INFORMATION TO USERS This manuscript has been reproduceâ from the microfilm master. UMI films oie text directiy from the original or copy submitkd. Thus, some thesis and dissertation copies are in t ypew i i t e r face, while others may be from any type of cornputer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Ovenize materials (e.g., maps, drawings, &arts) are reproduced by secüoning the original, beginning at the upper Ieft-hand corner and wntinuing from left to nght in equal sections with small overlaps. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographie prints are available for any photographs or illustrations appeating in this copy for an additional charge. Contact UMI diredly to order. Bell & Howell Information and Leaming 300 North Zeeb Road, Ann Arbor, Mi 481061348 USA 800-521 -06 00
176
Embed
INFORMATION TO USERS - Library and Archives Canada DHAR JUGESSUR, &Sc. DEPARTMENT OF SURGERY DMSION OF SURGICAL RESEARCH McGILL UNIVERSITY MONTREAL, QUEBEC CANADA DECEMBER 1997 A ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
INFORMATION TO USERS
This manuscript has been reproduceâ from the microfilm master. UMI films
oie text directiy from the original or copy submitkd. Thus, some thesis and
dissertation copies are in typewiiter face, while others may be from any type of
cornputer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality illustrations
and photographs, print bleedthrough, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
Ovenize materials (e.g., maps, drawings, &arts) are reproduced by
secüoning the original, beginning at the upper Ieft-hand corner and wntinuing
from left to nght in equal sections with small overlaps.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9" black and white
photographie prints are available for any photographs or illustrations appeating
in this copy for an additional charge. Contact UMI diredly to order.
Bell & Howell Information and Leaming 300 North Zeeb Road, Ann Arbor, Mi 481061348 USA
800-521 -06 00
THE CHONDROITIN SULPHATE EPITOPE 846 OF
AGGRECAN: ITS RELA'FIONSHIP TO AGGRECAN
SYNTHESIS AND IT.S PARTIAL CaARACTERIZATTON
BY
HTESHINI DHAR JUGESSUR, &Sc.
DEPARTMENT OF SURGERY
DMSION OF SURGICAL RESEARCH
McGILL UNIVERSITY
MONTREAL, QUEBEC
CANADA
DECEMBER 1997
A THESIS SUBMITTED TO THE FACUL'IY OF GRADUATE
STUDIES IN PARTIAL FULFILMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY
Q HITESHINI DHAR JUGESSUR
National Libmry B * I of Canada Bibliothèque nationale du Canada
Acquisitions and Acquisitions et Bibliographie Services services bibliographiques
395 Wellington Street 395, rue Wellington Ottawa ON KI A ON4 Ottawa ON K1A ON4 Canada Canada
The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant a la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sell reproduire, prêter, distribuer ou copies of this thesis in microfom, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/film, de
reproduction sur papier ou sur format électronique.
The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels may be printed or othewise de celle-ci ne doivent être imprimes reproduced without the author's ou autrement reproduits sans son permission. autorisation.
TABLE OF C O r n N T S
ABSTRACT........ .................................................................................... i
a.. R É s m ..................................... .o...o.......o.*...o....*......o..œ.....o......oui
.................................................................................................. DEDICAWN v
............................................................................ ACKNOWLEDGEMENTS vi
... ............................................................................. LIST OF TABLES ....oœ.....vlii
LIST OF FIGURES . . ~ ~ ~ . m . a a m ~ m . a . o a a ~ ~ ~ ~ ~ m o * ~ a ~ m ~ ~ ~ ~ o o a ~ a o o a a o ix
0.. ....................................................................... LIST OF ABBREVIATIONS xiil
l a I N T R O D U C T I O N . . . . . ~ a a ~ a a ~ ~ a o a a ~ . a ~ . o . o a o o 1
(A.R. Poole, 1993) Figure 2. Structure of Aggrecan Aggregate.
even smaller in aggrecan isolated 5om mature cartilage. It has been suggested that
this is the result of extracellular processing of this molecule.
The human aggrecan gene is composed of 19 excns ranging in size fkom 77 to 4224
bp with exon 1 being non-coding. The G1 domain has a characteristic three-looped
structure. This includes an N-terminal A loop, which shows amino acid sequence
similarity to the irnmunoglobulia superfandy, and two further loops B and B' which
are homologous to each other and are tamed proteoglycan tandem repeats (PTRs)
(Perkins et al., 1989). Link protein also contains these motifs, thus replicating the
structure of the G1 domain (Pakins et al., 199 1). The A loop of G1 is encoded by
exon 3, the B loop by exons 4 and 5, and the B' loop by exon 6 ('Vaihmu et al., 1995).
The G 1 domain of aggrecan contains sequences that can act as B and T-ceU epitopes
and are responsible for the antigmicity of this macromolede in diseases like
rheumatoid arthntis (Leroux et al., 1996). The G2 globular domain contains a similar
proteoglycan tandem repeat sequence but no immunoglobulin-like domain. This
domain has no known bction. The B loop of G2 is encoded by exons 8 and 9 and
the B' loop by exon 10 (Valhmu et ai., 1995).
The interglobular domain that links G1 and G2 is encoded solely by exon 7 and
provides numaous sites for the proteolytic attack of aggrecan during extracellulin
matrix m o v e r in normal physiology and in pathology. Proteolysis of aggrecan is
mediated b y eflzymes such as matrix metalloproteinases (MMPs), aggrecanase,
cathepsins, el astase and many others (Hardingham et al., 1994a). Cleavage within the
interglobdar domain results in the loss of the GAG bearing part of the molede
while G1 remains complexed with HA in the cartilage ma&.
AU of the CS and most of the KS chains are substituted onto the core protein between
the G2 and G3 globular domains. Most of these KS chahs are closer to the K
terminus of the protein core, with a KS-rich region being present on the C-terminal
side of the G2 globular domain. A few KS chains are also found in the GLG2
intergiobular domain (Barry et al., 1995). The KS-rich region of the human aggrecan
gene is encoded by exon 1 1 and the 5' end of exon 12. nie exon 12-encoded portion
of the KS-rich region contains 11 consecutive hexapeptide repeats of EEP(S,F)PS
(Doege et a!., 1991). Exon 12 also codes for the complete CS-attachent region
consisting of the CS 1 and the CS2 regions, the latter being situated in the C-terminal
half of the core protein. Over 100 CS chains are attached to specific serine residues in
the CS attachment region. The concaisus attachment signal for CS requires these
serine residues to occur adjacent to glycine residues; other re@anents for CS
attachent may include a nearby acidic amino acid as weîi as a nonpolar residue.
Human aggrecan contains a conserved sequence of 19 repeats of 19 amino acids
within the CS 1 region, where the majority of the CS chains are concentrated.
However it has recently bem reported that w i t h a single population, h u m a exhibit
genetic polymorphism in the number of ser-gly sequences and hence in the length of
the core protein that they express. This individual allelic variation in aggrecan repeat
length wouid a f k t the numba of CS chains substimted on the core protein (Doege et
al., 1997). This may have implications for the functionality of the cartilage of
different individuals.
The G3 region of human aggrecan consists of two alternatively spliced epidamal
growth factor (EGF)-like domains encoded by exons 13 and 14, a lectin (LEC)-like
domai. aicoded by exons 15, 16 and 17 and an dtematively spliced complement-
regulatory protein (CRP)-like domain encoded by exon 1 8 (Valhrnu et al., 1995). 0 3
appears to be involved in the intraceIlular aafncking of aggrecan, based on the
studies of the mcated aggrecan precursor produced by nanomelic chicks (Vatel et
al., 1993; Rimorac et al., 1994). It has also bem demonstrated that this domain has
the ability to bind to the chondrocyte cell surface. The two EGF-like domains may be
involved in Ca2+ binding and in regulating cell metabolism, although thae is little
evidence for this to &te. The LEC-like domain on this molecule has the capacity to
bind to sugars such as fucose and galactose (Halberg et al., 1 988) and bears sequence
sidarity to proteins such as the leucocyte adhesion molecde- 1 (LAM- 1), the human
macrophage mannose receptor and the chicken hepatic Iectin. This domain may
therefore interact with other rnatrix components such as the galactose residues on
type II collagen.
Exon 19 codes for the 25-amino acid C-terminus of the core protein and the 3' UTR
of the aggrecan gene ( Valhrnu et al., 1995).
