrecan (Karsdal et al. 2007, 2008, Rousseau et al. 2008,
G1
MMP
IPEN 341 342 FFGV
AF-28
ADAMTS
Aggrecanase- and MMP-mediated aggrecan degradation 267
IGD sGAG
TEGE 373 374 ARGS
BC-3
Sondergaard 2006a). Furthermore, we also showed that
MMP-2 and -9 activity increased over time in cytokinestimulated
cartilage explants (Sondergaard 2006a),
which is important when investigating the difference
between aggrecanase- and MMP-mediated degradation
of aggrecan.
Different research groups have suggested that articular
cartilage aggrecan may be present in the tissue in
two or more metabolic pools (Ilic et al. 1995, Lark et al.
1997, Lohmander et al. 1973, Mok et al. 1994, Struglics
et al. 2006). However, currently, the temporal and spatial
molecular degradation of aggrecan has not been
clarified, and various contradictory theories and models
have been proposed by different investigators (Fosang
et al. 2000, Lark et al. 1997, Sandy 2006, Struglics et al.
2006). It has been proposed by Fosang et al. (2000) that
aggrecan degradation by ADAMTSs and MMPs occurs in
independent locations and with different kinetics. More
recently Struglics et al. (2006) analysed human synovial
fluids and concluded that ADAMTSs and MMPs act
sequentially on the same substrate molecule. A review
of the area by Sandy (2006) attempted to clarify these
controversies. On the basis of available evidence, Sandy
(2006) concluded that ADAMTSs and MMPs degrade
different aggrecan structures. Specifically it was suggested
that ADAMTSs act primarily on full-length or
G1-G2-chondroitin sulfate1 aggrecan by a sequential
process which removes the chondroitin sulfate-bearing
domains and lastly at the Glu 373-374 Ala bond in the IGD.
In contrast it was proposed that MMPs degrade only the
calpain (calcium-dependent cysteine protease) product
G2 G3
F-78-HRP
F-78-HRP
Figure 1. Protease-mediated neoepitopes in aggrecan. Aggrecan molecule with three globular domains G1, G2 and G3. The interglobular domain
(IGD) is highly sensitive to proteolysis. Beneath the aggrecan molecule are the 342 FFGVG-G2 and 374 ARGSV-G2 fragments mediated by matrix metalloproteinase
(MMP) and ADAMTS, respectively. The antibodies BC-3 and AF-28 recognize the neoepitopes and F-78 is the detection antibody
binding to the G2 domain. The domain between G2 and G3 is heavily saturated with highly negatively charged glycosaminoglycans (GAGs).
79

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
268 Suzi Hoegh Madsen et al.
G1-G2-keratan sulfate (Maehara et al. 2007, Oshita et al.
2004) by cleavage at the 342 FFGVG site. In this scenario
the independent activities of ADAMTSs and MMPs are
seen as being due to a protease-specific preference for
different substrate structures. Furthermore, Lark et al.
(1997) described three hypothetical models for pathways
of proteolytic degradation of aggrecan, emphasizing the
complex nature of the generation of various fragments
from the same molecule.
To examine further the kinetics and molecular
dynamics of aggrecanolysis in cartilage explant cultures,
we investigated the degradation process by quantifying
the time-dependent release of ADAMTS- and MMPmediated
aggrecan fragments by the 374 ARGSVI-G2
and 342 FFGVG-G2 ELISA assays, respectively. These
new aggrecan fragments were categorized by size, via
Western blot, to identify molecular differences regarding
glycosylation of the core protein. Furthermore, we investigated
the specificity pattern of MMP-9 and ADAMTS-4
for degrading native and deglycosylated aggrecan in
vitro with respect to generating the aforementioned ex
vivo neoepitopes.
Materials and methods
Tissue preparation
Bovine articular cartilage explants were harvested by
using a scalpel to remove the articular cartilage without
adherent calcified cartilage from stifle joints of
18-month-old bovine heifers. The cartilage explants
(12 ± 2 mg) were washed in phosphate-buffered saline
(PBS) and cultured in 96-well plates (Nunc, Slangerup,
Denmark) in four replicas with 200 µl serum-free
medium (Life Technologies, Naerum, Denmark) plus
penicillin/streptomycin (Invitrogen, Taastrup, Denmark)
per well in the absence or presence of tumour necrosis
factor (TNF)-α (10 ng ml −1 ) (R&D Systems, Abingdon,
UK) and oncostatin M (OSM) (20 ng ml −1 ) (Sigma
Aldrich, Broendby, Denmark). These concentrations
had been optimized in previous ex vivo model studies
by Sondergaard et al. (2006a, b). As a negative control,
cartilage was metabolically inactivated (MI) by three
repeated freeze–thaw cycles in liquid N 2 , and at 37°C in
a water bath. The explants culture media were replaced
every 2nd or 3rd day for 21 days. The conditioned media
were kept at -20°C until tested for the presence of biochemical
markers.
Biochemical markers of aggrecan degradation
MMP-mediated aggrecan degradation was quantified
using the 342 FFGVG-G2 sandwich ELISA assay (Sumer
80
et al. 2007), with AF-28 (Fosang et al. 1995) as the
catching antibody. Quantification of the aggrecanasemediated
374 ARGSV-G2 neoepitope was obtained in
a similar way to the 342 FFGVG-G2 ELISA, except the
catching antibody was BC-3 (Abcam, Cambridge, UK)
to recognize the aggrecanase-mediated neoepitope
374 ARGSVI (Karsdal 2007). Both assays required the presence
of the G2 domain on the fragments being assayed,
with the detecting antibody being F-78. The efficacy of
the antibodies was confirmed by blocking experiments
in cytokine-stimulated explant media with the peptides
342 FFGVGEEDITVQT and 374 ARGSVILTVKPIF (Bachem,
Weil am Rhein, Germany).
Detection of sulfated glycosaminoglycans
Sulfated glycosaminoglycan (sGAG) levels were measured
in the conditioned medium using the alcian blue-binding
assay according to the manufacturer’s instructions (Euro-
Diagnostica, Malmö, Sweden).
Western blot on supernatants
Conditioned medium from four replicates of MI, nontreated,
or cytokine-stimulated cartilage explants from
day 7 or 21 were either deglycosylated for 4 h with keratanase
and chondroitinase lyase ABC (0.1 units/10 μg)
(Sigma Aldrich) with gentle shaking at 37°C in 100 mM
Tris/HCl and sodium acetate buffer at pH 6.5, or left
untreated. Thirty microlitres of the deglycosylated or
untreated sample was electrophoresed on 3–8% precasted
Tris–acetate gradient gels under reducing conditions
using Tris–acetate as running buffer (Invitrogen).
After transferring the proteins to polyvinyl difluoride
membranes overnight at room temperature (RT) at 120
mA in a 10 mM N-cyclohexyl-3-aminopropanesulfonic
acid buffer with 5% ethanol, the membranes were
blocked with 50 ml of 5% non-fat milk powder, 0.2%
Tween-20 in PBS with shaking for 2 h at RT. Antibodies
were applied in appropriate dilutions in 50 ml PBS
with 5% bovine non-fat milk, 0.2% Tween-20 for 1 h at
RT. BC-3 anti- 374 ARGSVI cell supernatants (Caterson
et al. 2000) diluted 1:50 or the anti- 342 FFGVG monoclonal
antibody AF-28 diluted 1:2250, giving a final
concentration of 1200 ng ml −1 IgG (Fosang et al. 1995).
After washing three times with PBS containing 0.2%
Tween-20, the membranes were incubated for 1 h at RT
with peroxidase-labelled rabbit antimouse antibody
(1:5000) (Millipore, Copenhagen, Denmark) in PBS
with 5% bovine non-fat milk powder, 0.2% Tween-20.
After washing three times with PBS (0.2% Tween-20),
the bands were visualized using an electrochemiluminescence
developer system (Amersham, Hilleroed,
Denmark).

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
ELISA for determining the total aggrecanase activity
The total aggrecanase activity in the conditioned medium
from days 7–21 was measured in four replicates using a
commercially available ELISA assay according to the
manufacturer’s instructions (MD Bioscience, Zurich,
Switzerland).
Digestion of aggrecan fragments in the conditioned
medium with ADAMTS-4
The conditioned medium from cytokine-stimulated
cartilage explants from day 21 was diluted 1:50 in
aggrecanase digestion buffer (50 mM Tris–HCI, pH 7.5,
150 mM NaCl, 5 mM CaCl 2 , 1 μM leupeptin, 1 μM pepstatin,
1mM Pefabloc, 0.05% Tween-20) and digested for
15 min with shaking at 37°C, using a total concentration
of 20 nM of ADAMTS-4 (Millipore). The digestions were
stopped with 0.5 mM EDTA and the media tested using
the 342 FFGVG-G2 and 374 ARGSV-G2 assays.
MMP-9 and ADAMTS-4 digestion on native and
deglycosylated aggrecan
Purified bovine aggrecan (Sigma Aldrich) was dissolved in
chondroitinase buffer (50 mM Tris, pH 8, 60 mM sodium
acetate, 0.02% BSA) to a concentration of 2 mg ml −1 . One
half of the solution was deglycosylated with chondroitinase
ABC (0.01 U/10 µg) for 4 h at 37°C with shaking.
The other half was left untreated but incubated along
the deglycosylated half. Two hours before terminating
the deglycosylation, the MMP-9 enzyme (Calbiochem,
Herlev, Denmark) was activated in 4-aminophenylmercuric
acetate (APMA) (1 mM) at 37°C for 2 h. The native
and deglycosylated aggrecan were each digested with
MMP-9 (1:100 enzyme/protein) and ADAMTS-4 (1:50
enzyme/protein). Digestion with MMP-9 continued for
3 days in a MMP buffer (100 mM Tris–HCl, 100 mM NaCl,
10 mM CaCl 2 , 2 mM Zn-acetate, pH 8) while digestion
with ADAMTS-4 took place over 22 h in a ADAMTS-4
buffer (75 mM Tris–HCl, 5 mM CaCl 2 .2H 2 O, 150 mM NaCl,
pH 7.5) at 37°C, and was stopped with 0.5 mM EDTA. The
MMP-9 and ADAMTS-4 digestions on native and deglycosylated
aggrecan were examined using the 342 FFGVG-G2
and 374 ARGSV-G2 ELISA assays.
Statistics
Results are shown as mean + standard error of mean
(SEM). Baseline was set at day 2 of culturing. Differences
between mean values were compared using the Student’s
t-test for unpaired observations, assuming normal distribution
where more than four replicates were used.
Differences were considered statistically significant if
p < 0.05.
Aggrecanase- and MMP-mediated aggrecan degradation 269
81
Results
Depletion of sulfated glycosaminoglycans in
cytokine-stimulated cartilage explants is mediated by
aggrecanases
As aggrecan is one of the first proteins to deplete from
articular cartilage in early OA, we investigated the timedependent
enzyme-mediated release pattern of aggrecan
molecules from cytokine-stimulated bovine articular
cartilage explants. The first cytokine-stimulated release
of aggrecan fragments from bovine articular cartilage
explants was caused by aggrecanase activity, as demonstrated
by an increase in fragments detected by the
374 ARGSV-G2 assay. The maximum level of aggrecan fragments
was detected at day 7 (Figure 2A) corresponding
to an approximate sevenfold increase compared with
non-stimulated bovine articular cartilage explants. After
the peak at day 7, 374 ARGSVI-G2 concentrations were
reduced to background levels.
To investigate if the reduction in 374 ARGSVI-G2 after
day 7 resulted from depletion of the aggrecan protein as a
substrate, we measured the MMP-mediated 342 FFGVG-G2
fragments. A time-dependent analysis showed that after
day 16, increased 342 FFGVG-G2 fragments were released
from cytokine-stimulated cartilage. At day 21, an increase
of more than 1600% in these released fragments was
observed in cytokine-stimulated cartilage compared
with non-stimulated cartilage (Figure 2B). This indicted
that aggrecan was present as a substrate during the entire
culture period.
Karsdal et al. (2008) showed a total depletion of sGAGs,
assessed by alcian blue staining from bovine cartilage
explants as early as after 7 days after cytokine stimulation,
while the non-stimulated cartilage retained the sGAGs in
the ECM. Measurements of the time-dependent release
of sGAGs to the conditioned medium (Figure 2C) showed
a similar pattern to the aggrecanase-mediated aggrecan
degradation. This indicated that the loss of sGAGs may
be linked to the loss of aggrecanase-mediated aggrecan
degradation.
Aggrecanase-mediated aggrecan degradation
fragments are predominantly glycosylated whereas
MMP-mediated aggrecan degradation fragments are
predominantly deglycosylated
Due to the strong correlation between aggrecanasemediated
aggrecan degradation and sGAG-substituted
molecules, we investigated aggrecan fragments in the
early period (day 7) and late period (day 21) by Western
blotting.
In the conditioned medium from cytokine-stimulated
cartilage from day 7, a high-molecular-weight band –
above 460 kDa – of the aggrecanase-mediated 374 ARGSVI