1 S,2. BIOSYNTHESIS OF AGGRECAN
Model systans of cultureci chondrocytes from the Swarm rat chondrosarcoma have
been used to study the biosynthesis of aggrecan. In this system, proteoglycan
synthesis can be described as occuring in two stages. Firstly, a large pool of core
protein precursor with a long haE4Se (- 90 mins) undergoes translation and K
glycosylation in the rough E R This is followed by a rapid modincation in the Golgi
complex in which GAG chains and O-linked oligosaccharides are added to the core
protein with subsequent rapid secretion into the extracellular ma& (half-Iife in the
cell is about 7-8 rnins) (Khura et a[. , 1 9 84; Hascall and Kimura, 1 98 1 ; Kimura et al.,
198 1).
During translation, the newly forming aggrecan core protein is translocated across the
endoplasmic reticulum (ER) by a signal peptide-mediated process. Withh the rough
ER lumen, N-asparagine-Wed hi&-mannose oligosaccharides are added CO-
translationaîly to the core protein fiom dolichol phosphate intemediates. Most of
these oligosaccharides are in or near the G1 and G2 domains. Disulphide bond
formation and correct folding of the globular domains also occurs in the ER, most
likely assisted by chaperone proteins. Trimming of the N-iinked oligosaccharides also
begins here. Xylosylation of the core protein begins in the late ER and possibly
continues to the early cis Golgi cornpartment (Kearns et al., 1993; Nuwayhid et al.,
1986; Lohmander et aL, 1989). Xylosylation occurs by the addition of xylose f?om
uridine diphosphate-xylose (UDP-xylose) onto the serines in the ser-gly sequences
and is an essential step in the initiation of CS ch& synthesis.
The late ER cornpartment is the site where defective aggrecan precursors are arrested
fiom exiting the ER and targetted for degradation. When the xylosylated core protein
amives in the Golgi, N-linked oligosaccharides are further processed, O-linked
oligosaccharides are added and KS and CS chahs are formed. Nucleotide sugar
precursors are required for the synthesis of both oligosaccharides and
glycosaminogiycans, and phosphoadaiosine phosphosulphate (PAPS), the high
energy sulphate donor, is required for sulphation. These are produced in the cytosol
and trrmsported into the Golgi cistanae by an antipon exchange-mechanism. 0-
linked oligosaccharide addition is initiated by the transfer of G a A c to Ser or Thr
residues and may be a medial- or lataGolgi event occuring concomitantly to CS
chain synthesis (Lohmander and Kimura, 1986).
Multienryme complexes within the Golgi membranes are involved in the processes of
chain elongation and sulphation. These glycosyltransferases and sulphotransferases
act CO-ordinately to achieve GAG synthesis. In the case of KS formation, Gd-
GlcNAc repeats are added on the terminal Gd of O-linked oligosaccharides. This is a
lateGolgi event (Lohmander and Kimwa, 1986) since O-linked oligosaccharides are
synthesised in the medial-tram Golgi compartments. Furthmore, there is evidence
for a lack of KS substitution on aggrecan molecules treated with brefeldin A (BFA), a
compound which dismpts the translocation of the core protein fiom the ER to the
trans-Golgi network (Wong-Palrns and Plaas, 1 99 5).
CS synthesis has been studied more extensively. The assembly of GlcA-GalNAc
repeats occurs very rapidly and has been shown to begin in the medial-tram Golgi
(Siibert and Sugumaran, 1995). This view is supported by the observation that whm
treated with BFA, chondrocytes h m nanomelic chickens synthesise a tmcated
aggrecan core protein that is substituted with CS and that remains in the ERfGolgi
campartment (Verte1 et ai., 1994). Moreover; BFA does not cornpletely inhibit CS
chah elongation (Sugumaran et al., 1992). Howeva, other studies using BFA favour
the Tram Golgi Network as the site of these events in bovine chondrocyte cultures
(Wong-Palms and Plaas, 1995). Prior to CS chah polymerisation, a M a g e region
consisting of Gal-Gd-GlcA is added sequentially to the xylosylated core protein.
Specific glycosyltransferases are responsible for the addition of each sugar in the cis
and medial compartments of the Golgi (Silbert and Sugumaran, 1995). Tt has been
reported that a large proportion of xylose residues in proteoglycans from the Swarm
rat chondrosarcoma contain are phosphorylated in their 2-position, however their
function is unknown (Oegema et al., 1984).
The newly secreted aggrecan molecules are unable to interact with hyaluronic acid
immediately, but are able to form stable aggregates together with lhk protein with
tirne in the extracellular ma&. It has been suggested that this phenornenon is
sensitive to the extracdufar environment (Melching and Roughley, 1990; Sandy et
ai., 1989) and that certain conformational changes in the hyaluronic acid-binding
regions may faaltate this process (Oegema, 1980; Bayliss et al., 1984).
1.5.3. AGGRECAN CATABOLISM
Aggrecan degradation is initiated in the ma& by proteolytic enzymes produced by
the chondrocytes thernselves. These include lysosomal enzymes such as cathepsin B
and various metalloproteinases (Fosang et al., 1992). The action of these enymes in
normal tissue hirnover gives rise to GAG-peptides which lose theu ability to interact
with hyaluronic acid and coasequently diffuse into the surrounding joint fluid, korn
where they enter the circulation, are further degraded in the liver, and are finally
eliminated via the kidneys into the urine. In pathology, howeva, these GAG-peptides
are taken up by phagocytic ce& and degraded in their lysosomes by various
proteinases, glycosidases and sulphatases. Meatlwhile, in the matrix, thae is an
accumulation of link protein as weli as of G1, resuiting fiom cleavage withh the G1-
G2 intaglobuiar domain. niese rem& attached to hyaluronan.
The aggrecan core protein can undergo proteolytic cleavage at several sites d o n g its
length, but the G142 interglobular domain is a prefared site of action of several
enzymes. The A S ~ " ~ - P ~ ~ P ' and peptide bonds are the cleavage sites of
metalloproteinases and the enzyme activity associated with ' aggrecanase' ,
respectively (Sandy et al., 1991b; Ilic et al., 1992). C-terminal fragments with the
Glu"=Ala3" cleavage site have been identifïed in the synovial fluids of arthritic
patients (Lohmander et al., 1993) and in the media from cartilage explant and
chondrocyte cultures (Sandy et al., 1991b; Ilic et al., 1992) stirnulated with
interleukin-1 or retinoic acid. However, the protehase mediahg cleavage at the
aggrecanase site has not been characterised to date. In vitro studies using a p d e d
G1-G2 subtrate danonstrate that MMP-8 can cleave at the Gh373-Ala374 aggrecanase
site (Fosang et al., 1994 ). However, it has also been demonstrated that this enzyme
will preferentidy cleave aggrecan at the metalloproteinase site. Furthennore, the
inhibition of MMP-8 using a potent inhibitor does not eliminate the detection of
degradation products generated by endogenous aggrecanase (Amer et al., 1997).
Ma& metalloproteinases MMP- 1 ,-2,-3,-7,-8,-9, and - 13 have been shown to cleave
at the A ~ n ~ ~ l - P h s ~ ' site in vitro (Fosang et al., 1996; 1994; 1993) Monoclonal
antibodies have been prepared that are specific for the neoepitopes generated by
proteolytic cleavage within the GLG2 interglobular domain. These have been useful
in the identification and quantitation of aggrecan catabolic products both Ni vitro and
in situ (Hughes et al., 1995).
The precise sites of aggrecan cleavage within the CS attachment regions are not W y
understood Antisera have been used to detect core protein species generated by
aggrecanase within the CS attachment regions of human aggrecan obtained fiom
dcu1a.r cartilages of diffkrent ages. These studies indicate that human aggrecan is
cleaved by aggrecanase in vivo at the Glu1 714-Gly 17 15 site, and that this cleavage
occurs in epiphysial cartilage during foetal development (Sandy et al., 1996).
1.5.4. CHANGES IN AGGRECAN S T R U a DURING NORMAL
DEVELOPMENT AND AGING
The composition of the extracellular matrix changes during foetal development as
well as after birth, during juvenile development to the adult and thereaftei: throughout
Core Protein Structure Glycosylation and GAG fine structure
' & size -1 size, number of CS ? IGD cleavage f size, number of KS & G3 content 6 O-linked oligosaccharides L Çer, Gly f total sulphation T Arg , Tyr f 6-sulphation of CS
& 6sulphation of CS -1 non-reducing terminal GalNAc4MlcA-of CS f non-reducing terminal GalNAc4,6S-GlcA- of CS
Table 2. Changes in Aggrecan Structure Durhg Normal Developrnent and Aging.
life. Proteoglycans, in particular, undergo various modincations in their size and
charge (Table 2) and these changes may subsequently alter the functional properties
of the cartilage. Adult hyaline cartilage therefore consists of a heterogeneous mixture
of aggrecan molecdes gaierated by age-related changes in biosynthetic and catabolic
events.
a') Chan~es in core rotei in structure
With increasing age up to the end of growth, there is a decrease in the aggrecan
content of cartilage which remains relatively constant thereafter. The hydrodynarnic
size of the proteoglycan subunit deaeases, whilst the protein to GAG content
increases. As a result of proteolytic cleavage within the O1 -G2 interglobular domain,
there is an increase in abundance of G 1 -containhg fragments in the cartilage, as these
cimnot diffuse fiom the tissue because of their interaction with hyaluronic acid
(Roughley et al., 1984; 1985).
The amino acid composition of the core protein also changes, with the serine and
&cine content decreasing whilst the argùline and tyrosine content inneases
(Roughley and White, 1980). This variability in amino acid composition may be
interpreted as being the resdt of proteolytic cleavage of the aggrecan core protein,
resulting in heterogaieous sues of the GAG attachment region (Heinegard, 1977). It
has also been proposed that different aggrecan populations bearing distinct core
proteins exist in cartilage mopwood and Robinson, 1975; Stanescu et al., 1977).
hdeed, using immunological and biochemical rnethods, two distinct populations have
beea identified (Champion et al., 1982; Bayliss et al., 1983; Heinegard et al., 1985).
However, since only a single gene for aggrecan has been identified, the existence of
these different populations has been atûibuted to differmces in post-translational
processing of the aggrecan.
During human foetal development, the content of the high-daisity foetal type PG
increases and drops after birth and with inaeasing age until it is barely detectable in
the adult. The core protein shows Little or no change in its composition during the
foetal period but changes considerably after birth. Analysis of human cartilage of
different ages using monoclonal antibodies to epitopes on the core protein and CS
chahs of aggrecan revealed the existence of a foetal type and an adult type of
proteoglycan molecules, which were thought to be products of different genes (Glant
et al., 1986). The foetal type proteoglycan decreased in content in cartilage with
increasing foetal age af'ter about 27 weeks until it was barely detectable in adult
cartilage by about 30 years of age. As the content of this population diminished, a
second population appeared a fk r birth, the content of which increased with
development and was maximal in adult cartilage. However, it was subsequently
demonstrated that these changes were due partly to differences in glycosylation
patterns of the core protein and partly to the outcome of proteolytic modification of
the initially produced core protein (Roughley et al., 198 7).
The content of the C-termitlai region of aggrecan containhg the G3 domain also
dbhhhes with aging as a result of proteolytic cleavage of thc core protein fkom the
C-terminus. The polyclonal antibody JD5 raised against a recombinant f o m of G3
has been used in immunoassays to qmtitate the amount of G3 relative to G1 in
cartilage extracts fkom newboms to individu& of up to 65 years of age. The results
show a 92% drop in the content of G3 with aging (Dudhia et al., 1996). The decrease
is smalî up to about 30 years of age but ttiereafter is very rapid This observation is
supported by the technique of rotary shadowing electron microscopy which has
previously shown that the G3 globdar domain is present on most aggrecan molecules
prepared fiom cartilage obtained f?om young subjects, but is often misshg fiom
aggrecan prepared from adult cartilage (Paulsson et al., 1987).
b) Chan~es in elvcosvlation and GAG fine structure
In humans, the abundance of CS and KS does not change before birth. During the
period of skeletal development, however, there is a decrease in the size and the
numba of CS chahs and an inaease in the size and number of KS chains (Bayliss
and Ali, 1978; Elliot and Gardner, 1979; Roughley and White, 1980). The number of
O-Wed oligosaccharides also decreases (Roughley et al., 198 1; Santer et al., 1982).
Such changes have also been reported in various animals (Sweet et al., 1979; Garg
and Swann, 198 1).
In the human foetus, there is an increase in the degree and position of sulphation of
the CS chahs. CS c h a h from the eariy foetal period contain 25% nonsulphated
disaccharides. By 19 weeks of gestation, 85% of the disaccharides are sdphated and
by binh 90-95% of disaccharides are sulphated. During this period the amount of 4-
sulphation increases mtil it reaches the level of 6-sulphation, which remains constant.
In contrat, after biah, the proportion of 6-sulphation inmeases whilst Csulphation
declines until it is about 5% in the adult (Rougidey et al., 1987). 0th- investigators
have also obsaved such a decrease in 4-sulphated disaccharides and increase in 6-
sulphated disaccharides with aging (Bayliss et al., 1995).
Disaccharide aiialysis of CS chains firom human aggrecan reveals that the ratio of
unsulphated disaccharides to 4-sulphated and 6-sulphated disaccharides is about
10:40:50 in the foetus, but changes to 0 5 9 5 in nomal adult tissue (Plaas et al.,
1995). In addition to changes in sulphation withui the CS chains, there are differences
in the non-reducing terminal residues of CS chains during cartilage maturation (Table
3). Approximately 75% of CS chains on foetal aggrecan terminate in GalNAc4S-
GlcA-, 5% in GaNAc4,6S-GlcA-, 8% in GlcA-GaINAc4S-GlcA- and 12% in GlcA-
GalNAc6S-GlcA-. The proportions are 47%, 49%, 1% and 3%, respectively, on adult
CS chains. Thus the relative number of CS chains terminating in GlcA is much higha
on foetal aggrecan (Hascd et al., 1995). In support of these observations, data
obtained from cultures of Swarm rat chondrosarcoma ceUs show that the majority of
aggrecan CS chahs terminate in GmAc4S or GaWAc4,6S and that there is a 60-
fold greater incidence of GalNAc4,6,S at the non-reducing end position as compared
to internai positions (Midura et al., 1995). Moreover, previous work on
biosynthetically labelled aggrecan CS fkom chick and rat chondrocytes shows the
presence of GaINAc4S and GdNAc4,6S as terminal components (Otsu et al., 1985).
Table 3. Chain Terminations of Chondroitin Sulphnte.
With increasing age, human aggrecan shows an increase in KS relative to CS
(Roughley and White, 1980) and the abundance of smaller, KS-nch molecules
increases (Webber et al., 1987). These age-related changes in KS have been further
studied in the G1 domain of mature and immature bovine aggrecan and show that in
the steer KS is attached to Thf2 in loop A of the G1 via an O-linkage. This is not the
case in the calf G1. KS is found N-linked to Asn= in the B loop of the 0 1 in both the
calf and the steer, but the chains are shorter in the calf. The B' loop of mature G1 is
substituted with an N-iînked KS at A d L 4 or with a complex type oligosaccharide,
while this site does not carry KS in the calf. In addition, the two Thr within the
sequence TIQWT located in the GLG2 intaglobular domain are substituted with KS
in the caIf and the steer, but in the latter further substitution occurs within the
NITEGEA sequence which contains the aggrecanase cleavage site (Barry et al.,
1995). Thus these features demonstrate further the reasons for polydispersity of
aggrecan fiom mature cartilage.
The region near the CS-protein linkage in aggrecan also contains various types of
substitutions which add to the complexity of these macromolecules (Shibata et al.,
1992; Cheng et ai., 1996). In humans, age-related changes in sulphation within these
linkages have been identified (Cheng et al., 1996). In CS fiom young human
cartilage, sulphate groups are mostly at the Cposition of GalNAc in the major part of
the ch&, but at the 6-position in the non-reduchg distal end. In CS fiom old
cardage, however, sulphation at the 6-position of GaWAc is predominant &om the
non-reducing end d o m to approximately positions 4 and 5 fiom the linkage region,
where GalNAc4S is cornmon. It has been proposed that these variations may encode
information which would detemiine the type of substitutions which would occur on
the remaining GAG c h a b however this remains to be proven.