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
270 Suzi Hoegh Madsen et al.
A
µg 374ARGSVI-G2/ml/ mg cartilage
B
µg 342FFGVG-G2/ml/ mg cartilage
C
µg GAGs/ml/
mg cartilage
125
100
10
75
50
25
8
6
4
2
0
60
40
20
0
0
**
*
2
*
*
Days
Cytokines
W/O
neoepitope was recognized by the monoclonal antibody
BC-3 (Figure 3A). This band was reduced to approximately
325 kDa upon deglycosylation (Figure 3B). In
contrast, the low-molecular-weight smear of bands
carrying the 342 FFGVG neoepitopes detected in the
supernatant at day 21 with monoclonal antibody AF-28
*
4 7 9 11 1416 18 21
Days
* **
* **
2 4 7 9 11 14 16 18 21
**
**
2 4 7 9 11 14 16 18 21
Days
MI
Cytokines
W/O
MI
Cytokines
W/O
Figure 2. Quantification of biochemical markers. Bovine articular
cartilage explants were isolated and cultured with the proinflammatory
cytokines oncostatin M (OSM) (10 ng ml −1 ) in combination with
tumour necrosis factor (TNF)-α (20 ng ml −1 ) (- ◆ -). The time-dependent
release of aggrecanase-mediated 374 ARGSV-G2 (A), matrix metalloproteinase
(MMP)-mediated 342 FFGVG-G2 fragments (B) or sulfated
glycosaminoglycans (GAGs) (C) were quantified. Non-stimulated
explants (- ■ -) were used as the vehicle. Metabolically inactivated
(MI) explants (- ● -) were used as a negative control to measure the
non-chondrocyte-mediated release. The asterisks indicate significant
difference compared with non-treated explants; *p < 0.05, **p < 0.01,
***p < 0.001; n = 4. W/O, non-stimulated controls).
MI
82
(Figure 3C) was relatively insensitive to deglycosylation
(Figure 3D). No bands were detected when BC-3 was
used for staining of the conditioned media from day 21,
or the monoclonal antibody AF-28 from day 7 (data not
shown), which resemble the time-dependent analysis by
the 374 ARGSV-G2 and 342 FFGVG-G2 assays in Figure 2A
and B.
The calculated theoretic molecular weight difference
between the neoepitopes 342 FFGVG-G2 and 374 ARGSV-G2
is approximately 3.6 KDa (32 aa). However, the molecular
weight difference between these two neoepitopes when
released as fragments to the conditioned media was far
from this. The 374 ARGSV-G2 fragment was glycosylated
and large, and the 342 FFGVG-G2 fragment was deglycosylated
and small. This explains the correlation between
the release patterns of sGAGs and aggrecanase-mediated
aggrecan degradation. Furthermore, the Western blots
clearly showed that the deglycosylated 374 ARGSV-G2
fragments were more than 200 KDa larger than the
deglycosylated 342 FFGVG-G2 fragments. This indicates
that we were not able to deglycosylate the 374 ARGSV-G2
fragments totally, and that the protein structure of the
aggrecanase-mediated aggrecan fragments may be larger
than the MMP-mediated 342 FFGVG-G2 fragments.
Lack of aggrecanase-mediated aggrecan degradation
is not a consequence of lack of aggrecanase activity
In the late culture period, after day 16, we saw high
levels of 342 FFGVG-G2 fragments and very low levels
of 374 ARGSV-G2 fragments. To investigate if the lack of
aggrecanase-mediated aggrecan degradation was caused
by depletion of aggrecanase activity, we tested the conditioned
medium at days 7 and 21. The concentration of
aggrecanases released at day 7 from cytokines-stimulated
cartilage did not show any increase compared with that
from the non-stimulated cartilage. In contrast, the concentration
of aggrecanases released at day 21 from cytokinestimulated
cartilage revealed an elevation of over 935%
compared with non-stimulated cartilage (Figure 4). This
reduced release of aggrecanases at day 7 is interesting,
because of the high aggrecanase-mediated degradation
of aggrecan at the same time. However, the aggrecanases
might still be sited in the extracellular matrix, which is
then released later due to the loss of sGAGs. It appeared
that the low concentration of 374 ARGSV-G2 fragments at
day 21 was not due to low aggrecanase activity.
The MMP-mediated deglycosylated aggrecan
degradation released in the late period is not
protected from further cleavage at the major
aggrecanase site in the IGD in vitro
Aggrecanases were present in the conditioned medium
at the end of the culture period. To investigate whether

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
Day 7
Day 21
the lack of 374 ARGSV-G2 fragments at day 21 was due to
the inability of aggrecanases to process the 342 FFGVG-G2
fragments further, we added recombinant enzymes to
the culture medium and measured further cleavage of
the 342 FFGVG-G2 aggrecan fragments by quantitative
measurements.
A
W/O MI
C
W/O
MI
Aggrecanase- and MMP-mediated aggrecan degradation 271
Untreated Deglycosylated
Cytokines
Cytokines
374 ARGSV
460 kDa
75 kDa
342 FFGVG
50 kDa
B
W/O
D
W/O
MI Cytokines
MI Cytokines
460 kDa
374 ARGSV
268 kDa
75 kDa
342 FFGVG
50 kDa
Figure 3. Western blot analysis of aggrecanase- and matrix metalloproteinase (MMP)-mediated fragments. Conditioned medium of four replicates
of non-stimulated, metabolically inactivated (MI) or cytokines stimulated explants from day 7 or day 21 were separately pooled by days and either
left untreated (A and C) or deglycosylated (B and D). Pooled supernatants were run in 3–8% precasted Tris–acetate gels under reducing conditions.
The two uniform membranes were stained with either the monoclonal antibody BC-3 (1:50 dilution) staining for the aggrecanase-mediated
374 ARGSVI neoepitope (A and B) or the monoclonal antibody AF-28 (1200 ng ml −1 ), staining for the MMP-mediated 342 FFGVG-sequence (C and D).
Peroxidase-labelled rabbit antimouse antibody (1:5000 dilution) was used as a secondary antibody. A standard molecular weight marker was also
run to determine the size of the detected fragments (not shown). W/O, non-stimulated controls).
nM Aggrecanase Conc.
/mg cartilage
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
W/O
MI
Cytokines
Day 7
**
Day 21
Figure 4. Quantification of aggrecanase activity in catabolically
stimulated bovine explants. Bovine articular cartilage explants were
isolated and cultured in the presence or absence of cytokines for 21
days. Non-stimulated explants were used as the vehicle. Metabolically
inactivated (MI) explants were used as a negative control to measure
the non-chondrocyte-mediated release. The mean value of aggrecanase
activity in the conditioned medium was determined by measuring
four different wells containing the same treatment. This was done
for all three different treatments at days 7 and 21. The aggrecanase
activity was significant increased (p < 0.01) in supernatants (n = 4) from
cytokine-stimulated cartilage compared with the non-stimulated controls
(W/O) at day 21.
83
The conditioned medium from cytokine-stimulated
explants from day 21 was digested in vitro with
ADAMTS-4 and measured by the 342 FFGVG-G2 and
374 ARGSV-G2 assays. The aggrecanase degradation index
( 374 ARGSV-G2/ 342 FFGVG-G2) was calculated and the
ADAMTS-4 digested media produced a 225% elevation
compared with the index from non-ADAMTS-4-digested
media (Figure 5). This indicates that the 373/374 cleavage
site in the aggrecan fragments present at day 21 (only
low-molecular-weight aggrecan) was functional, but only
cleaved in vitro, not ex vivo.
Substrate affinities by MMPs and aggrecanases may
be essential for the diverse generation of 342 FFGVG-G2
and 374 ARGSV-G2, and not the availability of aggrecan
But why were the aggrecanase-mediated aggrecan
fragments not present ex vivo at day 21, when both
aggrecanases and a functional aggrecanase cleavage site
were present in the conditioned medium? To investigate
the difference between in vitro and ex vivo cleavage in
aggrecan, we focused on the importance of glycosylation,
which was absent in the late culture period. We investigated
the different proteolytic potentials for ADAMTS-4
and MMP-9 on either native or deglycosylated aggrecan
in vitro and compared the results with the present ex vivo
study. These proteolytic digestions were assessed by the
374 ARGSV-G2 and 342 FFGVG-G2 assays.