1.5.5. CHANGES IN AGGRECAN STRUCTURE IN JOINT PATHOLOGY
a) Chan~es in core arotein structure
In osteoarthntis (OA), a . imbalance between metalloproteinase levels and their
inhibitors results in accelerated degradation of cartilage aggrecan molecules. Similar
changes in aggrecan size occur to those observed during aging, however the products
of degradation in early degeneration are smalla than those found in age- and site-
matched healthy addt cartilages. This early degenerative phase is followed by a
reparative phase during which new aggrecan molecules are synthesised having
properties similar to foetal type aggrecan in their larger size and can also aggregate
with HA (Rizkalla et al., 1992). These are subsequently lost fkom the matru< as the
disease progresses. Analysis of synovial fiuids f?om patients with recent knee injury,
early or late stage OA and fiom patients with Mammatory joint conditions has
demonstrated the presence of cleavage products generated by aggrecanase
(Lohmander et al., 1993; Sandy et al., 1992), but other proteases such as MMPs are
also involved in aggrecan catabolism in pathology (Flannery et al., 1992).
b) Chmees in GAG structure
The sulphation pattern of CS c h a h changes in joint diseases. These changes produce
aggrecan that bears resemblance to the foetal type moledes in their sulphation
pattern and in their non-reducing terminal components. Disacchande compositional
analysis of human aggrecan shows the ratio of unsdphated to 4-sulphated to 6-
sulphated disaccharides to be 0:5:95 in normal adult sarnples but 5:15:80 in OA
samples. Hence an increase in unsuiphated and 4-sdphated disaccharides is observed.
The non-reducing terminal structures in OA show an elevated proportion of GlcA-
GalNAc4S and a reduced level of GalNAc4,6S (Plaas et al., 1995).
Similar changes have been reported in racehorses with osteochondral fractures that
have led to cartilage degeneration (DJD). In DJD horses, there is an increase in 4-
sulphated disaccharides compared to normal horses. This is accompanied by an
increase in the proportion of non-sulphated CS chains kom 2.7% in normals to 9% in
DJD cartilage. In normal cartilage, 25% of the c h a h t e d a t e with GalNAc4S, and
75% with GalNAc4,6S. In foetal and DJD cartilage, the CS chains terminate
predominantly (90%) with GaNAc4S (Brown et al., 1996).
1 5 6 . MONOCLONAL ANTIBODIES TO SPECIFIC CS EPITOPES
The above described variation in sulphation patterns dong the CS chaius has enabled
the development of several monoclonal antibodies (MAbs) which c m recognise
distinct epitopes on the GAG chah. These antibodies can be aivided into two
general categories: a) those that require predigestion of the CS with endo- or exo-
glycosidases to generate their epitopes and b) those that recognise epitopes occuring
in the 'native' CS chains. Some of these antibodies have been used in the
irnmunolocalisation of CS epitopes in tissue sections and some have been suggested
as having potentiaî uses as markers for changes occuring during tissue development
or pathology (Catason et al., 1990a).
a) Monoclonal antibodies a~ainst endo- or exo-elycosidase penerated CS
epitones
MAbs in this group have been generated by immunising mice with chondroitinase
ABC-treated aggrecan. MAbs that recognise epitopes generated by complete
chondroitinase digestion of the CS, al l recognise a 4,s- unsahirated hexuronic acid,
unsaturated GlcA, at the non-reducing end of the residual disaccharide 'stub' of the
digested CS chain, which remains bound to the Mage oligosaccharide attached to
the core protein. This product results &om the 'efiminase' action of the
chondroithase (Hascd er al., 1995; Caterson et al., 1985) and is v a y antigenic.
Beiow, are some examples of MAbs with different specificities for different
sulphoesta substitutions on the adjacent GalNAc residue (Cataon et a[., 1990a).
These are siimmarised in Table 4.
MAb 2B6, 9A2 and 3D5 ail recognise epitopes containing a Csulphated G m A c
adjacent to a non-reducing terminal GlcA. 2B6 and 9A2, however, require the non-
reducing texminal GlcA to be a 4,s-unsaturatecl GlcA, whereas 3D5 also recognises
saturated GlcA at the non-reducing teminus (generated by the 'hydrolase' activity of
mammalian hyaluronidase). MAI, 3B3 recognises a 6-sulphated GalNAc adjacent to
eitha a saturated or unsaturated GlcA at the non-reducing tenninus, whereas MAb
1B5 recognises an unsulphated GalNAc next to a non-reducing t&al unsaturated
GlcA.
'Native' or Monoclonal antibody Epitope recognised chondroitinase generated
2.2.7, Tmmiinoprecipitation of cartilage extracts and media
In order to demonstrate that the 846 epitope was present on newly synthesised
aggrecan molecules, and to investigate what proportion of these navly synthesised
molecules contalied the epitope, cartilage extracts and media fiom the foetal bovine
explant cultures were immunoprecipitated. Magnetic immunoaffinity beads ( 1 . 2 ~ 1 Os,
Dynabeads M-450 Rat anti-Mouse IgM, Dynal, Lake Success, NY) were coupled to
antibody 846 ascitic fluid @repared at the Joint Diseases Laboratory, see 2.3.1 .)
diluted 1 : 100 in 10 ml of 0.2M triethanolamine, pH 9.0, contakiing 2ûmM dimethyl
pimelimidate dihydrochlonde (Sigma), with bidirectional mWng for 45 min, at room
temperature. The reaction was stopped by placing the tube containing the beads in a
magnet (Dynal MPC, Dynal) and discarding the supanatant. The beads were
resuspended in 10 ml of 0.2M triethanolamine, pH 9.0, and incubated for a M e r 2
hrs at room temperature. The beads w a e washed 3-4 Mies in PBS containing 0.1%
BSA. 3sS-suIphate labelled dialysed cartilage extracts or media (containing - 1 0000cpm) were added to the beads and the samples gently rocked at 4OC for 2 hrs.
The beads were eluted with 2M sodium iodide, pH 7.4 and the eluted material
counted in a scintillation counter. This method was tried on several occasions, using
varying amounts of beads and %sulphate labelled samples as well as with varying
incubation times, but was unsuccessfid at immunoprecipitating sufncient and
reproducible numbers of counts (cpm). It was therefore disconhued.
2.2,8, Determination of sxhhated GAG content
The total amount of sulphated GAG in the ceii layes and media was quâotitated
using the dimethylmethylene blue (DMMB) dye binding assay (Famdale et al., 1986)
using shark chondroitin sulphate (Sigma) as a stanâard Briefly, 10 pl of samples or
standards were added to 190pl of 0.1M DMMB (BDH Chemicals, Montreal, QC) in a
round bottom 96-well microtiter plate (Evergreen Scientific, Los Angeles, CA). The
absorbance was read within 5 mins at 525nm on a ELx808 microplate reader (Bio-
Tek Instruments Inc., Winooski, VT).
2-2.9. Determination of DNA content
A modified version of the method of LaBarca and Paigen (1980) was used to
determine the DNA content of the bovine chondrocyte cell layers. The cell layers
were digested with varying volumes of a stock concentration of proteinase K (2
mg/ml) to give a final amount of 0Smg protehase W 5 h g ce11 layer. The digestion
was c e e d out at 56OC for 24 hours in 0.1M sodium phosphate buffa, pH 6.5,
containhg 0.01% EDTA. Calf thymus DNA was used to prepare standards in the
range 2.5pg/ml-10pg/ml. In the assay, 2 d of a solution of 0.1pghnl bisbenPrnide
(Hoechst H 33246, Sigma) in 50mM sodium phosphate, 2M NaCl, pH 7.4, was added
to lOOpl of tissue digest or standard, and the mixture vortexed and placed in the dark
for 30 mins. Fluorescence was recorded with excitation at 356nm and emission at
45811x1, using a fluorescence spectrophotometer (mode1 650-IOS, Perkin-Elmer
Corporation, Norwalk, o.
2.3. Iodination of foetal human PG
Foetal human PG @TG) (AIDI, prepared by V. Vipparti, Joint Diseases
Laboratory, using a modifïed version of the method by Tang et al., 1979) was
iodinated using the chloramine T method of radioiodination (Sonada and
Schamowitz, 1970), in order to prepare '*I-HFPG for use as a cornpetitor in the
radioimmunoassay for the 846 epitope. HFPG was dissolved at 2mg/ml, ovdgh t , at
4"C, in fieshly prepared iodination buffa consisting of 50mM Tris-HC1, pH 7.5,
containing 150mM NaCl. A 10 ml pippette (Falcon) was packed with Sephadex G-25
(medium grade, P h a c i a , Uppsala, Sweden) in an elution buffer consisting of 0.1M
Tris-acetate containing 80mM sodium &de and 0.17M sodium acetate, pH 7.3. A
10% BSA solution was prepared in distilled water and 200p.l of it was applied to the
column. The column was washed with 20 ml of elution buffer. One hundred
microlitres of the HFPG solution was pippetted into a glas tube. Fresh solutions of
chloramine T (0.5 mm), sodium metabisulphite (1.2 mg/ml) and sodium iodide
(1 ûmg/ml) were prepared in the iodination buffer.
In the fume hood prepared for radioactive work, NalsI (supptied in dilute sodium
hydroxide solution at approximately 500mCi/ml, Amersham Life Sciences Inc.,
Oakville, ON) was diluted to 50 mCi/ml in 0.4M phosphate buffer, pH 7.6. Ten
microlitres of the SûmCi/ml solution of Na131 was added to the HFPG solution. The
tube was vortexed briefly, then lOpl of the chloramine T solution was added and the
mixture vortexed for 2 mins. One hundred microlitres of the sodium metabisulphite
solution was added into the tube to stop the reaction. The mixture was applied to the
G-25 column using a tram fa pippette. Two hundred microlitres of the sodium iodide
solution was added to the tube io which the iodination was canied out, and the tube
and tramfer pippette rinsed with this solution which was also applied to the column.
The column nin was started, eluting with the 0.1 M Tris-acetate elution buffer (see
above) and approximately 40 fractions w a e collected, with 15 drops (approximately
0.6 ml) per fraction. Ten microlitre aliquots fiom each hction were counted in a
gamma counter (Cobra Mode1 II, Packard, Meridan, CT). The void volume peak was
collected, pooling fiactions containing about 400,00Ocpm/10 pl. An equal volume of
radioimmunoassay (RIA) b d e r was added to the lSI-HFPG solution to stabilue the
preparation. The RIA buffer contained 0. lSM potassium dihydrogen phosphate
(BDH), 0.15M disodium hydrogen phosphate heptahydrate (BDH) at pH 8.1,
samples were pooled, dialysed overnight against distilled water, then lyophilised. The
sample was then redissolved at 2mg/ml in the above buffer for papain digestion.
Papain (Sigma) was added to all three samples at IOpg/mg PG, estimated by the
DMMB assay, and the samples were incubated at 37OC for 4 h after which papain
was added again and the samples left to digest ovemight. Papain was inhibited with
iodoacetadde (Sigma) added to a final concentration of 10mM.
2.6.3. Cetylpyridinium chloride KPCI mecioitation of ~ a ~ a i n digested samdes
CPC precipitation was performed in order to p e the single CS chains attached to
peptides generated by papain (Method adapted eorn Roughley and Bamett, 1977).
Briefiy, a 5% (w/v) CPC solution containing 0.25M MgC12 was added to the sampîes
to a f i a l conceneation of 1% CPC. The precipitate h e d was separateci fiom the
supernatant by centrifugation at 1500û1pm (15000g, Eppendorf Mode1 54 13 rotor) for
5 mins at 200C. It was then resuspended in 2ml of a 1% (w/v) solution of CPC
containing 0.05M MgClz and the washes repeated 3 times. The h a 1 precipitate was
dissolved in lm1 of propan-1-01 Iwater (3:2, vh), then re-precipitated ovemight by
the addition of Zml ethanol saturateci with potassium acetate. The samples were
centrifiiged as above, and the precipitate washed twice with 2 ml of ethanol. As mu&
of the ethanol was aspirated as possible and the remainder was left to evaporate at
room temperature ovemight.
2.6.4. Dot blots of CS chains Fom foetal bovine PG and addt human PG
CS chains prepared by CPC precipitation were dissolved ovemight in PBS, prior to
performing dot blots. The dot blots were performed to determine whether the 846
epitope could be detected if high concentrations of CS chains were applied to a
polyvinylidene difiuoride (PVDF) membrane. This method was also used to establish
whether or not addt human PG exhibited reactivity to the 846 antibody under these
conditions. PVDF membrane (Biorad) was soaked for 5 mins in methanol, then in
PBS, in order to hydrate it. The membrane was then carefidiy rested on a sheet of wet
filter paper, making sure that there were no air bubbles trapped in between the two
layers. Twenty microlitre drops of the Smg/ml CS solution were applied to the
membrane (making sure that each drop was M y absorbed into the membrane before
the next one was added), until a total of 500,250, 125 and 62 kg CS were applied
The membrane was blocked overnight at 4'C, with gentle rocking in PBS containing
3% BSA. The membrane was washed (2 x 10 min washes) in PBS containing 0.1%
Tween and 1% BSA (PBSA'BSA). Then antibody 846 ascitic fluid @repared at the
Joint Diseases Laboratory, see 2.3.1.), diiuted at 1:200 in PBS/T/BSA, was added and
the membrane rocked in the antibody solution at room temperature for 1 hour. The
antibody solution was removed, and the membrane washed twice as above, in
PBSiTIBSA. Goat anti-moue IgGAM (Zymed, San Francisco, CA) was diluted at
1 : 1000 in PBS/T/BSA, added to the membrane, and the membrane gently rocked for
1 hour at room temperature. Following t w o 10 minute washes in PBSRBSA, a
substrate solution of 5-bromo4chloro-3-indolyl phosphate (BCIP) and nitroblue
tetrazoleum (NBT) (Biorad) was added and the blot dowed to develop. A solution
containing a mixture of 1 0mgh.l foetal bovine PG and 1 :200 diluted 846 ascitic fluid
in PBS, which had been allowed to incubate ovemight at 4OC with gentle rocking,
was used in the control blot instead of the 846 antibody solution.
2.6.5. Sepharose CG6B chromatoaa~hy of CS chahs
In order to investigate the size distributions of the CS chains from foetal bovine PG
and OA cartilage PG, the CS c h a h prepared by papain digestion and CPC
precipitation were chromatographed on a CL-6B coIumn (100cm x 2.5cm). Five to
ten milligrams of GAG were loaded onto the colurnn in a total volume of no more
than 5 mls of 0.2M sodium acetate, pH 5.5. The column was eluted in the same bufkr
and 10 ml fiactions were collected at a flow rate of 20 mYhr. The Vo and V, of the
column were detennined using native foetal bovine PG monomer and potassium
dichromate, respectively. The hctions were dialysed overnight against distilled
water (membrane molecular weight cut off 3500), lyophilised thea reconstihited into
200pl0.2M sodium acetate, pH 5.5.
2.6.6. Dot blot of Saharose CL-6B fiactions
To determine whether or not the 846 epitope was present on CS c h a h of a specinc
la@, or whether its presence on CS was irrespective of chah laigth, fiactions
representing the differait parts of the elution peak of the CS chahs were applied to
PVDF membranes and bloaed with the 846 antibody, as described above. A total of
150pg CS (determined by DMMB assay) was applied from each fraction.
2-6.7. Treatment of foetal bovine PG with 8-ducuronidase
Foetal bovine PG was treated with various amounts of bovine iiver P-glucuronidase
(Sigma) in order to remove any non-reducing terminal glucuronic acid and to
determine whether this residue was a component of the epitope recognised by the 846
antibody. Briefly, a 10 mghi solution of foetal bovine PG was prepared in 0.1M
sodium acetate, pH 5.0. One milligram PG was digested with O, 1, 10 and lOOU of
cPyme in a total volume of I d , at 37OC for 4 hours. The enzyme was inactivated by
boiling for 5 rnins. Aliquots were taken fkom each treatment and assayed for 846
epitope content by radioimmunoassay. The ranaining samples were prepared for
unsaturated disaccharide analysis, analysis of non-reducing terminal components and
analysis of the PG hydrodynamic size in collaboration with Dr. Anna Plaas (Shriners
Hospital for Children, Tampa Unit, Tampa, FL) using previously established
technology (Plaas et al., 1996; Deutsch et al., 1995).
2.6.8. Remration of PG sarudes for disaccharide and non-reducine terminal sunar
analysis
Samples treated with p-glucufonidase were *han01 precipitated in order to separate
the PG fiom the products of digestion. A total of 3.2 ml of 100% ethano1 saturated
with potassium acetate, was added to 800 pl of sample (80% v/v &al concentration
of ethanol) and the mixture s h e d overnight at 4OC to precipitate the PG. The
samples were centrifuged at 15000 rpm (15000g, Eppendorf Mode1 5413 rotor) for 10
mins to separate the precipitate. This was foliowed with 2 x 10 min washes of the
precipitate in 100% ethanol, after which the supernatant was carefülly rernoved and
the precipitate dowed to air-dry overnight at room temperature.