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
272 Suzi Hoegh Madsen et al.
ADAMTS-4 digestion significantly increased the levels
of 374 ARGSV-G2 on both native and deglycosylated
aggrecan compared with the non-digested controls
(Figure 6A). Furthermore, the 374 ARGSV-G2 levels were
( 374 ARGSVI-G2)/( 342 FFGVG-G2)
1.2
1.0
0.8
0.6
0.4
0.2
0.0
Cytokines
Day 21
*
Cytokines +
ADAMTS-4
Figure 5. Further cleavage of released matrix metalloproteinase
(MMP)-mediated 342 FFGVG-G2 fragments to 374 ARGSVI-G2 by
ADAMTS-4. Supernatants from bovine articular cartilage explants cultured
in the presence of cytokines for 21 days (n = 4) were digested with
20 nM ADAMTS-4, diluted 1:50 and assessed by the immunoassays
374 ARGSV-G2 and 342 FFGVG-G2 quantifying aggrecanase- and MMPmediated
aggrecan fragments, respectively. The degradation index
( 374 ARGSV-G2)/( 342 FFGVG-G2) was made for the ADAMTS-4-treated
cytokine-stimulated supernatants (black bar) and the undigested
(ADAMTS) cytokine-stimulated control (grey bar); *p < 0.05.
A
ng 374-G2/ml
250000
200000
150000
100000
50000
0
Nativ, ADAMSTS −
ADAMTS mediated
aggrecan degradation
***
Nativ, ADAMSTS +
***
Deglyco, ADAMSTS −
***
Deglyco, ADAMSTS +
significantly higher for digested deglycosylated aggrecan
compared with digested native aggrecan. MMP-9
digestion showed a similar degradation pattern to
ADAMTS-4 digestion. The levels of 342 FFGVG-G2 were
significantly increased on both native and deglycosylated
aggrecan compared with the non-digested controls
(Figure 6B). Furthermore, the 342 FFGVG-G2 levels
were significantly higher for digested deglycosylated
aggrecan compared with digested native aggrecan.
This did not correlate with the detection of 374 ARGSV-G2
fragments at day 7 (high level) and day 21 (low level)
in the ex vivo study (Figure 2A), but highlights the
importance of distinguishing between in vitro and ex
vivo results.
Interestingly, the 374 ARGSV-G2 levels were approximately
50 times greater than the 342 FFGVG-G2 levels.
This detectable difference between the two neoepitope
assays measured from in vitro samples resembles the
difference we were able to detect in the individual peaks
at days 7 and 21 in the ex vivo samples (Figure 2A and B).
This indicates that the ex vivo measurements were reliable.
This confirms that the lower concentration level of
342 FFGVG-G2 compared with 374 ARGSV-G2 was not due
to limited aggrecan as a substrate.
We were unable to conclude if glycosylation played
an important role, based on the in vitro study. However,
the substrate affinities of MMPs and aggrecanases may
still be the essential factor for the diverse generation of
342 FFGVG-G2 and 374 ARGSV-G2 ex vivo.
B
ng 342-G2/ml
4000
3000
2000
1000
0
Nativ, MMP −
MMP mediated
aggrecan degradation
Figure 6. Matrix metalloproteinase (MMP)-9 and ADAMTS-4 digestion on native and deglycosylated aggrecan. Purified bovine aggrecan was
either left untreated or deglycosylated. The native and deglycosylated aggrecan were each digested with MMP-9 and ADAMTS-4 and assessed by
the 374 ARGSV-G2 (A) and 342 FFGVG-G2 (B) ELISA assays; *p < 0.001; n = 4.
84
***
Nativ, MMP +
***
Deglyco, MMP −
***
Deglyco, MMP +

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
Discussion
To elucidate the molecular mechanisms behind the
destruction of the ECM, a large number of studies have
reported the molecular nature of type II collagen and
aggrecan fragments as they are released into the synovial
fluid (Malfait et al. 2002, Sandy et al. 1992, Struglics
et al. 2006) or into explant culture supernatants (Caterson
et al. 2000, Fosang et al. 2000, Little et al. 1999, Sumer
et al. 2007, Sztrolovics et al. 2002). While members of the
MMP family have long been known to have the capacity
to cleave aggrecan (Fosang et al. 1996, Struglics et al.
2006), only recently were the aggrecanases identified to
have the same ability (Stanton et al. 2005, Struglics et al.
2006). Subsequently, the relative importance of MMPs
and aggrecanases in aggrecanolysis has been debated.
In the present study we demonstrated that in cytokinestimulated
ex vivo cultures of articular bovine cartilage,
both MMP- and aggrecanase-mediated aggrecan fragments
were released into the supernatant, but not
simultaneously. The catabolic stimulation of cartilage
explants induced an early release of aggrecanase-mediated
374 ARGSVI-G2 fragments, and only after returning to
background levels did we detect an increase in the MMPmediated
release of 342 FFGVG-G2 fragments. The peak at
day 7 and subsequent drop in the release of aggrecanasemediated
aggrecan fragments in the ex vivo culture
(Figure 2A) was not caused by depletion of the aggrecan
as a substrate, because we detected MMP-mediated
aggrecan degradation later in the culture period (Figure
2B), which verified the presence of the core proteins of
aggrecan. Furthermore, our data suggested that the drop
in aggrecanase-mediated aggrecan fragments was not
caused by the lack of the aggrecanases, as a high level
of their activity was detected in the late culture period
(Figure 4).
As presented, we found both evidence of the presence
of aggrecan as a substrate and aggrecanase activity in the
late period of culture, but interestingly no 374 ARGSVI-G2
fragments. We further demonstrated that it was possible
to generate 374 ARGSVI-G2 fragments from the MMPmediated
342 FFGVG-G2 fragments, when treated with
aggrecanases in vitro. This suggests that low concentrations
of 374 ARGSVI-G2 fragments in the late culture period
did not result from an inaccessible 373/374-cleavage site
in the 342 FFGVG-G2 fragments. However, this does not
explain why the cleavage only happens in vitro and not
ex vivo. The ambiguity regarding the lack of 374 ARGSVI-G2
fragments in the late culture period resulted in further
investigations. We detected an identical pattern of
aggrecanase-mediated aggrecan degradation as occurred
with the sGAG release. This link between 374 ARGSVI-G2
fragments and sGAGs was confirmed via Western blotting,
showing highly glycosylated 374 ARGSVI-G2 fragments
from the early culture period and deglycosylated
Aggrecanase- and MMP-mediated aggrecan degradation 273
85
342 FFGVG-G2 fragments in the late culture period. These
data suggest that two different pools of aggrecan molecules
contain an N-terminal neoepitope in the IGD
region ( 342 FFGVG and 374 ARGSVI), the G2 domain and a
C-terminus of considerable difference. In particular, the
C-terminus and glycosylation are obvious candidates to
explain the discrepancy in the molecular sizes of the two
aggrecan pools. The results from the present paper are
summarized in Figure 7.
Our high- versus low-molecular-weight data support
the theory described by Sandy (2006) that MMPs
and aggrecanases act independently, and aggrecanases
cleave intact aggrecan, whereas MMPs degrade aggrecan
molecules which have already been C-terminally cleaved.
Moreover, Lark et al. (1997) reported that MMPs are
located in the damaged cartilage area, whereas aggrecanases
are located in the non-damaged area. This is
consistent with our study showing that aggrecanases are
present in the beginning of the culture period and MMPs
are present in the later culture period when the cartilage
is further degraded.
Struglics et al. (2006) suggested that two distinct pathways
were responsible for aggrecanolysis in vivo: first,
the MMP-mediated 342 FFGVG-neoepitope in the pericellular
matrix and secondly the aggrecanase-mediated
374 ARGSVI neoepitope in the inter-territorial matrix, on
the same aggrecan molecule. This is not supported by
the results in the present paper. However, the theory of
the two distinct pathways is consistent with observations
made in the present study, and is furthermore supported
by identification of such aggrecan fragments in human
synovial fluid spanning molecular sizes of 80–335 kDa
(Struglics et al. 2006). Our Western blot data differ from
those of Struglics et al. (2006) who found MMP-mediated
high-molecular-weight fragments in the synovial fluid
samples and in cartilage explants stimulated with MMPs.
We only detected high-molecular-weight fragments from
aggrecanase-mediated degradation (Figure 3).
One possible explanation for this interesting discrepancy
is the differences between ex vivo and in vivo settings.
In the current experiments we used a synchronic
culture system, which presents a temporal predefined
order of events following induction of catabolic activity
against bovine cartilage (Karsdal 2007, 2008). This
experimental setting is in contrast to OA patients who
may have cartilage mal-metabolism ranging in severity at
different distinct local places in the same joint. Enzymes
are to some extent freely diffusible in the synovial fluid,
so damaged cartilage areas mediated by extensive MMP
activity may also ‘share’ some of these MMPs with other
areas which have less cartilage damage, and which may
primarily consist of aggrecanase activity. Such ‘enzyme
sharing’ in the local environment may indeed generate
the MMP-mediated high-molecular-weight fragments as
described by Struglics et al. (2006).

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
274 Suzi Hoegh Madsen et al.
Degradation fragments or activity
Time
Importantly, Struglics et al. (2006) also suggested that
aggrecanases cleaved MMP-mediated fragments in the
highly glycosylated region (between G2 and G3), and not
in the IGD. In addition, they presented data in agreement
with Fosang et al. (2000), who clearly demonstrated that
aggrecanases were unable to cleave 342 FFGVG fragments
at the NITEGE 373–374 ARGSVI site (Fosang et al. 2000). Our
data support these observations and conclusions ex vivo,
albeit not in vitro, as we observed that ADAMTS-4 was
capable of processing the MMP-mediated 342 FFGVG-G2
fragments further in the IGD (Figure 5). However, a high
concentration of aggrecanases was used to post-cleave
the fragments released into the supernatant, and thus
our data suggest that these fragments released ex vivo
were protected from aggrecanase-mediated cleavage at
the NITEGE 373–374 ARGSVI site. The molecular mechanism
behind this protection ex vivo is currently debated, but
it may be the lack of glycosylation of the MMP-mediated
aggrecan fragments. We support the suggestion by
Tottorella et al. (2000) that aggrecanases may favour
highly glycosylated forms of aggrecan, or depend on
the protein sequence upstream of the 374 ARGSVI site
for proper recognition and processing ex vivo. In fact,
another group reported that aggrecanase activity is promoted
by the presence of glycosylated aggrecan (Pratta
et al. 2000).
MMP activity
Aggrecanase
activity
342-G2
sGAG
374-G2
Figure 7. Schematic overview of aggrecanolysis in the explants model. The dotted lines represent protease activity of matrix metalloproteinase
(MMPs) and aggrecanases over time. The other lines represent the degradation fragments released from the cartilage explants to the conditioned
medium over time. Above the lines are cartoons of aggrecan attached to hyaluronic acid, which show the degradation and release of sulfated glycosaminoglycans
(sGAGs) from the aggrecan molecule over time. Aggrecanases cleave the glycosylated aggrecan molecules in the early culture
period, whereas MMPs degrade deglycosylated aggrecan molecules in the late culture period.
86
All the aforementioned biomarkers were measured in
conditioned medium synthesized by ex vivo cleavage. We
designed an in vitro experiment to investigate the differences
of in vitro and ex vivo cleavage in aggrecan. This
enabled us to investigate the different proteolytic affinities
for ADAMTS-4 and MMP-9 on native or deglycosylated
aggrecan, which we assessed by the 374 ARGSV-G2 and
342 FFGVG-G2 assays. The measurements showed significant
increased levels for 374 ARGSV-G2 and 342 FFGVG-G2
fragments when stimulated with ADAMTS-4 and MMP-9,
respectively. In vitro cleavage showed that 374 ARGSV-G2
levels were approximately 50 times greater than the
342 FFGVG-G2 levels. This difference between the two
neoepitope assays measured in the in vitro samples
resembled the difference we detected in ex vivo samples
(Figure 2A and B). This suggests that the approximately
ten times lower levels of 342 FFGVG-G2 compared with
374 ARGSV-G2 levels measured from the ex vivo cleavage,
were not due to limited aggrecan as a substrate, and therefore
the measurements ex vivo represent a reproducible
difference between the neoepitopes. Interestingly, the in
vitro measurements showed that ADAMTS-4 and MMP-9
both show a preference to degrade deglycosylated aggrecan
rather than native aggrecan. This is somewhat different
from the ex vivo measurements, which indicated
that ADAMTS-4 prefer to degrade glycosylated aggrecan