3.1. Foetal bovine exvlant cultures
It has previously been show that the 846 epitope is abundant in foetal cartilage,
barely detectabie in nomal adult cartilage (Glaat et al., 1986), but is elevated in the
cartilage of OA patients (Rizkalla et al., 1992). Foetal bovine explant cultures were
established to investigate whether the 846 epitope was associated with the synthesis
of aggrecan molecules and not a product of the modification of the existïng aggrecan
in the matrix. To identie such a relationship, foetal cartilage was cultured since it is a
tissue that contains a high content of the 846 epitope, and it was thought that any
changes in the levels of the epitope resulting f?om new synthesis would be detectable
in such a system. Fig. 4a shows a progressive drop in the levels of newly synthesised
PGs C5S-sulphate labelled) in cartilage, with time in culture. These counts represent
(compared to media levels) 98.8% of the total amount of newly synthesised
molecules on &y 1, 97.6% on &y 6 and 97.8% on &y 11, showing that almost ai l
newly s ynthesised PG molecules are incorporated int O the extracellular rnatrix, with
very little being released into culture medium. In the culture media, the amount of
"S-labelled PO molecules released is no more than 2% at any given t h e (Fig 4b) and
shows day to &y variation. The drop in PG synthesis in cartilage may be a result of
feedback regdation by the chondrocytes to genaate fewer aggrecan molecules once a
steady state is achieved. Altematively, it may refiect unsatisfactory cu lNe conditions
to maintain synthesis at elevated levels.
harvest day
Figure 4. Changes in the Levels of Newly Synthesised PGs CsS-Sulphate labelled) in Cartilage (4a) and in Culture Medium (4b), with time in culture. Cartilage pieces were labelled with 25 pWml 3SS-Sulphate for 6 hrs, every 24 hrs. Cartilage and medium were harvested on days 0, 1,2,3,4, 5, 6, 7, 8, 9, and 10. The tissue weight is expressed on a wet weight basis. The mean and standard deviation values from 4 wells is indicated for each harvest day.
The levels of 846 epitope in the cartilage during this period do not reflect aggrecan
biosynthesis (Fig. 5a) and show that up to day 5, the epitope level i~ the tissue is
relatively constant. Thereafter, the levels virry every 24 hrs. The 846 epitope levels in
the cartilage represent an average of 99.99% of the total epitope levels at all days in
culture, showing that nearly d l molecules bearing the epitope are incorporated into
the matrix. These changes are however difficult to interpret due to the hi&
endogenous amounts of th is epitope in foetal cartilage which may mask any real
changes in the levels of newly synthesised epitope. It is therefore difncult to know
whether the 846 epitope is actually on the newly synthesised aggrecan molecules. In
the culture media, the levels of the epitope are no more than 0.02% of the total
epitope levels at any given time in culture, and show a certain degree of variation
(Fig. 5b). Furthamore, no obvious relationship cm be observed between the amount
of 846 epitope and newly synthesised aggrecan molecules released into culture
medium (Figs. 4b & 5b). These observations are also difficult to interpret for reasons
mentioned above, but may suggest either that aggrecan molecules lacking the 846
epirope are released into culture medium, or that smaller non-aggregating PGs
deficient in the 846 epitope are being released
Figs. 6a and 6b show the amount of total GAG in the cartilage and in medium,
respectively. The levds in the cardage are almost constant and represent 99.7%,
99.8% and 100% of the total GAG in the systm on harvest days 1, 6 and 10,
respedvely. The figures indicate that as with the "S-labellecl PGs and the 846
O 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 harvest day
Figure 5. Changes in the Levels of 846 Epitope @y radioimmunoassay) in Cartilage (5a) and Culture Medium (Sb) with time in culture.The cartilage and medium were harvested on days O, 1,2,3,4,5,6,7,8,9 and 10. The tissue weight is expressed on a wet weight basis. The mean and standard deviation values from 4 wells is indicated for each harvest day.
O 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 HARVEST DAY
Figure 6. Changes in the Levels of Total GAG (by DMMB assay) in Cartilage (6a) and Culture Medium (6b) with tirne in culture.The cartilage and medium were harvested on days O, 1, 2, 3,4, 5, 6, 7, 8, 9 and 10. The tissue weight is expressed on a wet weight basis. The mean and standard deviation values from 4 wells is indicated for each harvest day.
epitope, only a smdl proportion of GAGS are released into culture medium at any
point in time in this culture system.
The ratio of 846 epitope to GAG in cartilage and culhue medium is illustrated in Fig.
7. This increased ratio in cartilage may indicate a higher density of epitope p a CS
chah, or more CS chah bearing the epitope on the aggrecan molecules which are in
the cartiiage, in cornparison to those that are in the medium. This observation may
indicate a possible role for 846 epitope bearing molecules in the assembly of the
cartilage matrix. Since the cartilage was only labelled with '3-sulphate in the last 6
hrs prior to harvest, the relative amount of 35S-sulphate to GAG (cpm/GAG) was not
detennined.
Attempts made at immunoprecipitating "S-sulphate labelled, 846 epitope-bearing
PGs using 846 antibody immobilised onto magnetic beads, were unsuccessful despite
repeated attempts (data not shown). This was due to the propaties of the 846
antibody which belongs to the IgM isotype. These antibodies bind weakly to their
ligands (see discussion).
These data nevertheless demonstrate that due to its high endogenous 846 epitope and
total GAG content, the foetal cartilage explant system poses several problems with
respect to the interpretation of any infoxmation obtained. Furthamore, the
experiments performed using this system w e e not reproducible. Therefore, it t a s
decided that a more suitable system for the study of newly synthesised aggrecan
O 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 HARVEST DAY
Figure 7. The Ratio of 846 Epitope to GAG present in the Cartilage (a) and Released into Culture Medium (O) on each Harvest Day. The mean and standard deviation values fiom 4 welIs is indicated for each harvest day.
molecules and of the 846 epitope would be a chondrocyte culture systan where the
resident matrix has first been removed and in which only newly synthesised matrix is
present.
3.2. Foetal bovine chondrocyte cultures
In orda to investigate the hypothesis that the 846 epitope is present on newly
synthesised aggrecan molecules, rather than being a product of the modification of
the existing aggrecan in the ma&, hi& density monolayer cultures of foetal bovine
chondrocytes in senun contalliing m e h were established to examine aggrecan
biosynthesis. This culture system was chosen, firstly, because foetal chondrocytes are
more biosyntheticaliy active compared to adult chondrocytes, and would therefore
actively synthesise aggrecan as they try to establish a new matnx around them. In
addition, foetal bovine cartilage is much more readily available than foetal human
cartilage. Furthamore, by studying aggrecan biosynthesis in an isolated foetal
chondrocyte culture system rather than in a foetal explant culture system, it may be
possible to overcome the problems of having a resident matrix containing high levels
of the 846 epitope, which would mask the detection of any newly synthesised
epitope as was observed above (see 3.1 .).
The accumulation of newly synthesised aggrecan molecules in the celi layer was
followed over a perïod of 9 days to investigate whether the 846 epitope would be
deposited in the matrix as newly synthesised aggrecan molecules were being
deposited. The cultureci chondrocytes were harvested every 48 hrs, after labeihg
with 5-sulphate for a pexiod of 24 hrs prior to harvest. The effect of t h e in culture
on the levels of newly synthesised radiolabded proteoglycans is shown in Fig. 8.
The results show that approximately 80% of the total "S-sulphate labded PGs are
incorporated into the cell layer and about 20% are released into the culture medium,
at any given t h e examined in culture. Thus there is a preferential retention of the
newly synthesised PGs in the extracellular ma& of the chondrocytes. Indeed, one
observes an increase in the amount of ma& deposited in the ceil layer over the 9
days in culture (Fig. 9a) which is concomitant with the accumulation of GAG (Fig.
9b) and of the 846 epitope (Fig. 9c). The 846 epitope levels in the cell laya represent
(compared to media levels) 8 1% of the total amount of epitope on day 1, 93% on day
5 and 95% on &y 9. These observations show that aggrecan molecules bearing the
846 epitope are being prefereatialiy retained within the cell layer and may potentially
have a role in the formation of ma&. These studies have been performed on more
than one occasion and are reproducible.