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
and MMP-9 favoured degrading deglycosylated aggrecan
(Figures 2 and 3). This highlights the need to distinguish
between in vitro and ex vivo results.
In conclusion, this study highlights the importance of
the separate and distinct proteolytic processing of aggrecan,
which varies over time. If the current findings are
to be useful in a clinical setting, it may be favourable to
target activity or expression of aggrecanases early in the
OA process, whereas at later stages, after the depletion of
sGAGs, it may be more efficacious to target expression or
activity of MMPs.
Acknowledgements
This study was supported in part by The Ministry of
Science Technology and Innovation, Denmark.
Declaration of interest
All authors declare that the affiliation declares full disclosure.
In addition, Drs Karsdal and Qvist own stock in
Nordic Bioscience.
References
Behrens F, Kraft EL, Oegema TR Jr. (1989). Biochemical changes in
articular cartilage after joint immobilization by casting or external
fixation. J Orthop Res 7:335–43.
Caterson B, Flannery CR, Hughes CE, Little CB. (2000). Mechanisms
involved in cartilage proteoglycan catabolism Matrix Biol
19:333–44.
Felson DT, Neogi T. (2004). Osteoarthritis: is it a disease of cartilage or
of bone? Arthritis Rheum 50:341–4.
Flannery CR, Lark MW, Sandy JD. (1992). Identification of a stromelysin
cleavage site within the interglobular domain of human
aggrecan. Evidence for proteolysis at this site in vivo in human
articular cartilage. J Biol Chem 267:1008–14.
Fosang AJ, Last K, Gardiner P, Jackson DC, Brown L. (1995).
Development of a cleavage-site-specific monoclonal antibody
for detecting metalloproteinase-derived aggrecan fragments:
detection of fragments in human synovial fluids. Biochem J
310:337–43.
Fosang AJ, Last K, Maciewicz RA. (1996). Aggrecan is degraded by
matrix metalloproteinases in human arthritis. Evidence that
matrix metalloproteinase and aggrecanase activities can be independent.
J Clin Invest 98:2292–9.
Fosang AJ, Last K, Stanton H, Weeks DB, Campbell IK, Hardingham
TE, Hembry RM. (2000). Generation and novel distribution of
matrix metalloproteinase-derived aggrecan fragments in porcine
cartilage explants. J Biol Chem 275:33027–37.
Glasson SS, Askew R, Sheppard B, Carito B, Blanchet T, Ma HL,
Flannery CR, Peluso D, Kanki K, Yang Z, Majumdar MK,
Morris EA. (2005). Deletion of active ADAMTS5 prevents cartilage
degradation in a murine model of osteoarthritis. Nature
434:644–8.
Hardingham TE, Fosang AJ. (1995). The structure of aggrecan and its
turnover in cartilage. J Rheumatol Suppl 43:86–90.
Ilic MZ, Haynes SR, Winter GM, Handley CJ. (1995). Kinetics of release
of aggrecan from explant cultures of bovine cartilage from different
sources and from animals of different ages. Acta Orthop
Scand Suppl 266:33–7.
Aggrecanase- and MMP-mediated aggrecan degradation 275
87
Karsdal MA, Madsen SH, Christiansen C, Henriksen K, Fosang AJ,
Sondergaard BC. (2008). Cartilage degradation is fully reversible
in the presence of aggrecanase but not matrix metalloproteinase
activity. Arthritis Res Ther 10:R63.
Karsdal MA, Sumer EU, Wulf H, Madsen SH, Christiansen C, Fosang
AJ, Sondergaard BC. (2007). Induction of increased cAMP levels
in articular chondrocytes blocks matrix metalloproteinasemediated
cartilage degradation but not aggrecanase-mediated
cartilage degradation. Arthritis Rheum 56:1549–58.
Kiani C, Chen L, Wu YJ, Yee AJ, Yang BB. (2002). Structure and function
of aggrecan. Cell Res 12:19–32.
Lark MW, Bayne EK, Flanagan J, Harper CF, Hoerrner LA, Hutchinson
NI, Singer II, Donatelli SA, Weidner JR, Williams HR, Mumford
RA, Lohmander LS. (1997). Aggrecan degradation in human cartilage.
Evidence for both matrix metalloproteinase and aggrecanase
activity in normal osteoarthritic and rheumatoid joints. J
Clin Invest 100:93–106.
Little CB, Flannery C R, Hughes CE, Mort JS, Roughley PJ, Dent C,
Caterson B. (1999). Aggrecanase versus matrix metalloproteinases
in the catabolism of the interglobular domain of aggrecan
in vitro. Biochem J 344:61–8.
Lohmander S, Antonopoulos CA, Friberg U. (1973). Chemical and
metabolic heterogeneity of chondroitin sulfate and keratin
sulfate in guinea pig cartilage and nucleus pulposus. Biochim
Biophys Acta 304:430–48.
Maehara H, Suzuki K, Sasaki T, Oshita H, Wada E, Inoue T, Shimizu
K. (2007). G1-G2 aggrecan product that can be generated by
M-calpain on truncation at Ala709-Ala710 is present abundantly
in human articular cartilage. J Biochem 141:469–77.
Malfait AM, Liu RQ, Ijiri K, Komiya S, Tortorella MD. (2002). Inhibition
of ADAM-TS4 and ADAM-TS5 prevents aggrecan degradation in
osteoarthritic cartilage. J Biol Chem 277:22201–8.
Mercuri FA, Maciewicz RA, Tart J, Last K, Fosang AJ. (2000). Mutations
in the interglobular domain of aggrecan alter matrix metalloproteinase
and aggrecanase cleavage patterns. Evidence that matrix
metalloproteinase cleavage interferes with aggrecanase activity.
J Biol Chem 275:33038–45.
Mok SS, Masuda K, Hauselmann HJ, Aydelotte MB, Thonar EJ. (1994).
Aggrecan synthesized by mature bovine chondrocytes suspended
in alginate. Identification of two distinct metabolic matrix pools.
J Biol Chem 269:33021–7.
Oshita H, Sandy JD, Suzuki K, Akaike A, Bai Y, Sasaki T, Shimizu K.
(2004). Mature bovine articular cartilage contains abundant
aggrecan that is C-terminally truncated at Ala719-Ala720 a site
which is readily cleaved by m-calpain. Biochem J 382:253–9.
Pratta MA, Tortorella MD, Arner EC. (2000). Age-related changes in
aggrecan glycosylation affect cleavage by aggrecanase. J Biol
Chem 275:39096–102.
Roughley PJ, Lee ER. (1994). Cartilage proteoglycans: structure and
potential functions. Microsc Res Tech 28:385–97.
Rousseau JC, Sumer EU, Hein G, Sondergaard BC, Madsen SH,
Pedersen C, Neumann T, Mueller A, Qvist P, Delmas P, Karsdal
MA. (2008). Patients with rheumatoid arthritis have an altered
circulatory aggrecan profile. BMC Musculoskelet Disord 9:74.
Sandy JD. (2006). A contentious issue finds some clarity: on the
independent and complementary roles of aggrecanase activity
and MMP activity in human joint aggrecanolysis. Osteoarthritis
Cartilage 14:95–100.
Sandy JD, Flannery CR, Neame PJ, Lohmander LS. (1992). 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. J
Clin Invest 89:1512–16.
Sondergaard BC, Henriksen K, Wulf H, Oestergaard S, Schurigt U,
Brauer R, Danielsen I, Christiansen C, Qvist P, Karsdal MA.
(2006a). Relative contribution of matrix metalloproteinase and
cysteine protease activities to cytokine-stimulated articular cartilage
degradation. Osteoarthritis Cartilage 14:738–48.
Sondergaard BC, Wulf H, Henriksen K, Schaller S, Oestergaard S, Qvist
P, Tanko LB, Bagger YZ, Christiansen C, Karsdal MA. (2006b).
Calcitonin directly attenuates collagen type II degradation by
inhibition of matrix metalloproteinase expression and activity
in articular chondrocytes. Osteoarthritis Cartilage 14:759–68.

Biomarkers Downloaded from informahealthcare.com by Panum Biblioteket on 08/31/10
For personal use only.
276 Suzi Hoegh Madsen et al.
Stanton H, Rogerson FM, East CJ, Golub SB, Lawlor KE, Meeker CT,
Little CB, Last K, Farmer PJ, Campbell IK, Fourie AM, Fosang AJ.
(2005). ADAMTS5 is the major aggrecanase in mouse cartilage in
vivo and in vitro. Nature 434:648–52.
Struglics A, Larsson S, Pratta MA, Kumar S, Lark MW, Lohmander
LS. (2006). Human osteoarthritis synovial fluid and joint cartilage
contain both aggrecanase- and matrix metalloproteinase-generated
aggrecan fragments. Osteoarthritis Cartilage
14:101–13.
Sumer EU, Sondergaard BC, Rousseau JC, Delmas PD, Fosang AJ,
Karsdal MA, Christiansen C, Qvist P. (2007). MMP and non-
MMP-mediated release of aggrecan and its fragments from
articular cartilage: a comparative study of three different
88
aggrecan and glycosaminoglycan assays. Osteoarthritis Cartilage
15:212–21.
Sztrolovics R, White RJ, Roughley PJ, Mort JS. (2002). The mechanism
of aggrecan release from cartilage differs with tissue origin and
the agent used to stimulate catabolism. Biochem J 362:465–72.
Tortorella M, Pratta M, Liu RQ, Abbaszade I, Ross H, Burn T, Arner E.
(2000). The thrombospondin motif of aggrecanase-1 (ADAMTS-4)
is critical for aggrecan substrate recognition and cleavage. J Biol
Chem 275:25791–7.
Tortorella MD, Burn TC, Pratta MA, Abbaszade I, Hollis JM, Liu R,
Rosenfeld SA, Copeland RA, Decicco CP, Wynn R and others.
(1999). Purification and cloning of aggrecanase-1: a member of
the ADAMTS family of proteins. Science 284:1664–6.