In the culture medium, the release of newly synthesised metabolicdly labelled PGs is
accompanied by the release of the 846 epitope, both occuring at similar rates (Fig. 1 0)
suggesting that the 846 qitope is released with the release of newly synthesised
aggrecan molecules in this 'repir' system. When rneasuring the relative amount of
the 846 epitope to GAG, one sees that this ratio is vezy similar in the celi layer and in
the medium up to &y 3 in culture (Fig. 11) afier which there is an increase in this
ratio in the cell layer. This may indicate that there is a higher proportion of epitope-
containhg CS chah, or a higher density of the epitope per CS ch&, in the
O 1 2 3 4 5 6 7 8 9 1 0 HARVEST DAY
Figure 8. Changes in the Levels of Newly Synthesised Proteoglycans (3sS-sulphate labelled) with time in culture, in the Ce11 Layer (a) and in Culture Medium (O). Cultures were Iabelled with 50 pCi/rnl 35S-sulphate for 24 hrs on days 0,2,4,6 and 8. Cultures were harvested on days 1,3,5,7 and 9. The mean and standard deviaîion of 3 wells is plotted for each time point.
O 1 2 3 4 5 6 7 8 9 1 0 HARVEST DAY
Figure 9. Accumulation of Extracellular Matnv (9a), GAGS (9b) and 846 epitope (9c) in the Ce11 Layer with time in culture. GAG content is determined by the DMMB assay and the 846 epitope content is determined by radioimrnunoassay. The mean and standard deviation of 3 wells is plotted for each time point.
0.020 - " 0.018
' 0.01 6 m
a : 0.012 P - 0.0lO ; -0.008 8 i0.006 9
m
0.004 I
rn
O 1 2 3 4 5 6 7 8 9 1 0 HARVEST DAY
Figure 10. Release of Newly Synthesised 35S-sulphate labelled PGs (@) and the 846 Epitope (O) into the Culture Medium with Time in Culture. Cultures were labelled with 50 pCi/ml 35S-sufphate for 24 hrs on days 0,2,4, 6 and 8. Cultures were harvested on days 1, 3, 5, 7 and 9. The mean and standard deviation of 3 wells is plotted for each time point. (Spearrnan rank correlation of the data points show statistical significance at p = 0.0006, r =0.9 18, n = 15 for the correlation between the 846 epitope and 35S-sulphate labelled PG content in medium.)
O f 2 3 4 5 6 7 8 9 1 0 HARVEST DAY
Figure 11. The Ratio of 846 Epitope to GAG Present in the Celi Layer (a) and Released into Culture Medium (O) on each Harvest Day. The mean and standard deviation of 3 wells is plotted for each t h e point.
extracellular matrix of the c d laya: These data irnply that aU newly synthesised PGs
are not necessarily equal in structure, and that those containing the 846 epitope are
preferentially retained in the ma&. Since the ceil layw were only labelled with "S-
sdphate in the last 24 hrs pnor to hmest, the relative amount of US-sulphate to GAG
(cpm/GAG) was not determined.
Upon gel filtration analysis of the hydrodynamic sizes of the newly synthesised PGs
('3-sulphate labelled), in extracts of the ce11 layer from &y 1, it can be obsemed that
these molecules are large, eluting with a Kav of 0.30 (Fig. 12a) under dissociative
conditions. The 846 epitope is present on the largest of these newly synthesised
molecules (Kav = 0.26), and under associative conditions, these molecules can
aggregate with hyaluronan (Fig. 12b). Similar observations were made for PGs
extracted fkom the ce11 layer on otha harvest days (data not shown). The newly
synthesised PG molecules released into the culture medium on &y 1 are also large,
eluting with a Kav of 0.30 (Fig. 13a). The 846 epitope-bearing population also elutes
with a Kav of 0.3. As in the cell layer, these molecules also have the ability to interact
with hyaluronan (Fig. 13b), suggesthg therefore that they represent newly
synthesised aggrecan molecules which are not incorporated into the ma& (sec Fig.
8), possibly because they have a lower affinity for matrix retention or because there
are not enough 'binding' sites for their attachment onto hyduronan molecules in the
ma&. While "S-sulphate IabelIed PG molecules in the ceU layer and in culture
medium are similar in size, both e1uting with a Kav of 0.3 (Fig. 14), the ciifference in
Kavs of 846-bearing molecules in the celi layer and in culture medium (Fig. 15) may
-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Kav
Figure 12a. Sepharose CL-2B Chromatography, Under Dissociative Conditions of PGs Extracted From the Cell Layer on Day 1. The profiles of the 3SS-sulphate labelled PGs (a) and those bearing the 846 epitope (O) are illustrated. They elute with Kavs of 0.3 and 0.26, respectively. Ce11 layer PGs extracted on other harvest days showed a similar profile (data not shown).
Figure 12b. Sepharose CL-2B Chromatography, Under Associative Conditions, of PGs Extracted From the CeU Layer on Day 1. The profiles of the 35S-sulphate labelled PGs (a) and those bearing the 846 epitope (O) are illustrated. Ce11 layer PGs extracted on other harvest days showed a similar profile (data not shown) and eluted with Kavs of 0.0, showing complete aggregation with HA.
-0.4 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Kav
Figure 13a. Sepharose CL-2B Chromatography, Under Dissociative Conditions, of PGs Released into Culture Medium on Day 1. The profiles of the 3sS-sulphate labelled PGs (O) and those bearing the 846 epitope (O) are illustrated. PGs released into medium on other harvest days showed a similar profile (data not shown) and eluted with a Kav of 0.3.
Figure 13b. Sepharose CL-2B Chromatography, Under Associative Conditions, of PGs Released into Culture Medium on Day 1. The profiles of the 35S-suIphate labelled PGs (a) and those bearing the 846 epitope (O) are illustrated. PGs released into medium on other hamest days showed a similar profile (data not shown) and eluted with a Kav of 0.0, showing complete aggregation with HA.
-0.2 0.0 0.2 0.4 0.6 0.8 1 .O Kav
Figure 14. Cornparison of the Relative Hydrodynamic Sizes of 3SS-sulphate LabeLled PG Molecules From the Ceil Layer (a) and Culture Medium (O), harvested on Day 1. Both are shown to elute with a Kav of 0.3 and indicate that newly synthesised PG molecules in the cell layer and in culture medium are of the same size.
-0.2 0.0 0.2 0.4 0.6 0.8 1 .O Kav
Figure 15. Comparison of the Relative Hydrodynamic Sizes of PG Molecules Bearing the 846 Epitope from the CeU Layer (a) and Culture Medium (O) harvested on Day 1. They are shown to elute with Kavs of 0 .î6 and 0.3 0 respec tively, and uidicate that 846 epitope bearhg molecules in the ce11 layer are generally larger in size.
suggest that the molecules in the cell laya have a higher GAG content and are
therefore slightly Iarga in size. On the other hand, it may be possible that the
molecules released into the culture medium have lost their G3 globular domain by
proteolysis, since a role for this domain in the retention of newly synthesised
aggrecan molecules in the matrix of the cell layer has been proposeci (Flatltlery et al.,
1992).
This experiment demonstrates therefore, that aggrecan molecules bearing the 846
epitope are synthesised by isolated foetal chondrocytes as they establish a new
extracellular ma&, and that the majonty of these molecules are incorporated into the
cell layer whilst about 20% are released into the culture medium in a functional form.
It also suggests that there may be an enrichment of CS chains containing the 846
epitope, in the ceIl layer, with t h e in culture, in cornparison tu those released into the
medium. This may indicate that this epitope has a role in the repair process of the
extraceUular ma&.
3.3. a) Normal aduit human explant cultures
In view of the above observations that there was an increased retention of aggrecan
molecules bearing the 846 epitope in the c d layer of foetal bovine chondrocyte
cultures and that these molecules were mriched in this epitope, we wanted to
investigate whether the 846 epitope was reflective of the synthesis of matrix
molecules that may be incorporated as part of a "repair" process in human articular
cartilage. Iti these experiments, we cultureci normal adult human cartilage, which
contains a very low content of the 846 epitope (in comparison to foetal cartilage),
having treated it with or without trypsin prior to culture under various conditions.
This was an attempt at causing injury to the tissue and aeating an in viho arthritis
mode1 by depleting the cartilage ma& of its resident proteoglycans and otha non-
colîagenous proteins. The aim was to see whether the 846 epitope would be produced
in iacreased amounts in trypsin-treated cultures in comparison to cultures not treated
with trypsin.