CHAPTER 7
CHAPTER 7: PAPER IV
PAPER IV
___________________________________________________________________________
Introductory remarks
This manuscript describes the endogenous proteases in human OA cartilage, and the biochemical
markers that are able to assess the different proteolytic pathways for both collagen type II and
aggrecan degradation. More investigation on this topic may have significant value for
development of new treatments for OA.
89

CHAPTER 7: PAPER IV
Metalloproteinase and cysteine protease
profiling in normal and Osteoarthritic cartilage
Suzi Hoegh Madsen 1 , Kim Henriksen 3 , Søren Tvorup Christensen 2 , Morten Asser Karsdal 1 , and
Anne-Christine Bay-Jensen 1
In review in Proteome Science
1 Cartilage Biology and Biomarkers R&D, Nordic Bioscience A/S, Herlev, Denmark
2 Department of Molecular Biology, University of Copenhagen, Copenhagen, Denmark
3 Bone Biology R&D, Nordic Bioscience A/S, Herlev, Denmark
___________________________________________________________________________
Abstract
Treatment for cartilage degenerating diseases is either analgesic or joint replacement. Thus, the
discovery and development of new treatments are needed. Proteases, such as metalloproteinases
and cathepsins, play a major role in the net degradation of cartilage, and are therefore a natural
target for drug development. By testing the ability of a library of proteases to degrade cartilage
explants, and by inhibiting one type of protease, we aim to find the proteases responsible for
collagen type II and aggrecan degradation and possible masking effects by other proteases.
The cysteine inhibitor, E64, and the proteases; MMP-3 and -7, and ADAMTS-4
and -5, all increased aggrecan degradation (AGNx-II), in human metabolic inactivated
osteoarthritic (OA) cartilage. The MMP inhibitor, GM6001, abrogated all the increased aggrecan
degradation. Collagen type II was degraded with MMP-7 (CTX-II) and MMP-9 (NBC2), which
both could be inhibited by GM6001. Human OA cartilage cultured in buffer optimal for cysteine
proteases showed an increase in levels of CTX-II and NBC2 fragments when incubated with
GM6001. Furthermore, we observed altered levels of the biomarkers between normal and pre-
cytokine-stimulated bovine cartilage, indicating an induced alteration of the endogenous
proteases.
In human OA cartilage, the degradation markers are released in part by different
proteolytic pathways. Inhibition of cysteine proteases unmasks the levels of MMP-mediated
aggrecan degradation in OA cartilage. Inhibition of MMPs also unmasks the levels of the collagen
type II degradation mediated by non-MMPs. The role of endogenous proteases probably
depends on the stage of OA, as we did not observe the same release of biomarkers in normal
versus cytokine-stimulated and OA cartilage.
90

Introduction
CHAPTER 7: PAPER IV
Cartilage turnover is a complex process in which proteases play a prominent role in both health
and disease. The extracellular matrix (ECM) in healthy articular cartilage is expressed and
regulated by the chondrocytes, which include expression of matrix proteins, proteases, and
protease inhibitors 1 . One of the major mediators of cartilage destruction in osteoarthritis (OA) is
the metalloproteinases; matrix metalloproteinases (MMPs) and A Disintegrin And
Metalloproteinase with Thrombospondin Motifs (ADAMTS’) 2 . In few words, the destruction of
ECM in OA results from initial aggrecan degradation by ADAMTS’ and subsequently
degradation by diverse MMPs that continues to degrade the major components of the ECM, e.g.
collagen type II and aggrecan 2,3,4 . However, cysteine proteases are also essential players in the
proteolytic cleavage of ECM in OA, which are synthesized both by the chondrocytes and
synovial cells in response to cytokines and growth factors 5 .
Currently, treatment of cartilage degenerative diseases is either analgesic or joint
replacement. Thus, the discovery and development of new treatments are needed. Since proteases
- such as MMPs and cathepsins - play a major role in the net degradation of cartilage, they are a
natural target for drug development. Early treatment with protease inhibitors to either reduce
inflammation and/or reduce aggrecan/collagen loss may have the potential to reduce the onset
of future idiopathic or posttraumatic osteoarthritis. However, to target the right protease at the
right time during the progression of OA, a better understanding of the proteolytic processes is
needed. ECM-neoepitope biomarkers are great tools for investigating the proteolytic processes 6 .
The collagen type II degradation marker, C-terminal telopeptide degradation fragment (CTX-II) 7
and the aggrecan degradation marker, AGNx-II 8 , are both MMP-specific. The novel collagen
type II intra-helical degradation markers, NBC2, are supposed to represent cathepsin-specificity.
However, the biomarker will be evaluated in the present study.
In this study, we identified which MMPs, ADAMTS’ and cathepsins were able to
generate CTX-II, NBC2, and AGNx-II fragments using an in vitro and ex vivo model of human
OA cartilage. Such data could provide information on the specificity of the different proteases
for cleavage of collagen and aggrecan, and also inform us about the proteolytic pathways for each
biomarker release. Furthermore, the endogenous proteases in normal bovine cartilage were
compared with cytokine-stimulated bovine and human OA cartilage. The cytokines, oncostatin M
(OSM) and tumor necrosis factor-α (TNF-α), are inducers of cartilage degradation in vivo and in
vitro 9 , and would allow us to induce a controlled catabolic system towards the one seen in human
OA cartilage. I.e. induction of aggrecanases after 4 days with cytokine stimulation and induction
of MMPs after 11 days with cytokine stimulation 3 . By inhibiting one type of protease, we aimed
to find the proteases responsible for collagen type II and aggrecan degradation and possible
masking effects by other proteases.
91

Methods
Sources of proteases, inhibitors and cytokines
CHAPTER 7: PAPER IV
Purified human matrix metalloproteinase-1 (MMP-1) and recombinant human MMP-7 were
purchased from Enzo life science, DK. Recombinant human Cathepsin-L and purified human
Cathepsin-S and Cathepsin-B were from BioVision, US. Recombinant human aggrecanases;
ADAMTS-4 and ADAMTS-5, were from Chemicon international, DK. Purified human MMP-9
and recombinant human MMP-3, MMP-13 and Cathepsin-K, were purchased from Biocol, DE.
The matrix metalloprotease inhibitor (R)-N4-Hydroxy-N1-[(S)-2-(1H-indol-3-yl)-1-
methylcarbamoyl-ethyl]-2-isobutyl-succinamide (GM6001) and the cysteine protease inhibitor
trans-Epoxysuccinyl-L-leucylamido(4-guanidino)butane, L-trans-3-Carboxyoxiran-2-carbonyl-L-
leucylagmatine, N-(trans-Epoxysuccinyl)-L-leucine 4-guanidinobutylamide (E64) were both
purchased from Sigma, DK. The cytokines, oncostatin M (OSM) was from Sigma, DK, and the
Tumor Necrosis Factor-α (TNF-α) was from R&D systems, UK.
Buffers
To evaluate endogenous activity of metalloproteinases (MMPs and ADAMTS’), cartilage explants
were incubated in an MMP buffer [50 mM Tris-HCL, 200 mM NaCl, 5 mM CaCl 2, 1 µM ZnCl 2,
0.05% NaN 3, 0.05% Brij-35, pH 7]. Activation of latent MMPs was performed by adding 1 mM
freshly made p-Aminophenylmercurie (APMA). To evaluate endogenous activity of cysteine
proteinases, cartilage explants were incubated in a cysteine buffer [100 mM Na-Acetate, 5 mM
EDTA, 0.05% NaN 3, pH 5]. Activation of latent cysteine proteinases was performed by adding
20 mM cysteine just before use. Both buffers were prepared with inspiration from Tabassi et al. 13 .
Ex vivo models
Human articular cartilage experiments
The human OA cartilage was obtained from female patients between the ages of 65-77 years
undergoing total knee arthroplasty. The patients were informed by both oral and written
communication at the hospital and gave their written consent before entering study, which was
approved by the Capital Region of Denmark, DK-3400, approval number HD-2007-0084. The
cartilage was metabolically inactivated (MI) by freeze-thaw cycles and cut into homologous
explants. Half of the MI explants were incubated in MMP buffer for 3, 6, and 9 h in the presence
or absence of the inhibitors, GM6001 [10 µM] and E64 [40uM], and the proteases [100 ng/ml];
MMP-1, -3, -7, -9, -13 and ADAMTS-4, -5. The other half were incubated in cysteine buffer for
3, 6, and 9 h in the presence or absence of the inhibitors, GM6001 [10 µM] and E64 [40 µM], and
the proteases [100 ng/ml]; Cathepsin-K, -B, -L, -S. At the end of each incubation time (3, 6, 9h),
conditioned buffer (about 300 µl) was extracted and stored frozen at -20°C until assayed for
specific biomarkers.
92

Bovine articular cartilage experiments
CHAPTER 7: PAPER IV
Bovine knees (Hereford) were bought at the local slaughter house (Slangerup, DK) just after
slaughtering. Articular cartilage explants were isolated under semi-sterile conditions and washed
in PBS. One third of the cartilage explants were MI (t=0) and stored in -80°C until use. The rest
of the cartilage explants were pre-cultured with pro-inflammatory cytokines, OSM [10 ng/ml]
and TNF-α [20 ng/ml] (O+T), for 4 or 11 days to increase aggrecanase or MMP activity,
respectively 3,4 . The cartilage explants pre-cultured in cytokines for 4 or 11 days were after each
termination washed in PBS and MI, and stored in -80°C until use. The frozen explants without
cytokine pre-culture (t=0) and the pre-cultured catabolic induced explants (t=4 and t=11), were
thawed and incubated in MMP or cysteine buffer for 3, 6, and 9 h in the presence or absence of
the inhibitors, GM6001 [10 µM] and E64 [40 µM]. At the end of each incubation time (3h, 6h,
9h), conditioned buffer (about 300 µl) was extracted and stored frozen at -20°C until assayed for
specific biomarkers.
Biomarkers of cartilage degradation
The type II collagen telopeptide fragment, CTX-II (urine cartilaps, IDS Ltd, DK), were measured
accordingly to the manufacturer’s instructions 7 . AGNx-II measures MMP-mediated aggrecan
degradation accordingly to Sumer et al. 8 . The type II collagen inter helical neo-epitope fragment
(…PKGARG 707 ), NBC2, was measured accordingly to Bay-Jensen et al. 2011 (unpublished).
Statistic analysis
Results in fig. 1 are shown as mean + standard error of mean (SEM). Statistic analysis in fig. 1
was performed by ANOVA; nonparametric test, don’t assuming Gaussian distribution, n=4,
Kruskal-Wallis test with Dunn’s multiple comparison post test. Statistic significance is indicated
by the number of asterisks, * p < 0.05, ** p < 0.01, and *** p < 0.001. The Tables show mean
value of n=4 ± standard deviation (SD).
Results
Detection of key proteases to degrade human OA cartilage ex vivo; evaluated by collagen type II
degradation markers
Human OA cartilage explants incubated in buffer optimal for MMPs, increased the levels of
CTX-II when incubated with MMP-7 (p

AGNx-II ng/mL
CTX-II ng/mL
NBC2 ng/mL
2.4
1.8
1.2
0.6
Vehicle
MMP1
MMP3
MMP buffer
**
MMP7
MMP9
MMP13
CHAPTER 7: PAPER IV
A B
C
E
15
10
5
12000
9000
6000
3000
Vehicle
Vehicle
MMP1
MMP1
MMP3
MMP buffer
***
MMP3
MMP7
*
*
MMP9
MMP13
MMP buffer
*
MMP7
MMP9
ADAMTS4
ADAMTS4
MMP13
ADAMTS5
ADAMTS5
*
ADAMTS4
94
*
ADAMTS5
NBC2 ng/mL
CTX-II ng/mL
1.6
1.2
0.8
0.4
200
150
100
50
Vehicle
Vehicle
Vehicle
Cysteine buffer
CatK
*
CatB
CatL
Cysteine buffer
CatK
CatB
CatL
Cysteine buffer
Fig. 1. Collagen type II and aggrecan degradation are mediated by different pathways.
Human OA cartilage explants were incubated in MMP and cysteine buffer in the presence or
absence of different metalloproteinases [100 ng/ml] and cysteine proteases [100 ng/ml] (Cat =
cathepsin). Biomarkers for collagen type II degradation (CTX-II and NBC2) and aggrecan
degradation (AGNx-II) were measured after 3, 6 and 9h. Data are shown as accumulated for the
three measurements. N=4.
AGNx-II ng/ml
D
F
400
300
200
100
CatK
CatB
CatL
CatS
CatS
CatS

CHAPTER 7: PAPER IV
Detection of key proteases to degrade human OA cartilage ex vivo; evaluated by an aggrecan
degradation marker
In human metabolic inactivated OA cartilage incubated in MMP buffer, MMP-3 and -7, and
ADAMTS-4 and -5 all increased aggrecan degradation (AGNx-II) compared to non-stimulated
vehicle (fig. 1E). Addition of the MMP inhibitor, GM6001, to the proteases attenuated this
release of the aggrecan fragments (data not shown). We did not observe any changes in the
release of AGNx-II fragments by MMP-1, -9 and -13, compared to vehicle (fig. 1E).
The cathepsins did not affect the levels of AGNx-II significantly; however, we
observed a tendency towards a decrease in the levels of AGNx-II compared to vehicle (fig. 1F).
Collagen type II degradation in normal, cytokine-induced, and OA cartilage
In normal bovine cartilage, the endogenous proteinases activated by the MMP-buffer, degraded
the collagen type II in to fragments that were able to be detected by the biomarker, CTX-II. The
MMP inhibitor, GM6001, was able to attenuate CTX-II (Table 1). The bovine cartilage pre-
cultured with cytokines for 4 and 11 days before metabolic inactivated and incubated in MMP
buffer, increased the levels of CTX-II compared to normal bovine cartilage, which were also
attenuated in the presence of GM6001. However, the human OA cartilage did not generate CTX-
II fragments as in the bovine samples. The cysteine protease inhibitor, E64, did not affect the
CTX-II levels. The biomarker, NBC2, was not detectable in the MMP buffer (Table 1).
The proteinases activated by the cysteine buffer in both bovine and human
cartilage, generated fragments that could be detected by both biomarkers of collagen type II
degradation; CTX-II and NBC2 (Table 2). The MMP inhibitor, GM6001 showed a tendency to
decrease CTX-II levels (P=0.057) in normal bovine cartilage, whereas the same inhibitor showed
a tendency to increase CTX-II levels in pre-cultured bovine cartilage (11 days) and in human OA
cartilage (Table 2). Additionally, GM6001 increased NBC2 in human OA and bovine normal and
pre-cultured cartilage. The cysteine protease inhibitor, E64, decreased the levels of CTX-II in
human OA and bovine normal and pre-cultured (11 days) cartilage. E64 also decreased the levels
of NBC2 in normal bovine cartilage, whereas the cysteine protease inhibitor increased the levels
of NBC2 in bovine cartilage pre-cultured with cytokines for 4 days (Table 2). The bovine
cartilage pre-cultured with cytokines for 4 and 11 days before metabolic inactivated and
incubated in cysteine buffer, decreased the levels of CTX-II compared to normal bovine cartilage
(Table 2), which was the opposite effects from the one seen in MMP buffer (Table 1).
Furthermore, the human OA cartilage had CTX-II levels corresponding to the cartilage pre-
cultured with cytokines for 11 days (Table 2). Moreover, the levels of NBC2 were also lower for
the pre-cultured cartilage (11 days) compared with the normal cartilage. However, the human OA
cartilage had greatly higher levels of NBC2 compared to bovine cartilage in general (Table 2).
95

CHAPTER 7: PAPER IV
Aggrecan degradation in normal, cytokine-induced, and OA cartilage
In both bovine and human OA cartilage, the endogenous proteinases activated by MMP-buffer,
degraded the aggrecan in to fragments that were able to be detected by the biomarker, AGNx-II.
The MMP inhibitor, GM6001, was able to attenuate the levels of AGNx-II in both bovine and
human OA cartilage (Table 1). The bovine cartilage pre-cultured with cytokines for 4 and 11
days before metabolic inactivation and incubation in MMP buffer, increased the levels of AGNx-
II compared to normal bovine cartilage (Table 1). Furthermore, the human OA cartilage had
greatly higher levels of AGNx-II than the bovine samples (Table 1). The cysteine protease
inhibitor, E64, did not affect the normal or pre-cultured bovine samples, whereas E64 increased
the levels significantly in the human OA samples (Table 1).
The proteinases activated by the cysteine buffer in both bovine and human cartilage, also
degraded the aggrecan in to fragments that were able to be detected by the biomarker, AGNx-II.
96

CHAPTER 7: PAPER IV
The MMP inhibitor, GM6001, was able to attenuate the levels of AGNx-II in normal bovine
cartilage, whereas the pre-cultured bovine cartilage was not affected by the inhibitor (Table 2).
The bovine cartilage pre-cultured with cytokines for 4 and 11 days before metabolic inactivation
and incubation with cysteine buffer, decreased the levels of AGNx-II compared to normal
bovine cartilage (Table 2), which was the opposite effects from the one seen in the MMP buffer.
Furthermore, the human OA cartilage had levels of AGNx-II as the normal bovine cartilage
(Table 2). The cysteine protease inhibitor, E64, did not affect the normal or pre-cultured bovine
samples, whereas E64 increased the levels significantly in the human OA samples (Table 2).
Discussion
The aim of our study was to understand the molecular foundation for the progression of articular
cartilage degradation that leads to OA. We analyzed the release of cartilage degradation
biomarkers from human OA cartilage by the proteinases reported to play a role in cartilage
degradation 5,10 . Furthermore, we investigated how diseased cartilage is altered from normal
cartilage with regards to the endogenous proteases, and we were observant to possible masking
effects among the proteases.
We showed that the cartilage degradation markers, CTX-II, NBC2, and AGNx-II,
are released in part by different proteolytic pathways in human OA cartilage, as different
proteases generated different biomarkers (fig. 1). The biomarkers were only released by specific
metalloproteinases, and none of the supplied cathepsins (fig. 1). Moreover, Cathepsin-B (and a
tendency for Cathepsin-K and Cathepsin-L) inhibited the release of CTX-II in cysteine buffer,
suggesting either a protective effect by the cathepsins on cartilage degradation or more likely a
masking effect (fig. 2) from post-cleavage of CTX-II fragments by the cathepsins. The aggrecan
degradation increased in the presence of both MMPs and aggrecanases, although AGNx-II is
MMP-specific (fig. 1G), suggesting pre- or post-cleavages by the ADAMTS’ may aid the
degradation.
In human OA cartilage explants, the collagen type II and aggrecan may already be
severely degraded by the increased production of proteases during the disease 10 , so further
protease-cleavage ex vivo may be difficult to detect, as specific targets measurable by our
biomarkers have already been cleaved and further processed, thereby masking the effect from the
MMPs and cysteine proteases we tested. This possible masking effect could be investigated by a
follow up study, testing the library of proteases in normal cartilage and comparing the results
with those from the present study in human OA cartilage.
The endogenous proteases in human OA samples activated in the MMP buffer
(Table 1) failed to increase the levels of CTX-II and NBC2. One suggestion could be that the
damaged OA cartilage had less chondrocytes and ECM proteins, and thus fewer enzymes
present, compared to normal cartilage. This is supported by the increased levels of CTX-II and
NBC2, when supplying MMP-7 and MMP-9 (fig.1). However, as we were able to decrease MMP-
97

CHAPTER 7: PAPER IV
mediated aggrecan degradation with GM6001 in the MMP buffer, demonstrating that the human
OA cartilage is not depleted for metalloproteinases (table 1).
Fig. 2. Schematic overview of the masking effect of endogenous
proteases in human OA cartilage. The MMP-mediated neoepitope,
AGNx-II, is further processed by cysteine proteases. Thus,
Inhibition of cysteine proteases result in increased AGNx-II.
Several studies by our group have shown that a combination of the cytokines, OSM + TNF-α,
induce a damaging effect on cartilage by increasing the production of proteases, e.g. aggrecanases
and MMPs 3,4 . This induces a shift from normal turnover towards an uneven turnover - resulting
in excessive protein degradation. In the present study, we pre-stimulated normal bovine cartilage
explants with the cytokines to induce a catabolic system that resemble human OA cartilage
regarding the proteases. This would allow us to have a system representing the proteases in
normal and diseased cartilage, from the same cartilage explant. The cartilage explant pre-
stimulated with cytokines for 4 days increased the aggrecanase-mediated aggrecan degradation,
whereas the cartilage pre-stimulated with cytokines for 11 days increased the MMP-mediated
aggrecan degradation, and not vice versa (data not shown). This confirm that our proteases has
been induced as expected 3 .
98

CHAPTER 7: PAPER IV
The pre-stimulated cartilage incubated in MMP buffer increased the levels of CTX-II, and
AGNx-II, suggesting aggrecanases and MMPs to be the key proteases to generate these two
biomarkers. The MMP inhibitor, GM6001, decreased the increased levels of the two biomarkers
in both 4 and 11 days pre-stimulated cartilage explants, which correlates with the literature that
GM6001 inhibits both aggrecanases and MMPs 11 . The biomarker levels varied significantly
between the normal, cytokine-stimulated and OA cartilage explants, showing that the pre-
stimulated bovine cartilage explants did not represent human OA cartilage. However, the results
informed us instead that the presence and role of endogenous proteases vary from healthy
cartilage and during the progression of OA. Thus it is important to target the right proteases at
the right time, to receive the optimal treatment.
E64 indicated that cysteine proteases mask the levels of the biomarker representing
MMP-mediated aggrecan degradation, whereas GM6001 indicated that MMPs masks the levels of
the collagen type II degradation biomarkers in human OA (Table1/2). Sondergaard et al. 12 have
shown that CTX-II levels increases in the presence of E64 in bovine cartilage explants (living
tissue), suggesting that it is a compensatory degrading effect from other proteases, like MMPs,
when cysteine proteases are absent to degrade the proteins. However, it could also be a masking
effect from the cysteine proteases that cleaves MMP-mediated CTX-II fragments further, and
thus masks the levels, which is actually degraded by the MMPs. Our study confirms that the study
by Sondergaard et al. 12 must have been a compensatory effect, as we did not see any increase of
CTX-II by E64 in cytokine-stimulated bovine explants, in which the chondrocytes were
inactivated. If E64 should have conducted a masking effect in the Sondergaard 12 study, our E64
should have been increased as well.
The present brief report will be followed by a thorough study with regards to the
dynamics of proteases in diseased and normal cartilage, and potential compensatory effects of the
different proteases herein. A complete understanding of the common molecular sequence of
events underlying the development of OA will expand our knowledge of the pathogenesis of
OA. These events will also contribute with important information to search for novel therapeutic
targets for the prevention and treatment of OA at the right time during the progression of the
disease.
Conclusions
Taken together, these experiments suggest that CTX-II, NBC2 and AGNx-II are released in part
by different proteolytic pathways in human OA cartilage. Furthermore, the presence and role of
endogenous proteases probably depend on the stage of OA, as we did not observe the same
release of biomarkers during inhibition of one type of protease, in normal versus cytokine-
stimulated and OA cartilage. Thus more investigation on this topic may have significant value for
development of new treatments for OA.
99

References for CHAPTER 7 (PAPER IV)
1 Archer C.W. and Francis-West P. The
chondrocyte. Int J Biochem Cell Biol.
2003;35(4):401-404.
2 Rengel Y., Ospelt C., and Gay S. Proteinases
in the joint: clinical relevance of proteinases in
joint destruction. Arthritis Res Ther.
2007;9(5):221-21
3 Madsen S.H., Sumer E.U., Bay-Jensen A.C.,
Sondergaard B.C., Qvist P., and Karsdal M.A.
Aggrecanase- and matrix metalloproteinasemediated
aggrecan degradation is associated
with different molecular characteristics of
aggrecan and separated in time ex vivo.
Biomarkers. 2010;15(3):266-276.
4 Karsdal M.A., Madsen S.H., Christiansen C.,
Henriksen K., Fosang A.J., and Sondergaard
B.C. Cartilage degradation is fully reversible in
the presence of aggrecanase but not matrix
metalloproteinase activity. Arthritis Res Ther.
2008;10(3):R63-
5 Morko J.P., Soderstrom M., Saamanen A.M.,
Salminen H.J., and Vuorio E.I. Up regulation
of cathepsin K expression in articular
chondrocytes in a transgenic mouse model for
osteoarthritis. Ann Rheum Dis.
2004;63(6):649-655.
6 Leeming D.J., Bay-Jensen A.C., Vassiliadis E.,
Larsen M.R., Henriksen K., and Karsdal M.A.
Post-translational modifications of the
extracellular matrix are key events in cancer
progression: opportunities for biochemical
marker development. Biomarkers.
2011;16(3):193-205.
7 Olsen A.K., Sondergaard B.C., Byrjalsen I.,
Tanko L.B., Christiansen C., Muller A., Hein
G.E. et al. Anabolic and catabolic function of
chondrocyte ex vivo is reflected by the
metabolic processing of type II collagen.
Osteoarthritis Cartilage. 2007;15(3):335-342.
CHAPTER 7: PAPER IV
100
8 Sumer E.U., Sondergaard B.C., Rousseau J.C.,
Delmas P.D., Fosang A.J., Karsdal M.A.,
Christiansen C. et al. MMP and non-MMP
mediated release of aggrecan and its fragments
from articular cartilage: a comparative study of
three different aggrecan and
glycosaminoglycan assays. Osteoarthritis
Cartilage. 2007;15(2):212-221.
9 Hui W., Rowan A.D., Richards C.D., and
Cawston T.E. Oncostatin M in combination
with tumor necrosis factor alpha induces
cartilage damage and matrix metalloproteinase
expression in vitro and in vivo. Arthritis
Rheum. 2003;48(12):3404-3418.
10 Mort J.S. and Billington C.J. Articular cartilage
and changes in arthritis: matrix degradation.
Arthritis Res. 2001;3(6):337-341.
11 Burrage P.S. and Brinckerhoff C.E. Molecular
targets in osteoarthritis: metalloproteinases
and their inhibitors. Curr Drug Targets.
2007;8(2):293-303.
12 Sondergaard B.C., Henriksen K., Wulf H.,
Oestergaard S., Schurigt U., Brauer R.,
Danielsen I. et al. Relative contribution of
matrix metalloprotease and cysteine protease
activities to cytokine-stimulated articular
cartilage degradation. Osteoarthritis Cartilage.
2006;14(8):738-748.
13 Charni-Ben T.N., Desmarais S., Bay-Jensen
A.C., Delaisse J.M., Percival M.D., and
Garnero P. The type II collagen fragments
Helix-II and CTX-II reveal different
enzymatic pathways of human cartilage
collagen degradation. Osteoarthritis Cartilage.
2008;16(10):1183-1191.

CHAPTER 8
CHAPTER 8: General discussion and conclusions
General discussion and conclusions
The main aim of this PhD project was to investigate the pathogenesis of OA with respect to
bone and cartilage and the importance of the cellular interactions between these two tissues.
One objective was to develop a pre-clinical model that can be used partly to
investigate the interaction between osteoblasts, osteoclasts, and chondrocytes, and partly to test
potential drugs for OA. This objective was successfully achieved by the development of the pre-
clinical ex vivo model that included both subchondral bone and articular cartilage (fig. 13). Femur
heads isolated from 3- to 12-week-old female mice were cultured as explants. These femur heads
comprise a closed system, which are easy to isolate from their host without damaging the
explants. Furthermore, the use of different ages of mice in the model allowed us to monitor
changes simultaneously in both bone and cartilage, as the mice age. Hence, the femur head ex vivo
model has advantages compared to the cartilage ex vivo model (only cartilage). However, the
femur head model also has a disadvantage. Even though we characterized different ages of mice,
we were not able to mimic human mature articular cartilage. Mice are short-lived (compared to
human) and always in development, and therefore mimics immature cartilage. However, the
immature cartilage in the murine femur heads contains hypertrophic chondrocytes, which are a
hallmark of human OA, suggesting that this system could serve as an important disease model. It
is debateable whether the hypertrophic chondrocytes from human OA are the same hypertrophic
chondrocytes we see in the growth plate of growing mice.
The established novel ex vivo femur head model was used as a screening model, to
test prednisolone. Our results showed that prednisolone decreased cartilage degradation and
decreased the turnover of bone (fig. 13). We also tested the effects of prednisolone and
dexamethasone on bone or cartilage individually in other pre-clinical models, which confirmed
that GCs protect cartilage, but not bone. Additionally, GCs did not appear to affect the lifespan
and function of resting chondrocyte (fig. 13). These results suggest that the use of GC injections
for OA is not preferable without considering the damaging effects on bone. Data obtained from
this study also propose that the interaction between bone and cartilage is of less importance in
regard to GC-treatment.
Another objective was to gain greater insight into the proteolytic degradation
processes of pathogenic cartilage since cartilage degradation and subsequently loss of the tissue is
one of the hallmarks of the pathogenesis of OA. Loss of sGAGs from aggrecan is one of the first
steps in the pathogenesis of OA. We found that there are different molecular structured pools of
aggrecan that are degraded by different proteases (fig. 13), and we suggest that fully saturated
aggrecan molecules, which are normally found in normal cartilage, are preferentially degraded by
101

CHAPTER 8: General discussion and conclusions
aggrecanases. In contrast, MMPs primarily degrade low saturated aggrecan molecule. This
suggests that it may be favourable to target aggrecanases in early OA and MMPs in later stages of
OA.
We have also hypothesized that the endogenous proteases found in human OA
depend on the stage of disease. This assumption is based on the fact that MMPs, and probably
several other proteases, increase over time during the development of OA. However, the
importance of cathepsins in this relation is unknown. Thus we used biochemical markers to
assess the different endogenous proteases in normal vs. OA-cartilage. We found that CTX-II,
NBC2 and AGNx-II are released in part by different proteolytic pathways in human OA
cartilage. In addition, we suggest that the presence and role of endogenous proteases depend on
the stage of OA (fig. 13). Further investigations on this topic may have significant value for
development of new treatments for OA.
Figure 13. Schematic illustration summarizing the four papers. The figure shows a femur head divided in a
diseased and a normal half. The development of the whole femur head model is seen in PAPER I.
Glucocorticoids (GC) decrease the degradation of damaged cartilage, but do not affect normal cartilage. GCs
induce apoptosis in both normal and diseased bone (PAPER II). Aggrecan is divided in two different pools, with
different molecular structures, which are degraded by different proteases (PAPER III). Proteases vary between
normal and OA cartilage. Furthermore, the presence and role of endogenous proteases probably depend on the
stage of OA (PAPER IV). The figure was produced by Madsen, S.H.
102

Future perspectives
CHAPTER 8: Future perspectives
One of the purposes for developing the femur head model was to investigate the cellular
interactions between bone and cartilage, although the GC-study did not provide evidence for
such an interaction between the two tissues. Evidently, this needs to be investigated in more
detail in the femur head model. In particular, it is not unlikely that some drugs may affect one
tissue directly, but impinge on another tissue indirectly, i.e., through the other tissue. If a
close coupling between cartilage and bone turnover exists, future treatment of OA most likely
need to restore or rebalance cellular communication in the joint. Otherwise, if the bone or
cartilage does not affect one another in the pathogenesis of OA, the optimal drug for OA
may be a cocktail of tissue-specific drugs.
Future studies could involve the
disassembly of the femur head explants in
bone and cartilage and compare fractions to
the intact femur head. In this way we may
be able to delineate how the drugs affect
bone and cartilage individually as compared
to the entire explants. In this regard, it would be critical to include femur heads from knock-
out mice, which lack the functional role of a specific cell type and/or signalling system in
order to investigate cellular crosstalk between the two tissues.
In line with our findings in the ex vivo femur head model, it is critical to investigate whether
hypertrophic chondrocytes in this model share similarities to cells in human OA. If the
hypertrophic chondrocytes have the exact same phenotype, it will increase the usefulness of
the ex vivo model in regard to mimicking human OA cartilage. If we find that the cells are not
exactly alike, we can use the ex vivo model to investigate the hypertrophic chondrocytes we
observe during development.
The present clinical use of intra-articular GC injections in OA knees, and the damaging
effects we find on bone by GCs in our study, highlights the importance of finding a
combination of drugs or a new GC, which do not have the detrimental effects on bone.
Thus, further studies using the novel femur head model could screen for potential drugs with
no detrimental effect on bone or cartilage.
Lastly, it is very important to understand the proteolytic processes of pathogenic cartilage in
order to developed new treatments for OA, which target the exact proteases that exert the
damaging effects at a given time during the pathogenesis of OA. We have already suggested
that it may be favourable to target aggrecanases in early OA and MMPs in later stages of OA.
However, more investigation on this topic is needed, especially in regard to investigate other
proteases than aggrecanases and MMPs, e.g., how important is the cathepsins in OA?
103

Abbreviations
Abbreviations
342-G2 MMP-mediated 342 FFGV-G2 fragment from aggrecan
374-G2 Aggrecanase-mediated 374 ARGS-G2 fragment from aggrecan
ADAMTS Aggrecanase/A disintergrin and metalloproteinase with thrombospondin motifs
AGNx-II MMP-mediated 342 FFGV-G2 fragment from aggrecan
ALP Alkaline phosphatase
BMP Bone morphogenetic proteins
CC Chondrocyte
COMP Cartilage oligomeric matrix protein
CIIM Collagen type II
CTX-I C-telopeptide of type I collagen
CTX-II C-telopeptide of type II collagen
DEX Dexamethasone
DMOADs Disease modifying osteoarthritis drugs
ECM Extra-cellular matrix
ELISA Enzyme-linked Immuno Sorbent Assay
E64 Irreversible inhibitor of cysteine proteases
FDA The US Food and Drug Administration
G1, G2, G3 Globular domain-1, -2, and -3
GAG Glycosaminoglycan, a proteoglycan
GC Glucocorticoid
GH Growth hormone
GM6001 General MMP inhibitor
GR Glucocorticoid receptor
HA Hyaluronic acid
IGF Insulin-like growth factor-1
IHC Immunohistochemistry
IL Interleukin
mAb Monoclonal antibody
M-CSF Macrophage - colony stimulating factor
MMP Matrix metalloproteinase
MRI Magnetic resonance imaging
NSAIDs Non-steroidal anti-inflammatory drugs
OA Osteoarthritis
OB Osteoblast
OC Osteoclast
ON Overnight
OPG Osteoprotegerin
OSM Oncostatin M
OVX Ovariectomy
PBS Phosphate buffered saline
PICP C-terminal propeptide of type I procollagen
PIIANP N-terminal propeptide of type II procollagen, splice variant A
PIINP N-terminal propeptide of type II procollagen
PINP N-terminal propeptide of type I procollagen
PRED Prednisolone
PTH Parathyroid hormone
RA Rheumatoid arthritis
RANK Receptor activator of nuclear factor-ĸβ
RANKL Receptor activator of nuclear factor-ĸβ ligand
sGAG Sulfated glycosaminoglycans
SD Standard deviation
TGF-β Transforming growth factor β
TIMP Tissue inhibitor of metalloproteinase
TNF-α Tumour necrosis factor-α
TRAP Tartrate resistant acid phosphatase
104

Appendix I
Co-author declarations for paper I - IV
Appendix I
105

Appendix I
106

Appendix I
107

Appendix I
108

Appendix I
109

Appendix I
110

Appendix I
111

Appendix II
Publication list
4 first author papers
11 co-Author papers
1 Submitted (first author)
Induction of Increased cAMP Levels in Articular
Chondrocytes Blocks Matrix Metalloproteinase–
Mediated Cartilage Degradation, but Not
Aggrecanase-Mediated Cartilage Degradation
Morten Asser Karsdal, Eren Ufuk Sumer, Helle Wulf,
Suzi H. Madsen, Claus Christiansen, Amanda J.
Fosang, and Bodil-Cecilie Sondergaard.
ARTHRITIS & RHEUMATISM, Vol. 56, No. 5,
May 2007, pp 1549–1558.
Patients with Rheumatoid Arthritis have an
altered circulatory Aggrecan Profile
Jean Charles Rousseau, Eren U Sumer, Gert Hein,
Bodil C Sondergaard, Suzi H. Madsen, Christian
Pedersen, Thomas Neumann, Andreas Mueller, Per
Qvist, Pierre Delmas and Morten A Karsdal.
BMC Musculoskeletal Disorders, 2008 May 28;9:74.
Cartilage Degradation is Fully Reversible in the
Presence of Aggrecanase but not Matrix
Metalloproteinase Activity
Morten Karsdal, Suzi Madsen, Eren Sumer, Claus
Christensen, Kim Henriksen, Amanda Fosang, Bodil
Sondergaard.
Arthritis Res Ther. 2008;10(3):R63. Epub 2008 May 30
Cartilage formation measured by a novel PIINP
assay suggests that IGF-I does not stimulate but
maintains cartilage formation ex vivo
Suzi H. Madsen, Sondergaard BC, Jensen AC,
Morten A Karsdal.
Scand J Rheumatol. 2009 May-Jun;38(3):222-6
Appendix II
112
Biochemical markers and the FDA Critical Path:
how biomarkers may contribute to the
understanding of pathophysiology and provide
unique and necessary tools for drug
development.
Karsdal MA, Henriksen K, Leeming DJ, Mitchell P,
Duffin K, Barascuk N, Klickstein L, Aggarwal P,
Nemirovskiy O, Byrjalsen I, Qvist P, Bay-Jensen AC,
Dam EB, Madsen SH, Christiansen C.
Biomarkers. 2009 May;14(3):181-202. Review.
Is bone quality associated with collagen age?
Leeming DJ, Henriksen K, Byrjalsen I, Qvist P,
Madsen SH, Garnero P, Karsdal MA.
Osteoporos Int. 2009 Sep;20(9):1461-70.
MAPKs are essential upstream signaling
pathways in proteolytic cartilage degradation -
divergence in pathways leading to aggrecanase
and MMP-mediated articular cartilage
degradation.
Sondergaard BC, Schultz N, Madsen SH, Bay-
Jensen AC, Kassem M, Karsdal MA.
Osteoarthritis Cartilage. 2009 Nov 11. [Epub ahead of
print]
Suppression of MMP activity in bovine cartilage
explants cultures has little if any effect on the
release of aggrecanase-derived aggrecan
fragments.
Wang B, Chen P, Jensen AC, Karsdal MA, Madsen
SH, Sondergaard BC, Zheng Q, Qvist P.
BMC Res Notes. 2009 Dec 18;2:259.

Biochemical Markers of Joint Tissue Turnover.
Bay-Jensen AC, Sondergaard BC, Christiansen C,
Karsdal MA, Madsen SH, Qvist P.
Assay Drug Dev Technol. 2010 Feb;8(1):118-24
Application of biochemical markers in
development of drugs for treatment of
osteoarthritis.
Qvist P, Christiansen C, Karsdal MA, Madsen SH,
Sondergaard BC, Bay-Jensen AC.
Biomarkers. 2010 Feb.;15(1):1-19.
Which elements are involved in reversible and
irreversible cartilage degradation in
osteoarthritis?
Bay-Jensen AC, Hoegh-Madsen S, Dam E,
Henriksen K, Sondergaard BC, Pastoureau P, Qvist
P, Karsdal MA..
Rheumatol Int. 2010 Feb;30(4):435-42.
Aggrecanase- and matrix metalloproteinase-
mediated aggrecan degradation is associated
with different molecular characteristics of
aggrecan and separated in time ex vivo.
Madsen SH, Sumer EU, Bay-Jensen AC,
Sondergaard BC, Qvist P, Karsdal MA.
Biomarkers. 2010 May 15(3):266-76
Investigation of the direct effects of salmon
calcitonin on human osteoarthritic chondrocytes.
Sondergaard BC, Madsen SH, Segovia-Silvestre T,
Paulsen SJ, Christiansen T, Pedersen C, Bay-Jensen
AC, Karsdal MA.
BMC Musculoskelet Disord. 2010 Apr. 5;11:62.
Characterization of an ex vivo femoral head
model using bone and cartilage turnover.
Suzi Hoegh Madsen, Anne Sofie Goettrup, Gedske
Thomsen, Søren Tvorup Christensen, Nikolaj
Schultz, Kim Henriksen, Anne-Christine Bay-Jensen,
Morten Asser Karsdal.
Cartilage. 2011 July; vol. 2, 3: pp. 265-278, first
published on October 2010.
Appendix II
113
Glucocorticoids exert context-dependent effects
on cells of the joint in vitro.
Suzi Hoegh Madsen, Kim Vietz Andreassen, Søren
Tvorup Christensen, Morten Asser Karsdal, Francis
Michael Sverdrup, Anne-Christine Bay-Jensen, Kim
Henriksen.
Steroids. Acceptet, July 2011
Metalloproteinase and cysteine protease profiles
in normal and OA cartilage.
Suzi Hoegh Madsen, Kim Henriksen, Søren
Tvorup Christensen, Morten Asser Karsdal, Anne-
Christine Bay-Jensen.
Proteome Science. Submitted, September 2011
Acknowledged for creating figures for papers
Matrix metalloproteinase and aggrecanase
generated aggrecan fragments: implications for
the diagnostics and therapeutics of destructive
joint diseases.
Sumer, E. U., Qvist, P. and Tankó, L. B.
Drug Development Research, 2007, 68: 1–13.
Post-translational modifications of the
extracellular matrix are key events in cancer
progression: Opportunities for biochemical
marker development.
D. J. Leeming, A. C. Bay-Jensen, E. Vassiliadis, M. R.
Larsen, K. Henriksen, and M.A. Karsdal.
Biomarkers, 2011; 16(3): 193–205.