The treatment of the cartilage explants with trypsin results in the loss of about 50% of
the total GAG fiom the tissue (Fig. 16). Further loss of GAG is observed afta 2 days
in culture, possibly due to some residual trypsin activity, but is thereafta maintained
at a steady level of approximately 10% of the starting PG content in ail three c d ~ e
conditions. Although PG synthesis does occur after the depletion of GAGS fiom the
cartilage matrix by trypsin, the level of synthesis is only about 3040% of that which
is observed in the cultures not treated with trypsin (Fig. 17), indicating that the
chondrocyte metabolism may have been affected due to possible loss of cell surface
molecules e.g. growth factor receptors, which enable the ceils to respond to their
culture conditions. The relative amount of 846 epitope to GAG is higher in trypsin-
treated cultures than in untreated cultures on &y 6 (Fig.
explants &ed in the presence of 10% FCS show the
showing that the presence of s a u n in the medium favours
expgMents wae perfomed on more than one occasion and
1 8) and trypsin-treated
highest 846GAG ratio,
epitope synthesis. These
were reproducible.
Before trypsin treatment fa After trypsin treatment
Control DMEM DMEM DMEM DMEM Day O
Day 6
Figure 16. The Effect of Trypsin Treatment on the Depletion of GAGS from the Cartilage Explants. The total amount of GAG (measured by the DMMB assay) in the cartilage before and after trypsin treatment is illustrated. The tissue weight is expressed on a wet weight basis. The control samples represent those that were not cultured and harvested on day 0. Day 2 samples are those that were cultured in DMEM for 2 days. Day 6 sarnples are those that were cultured in the indicated conditions for 4 days after initial culture for 2 days in DMEM alone. The mean and standard deviation values fiom 3 wells is indicated.
T Before trypsin treatment After trypsin treatment
1 DMEM DMEM + ITS DMEM + FCSl
Day 6
Figure 17. The Effect of Trypsin Treatment on the Synthesis of PGs in Different Culture Conditions. The amount of 35S-sulphate incorporated into the cartilage before and after trypsin treatment is illustrated. The tissue weight is expressed on a wet weight basis. Samples harvested on day 6 are shown. The mean and standard deviation values fiom 3 wells is indicated.
1 Before trypsin treatment After trypsin treatment
Control DMEM DMEM DMEM Day O I + ITS + FCS 1
Figure 18. The Effect of Trypsin Treatrnent on the Ratio of 846 Epitope to GAG Content of the Cartilage. The control samples represent those that were not cultured and harvested on day 0. Day 6 samples are those that were cultured in the indicated conditions for 4 days after initial culture for 2 days in DMEM alone. The mean and standard deviation values from 3 wells is indicated.
This experiment suggests that the 846 epitope is maximdy upregulated in serum
containing medium in response to tissue 'injury'. Although the epitope levels
measured in this system are much smalla (100-fold less) than those measured in the
foetal bovine system desnibed above, these changes can be clearly detected. These
observations are supportive of the relationship of the 846 epitope to the cartilage
repair process.
b) OA enolant cultures
The relationship between proteoglycan synthesis and the expression of the 846
epitope was studied in cultures of OA cartilage, since this tissue has previously been
shown to express elevated arnounts of the epitope (Rizkalla et al., 1992). In addition,
since very s m d amounts of the 846 epitope are detected in normal addt cartilage, in
cornparison to foetal cartilage, any increase in epitope detected in OA cartiiage is a
consequence of the disease. Studies of cuitures of OA cartilage in which aggrecan
synthesis has been measured @y labelling the cartilage with 35S-sulphate for 24 hrs)
would aiiow us to investigate whether the increase in the 846 epitope levels is related
to an increase in the synthesis of aggrecan molecules bearing this epitope in disease.
In the OA samples studied, a mean of 80% (+/- 16% SD) of 846-epitope bearing
molecules and 79% (+/- 15% SD) of YS-sulphate labeiled molecules are present in
the cartilage, whiie the remahhg are released into the medium. Fig. 19 shows that
there is no correlation betweai the synthesis of aggrecan, as determined by 3-
sdphate incorporation, and the amount of 846 epitope detected in the cartilage. This
may be due to the variability in metabolkm of the cartilage h m the 10 different
patients (due to the state of the disease, age, medication, mobility, etc.). In addition, it
may be due to variation in the amount of 846 epitope already present in the tissue
pnor to culture, and to the variability in sampling of the cartilage pieces for culture
(because different areas of the cartilage may be biosynthetically active or degradative
to diffaent levels and since the macroscopic appearance and degree of visible tissue
damage of the cartilage fiom the different patients also varied). Although these
experirnents wae not pdormed on normal addt cartilage, the 846 epitope content of
the cultured OA cartilages are approximat ely 1 00-fold higha compared to the epitope
levels described in normal cartilage by Rizkalla et ai. (1992). In the media fiom these
cultures, there is a statistically significant positive correlation between the contait of
"S-labelled PG molecules and of the 846 epitope (Fig. 20% with p = 0.0023 and r =
0.379). This correlation is indeed present even when d o s a analysis of the samples
near the intercept region is paformed (Fig. 20b, n = 47, p = 0.0001 and r = 0.708).
This suggests that the synthesis of aggrecan molecules is accompanied by the release
of some of these moledes h m the cartilage and that these molecules bear (in part at
least) the 846 epitope. This observation is reminiscent of the foetal bovine
chondrocyte cultures, in which some of the newly synthesised aggrecan molecules
bearing the 846 epitope are also released into the culture medium (Fig. 10). Fig. 21
shows no correlation between the amount of newly synthesised aggrecan ("S-
sulphate labelled) molecules in the cartilage and those that are released into the
medium @ = 0.9838, r = 0.003). No correlation is observed between the newly
synthesised aggrecan C5S-sulphate labelled) content of the cartilage and the mount
cpmfpg DNA (cartilage)
Figure 19. Spearman Rank Correlation Analysis of Newly Synthesised PG C5S-sulphate labeiled) and 846 Epitope Content of Cultured Cartilage from 10 OA Patients. Cartilage pieces were cultured for 48 hrs prior to labelhg with 50 pCi/mI 3SS-sulphate for 24 hrs. Cartilage fiom 66 wells were analysed.
cpmlpg DNA (medium)
cprnlpg DNA (medium)
Figure 20a. Spearman Rank Correlation Analysis of Newly Synthesised PG CsS-sulphate labelled) and 846 Epitope Content of Culture Media From Cartilage Cultures of 10 OA Patients. Cartilage pieces were cultured for 48 hrs pnor to labelling with 50 @/ml 35S-sulphate for 24 hrs. Media from 66 wells were analysed. ('Outliers' are identified as O)
Figure 2Ob. Spearman Rank Correlation Analysis of Newly Synthesised PG CsS-sulphate labelied) and 846 Epitope Content of Culture Media From Cartilage Cultures of 10 OA Patients. Data fiom figure 20a are analysed by omitting 19 samples (a).
Figure 21. Spearman Rank Correlation Analysis of Newly Synthesised PG C5S-sulphate labeiled) Content of Cartilage and Culture Media From Cartilage Cultures of 10 OA Patients. Cartilage pieces were cultured for 48 hrs pior to labelling with 50 pCi/ml 35S-sulphate for 24 hrs. Cartilage and media fiom 66 wells were analysed.
of 846 epitope released into culture medium @ = 0.308, r = O. 126, n = 66; no Fig.
shown) eithes. However, there is a correlation between the 846 epitope content in
cartilage and those molecules bearing this epitope released fiom the cartilage (Figs.
22a and 22b; p = 0.0018, r = 0.405, after closa analysis of sarnples near the
intercept). This may be representative of newly synthesised aggrecan molecules that
are not incorporated into the matrix, as has previously been observed in the foetal
bovine chondrocyte cultures desuibed above (see 3.2.).
B. 846 EPITOPE CONTENT OF ARTICULAR CARTILAGE AND SYNOVIAL
FLUIDS FROM OA PATIENTS
3.4.846 e ~ i t o ~ e content of OA cartilage extracts and svnovial fluids.
The 846 epitope has been shown to be presait in elevated amounts in the synovial
nui& of OA patients (Poole et al., 1994). It has been suggested that it may be a
useful rnarka in the monitoring of the disease process, since its increased detection in
the body fluids of arthritic patients may be indicative of increased synthesis of
aggrecan by chondrocytes. Analyses were made of the 846 epitope contents of a
single fidi depth specimen of the cartilage (between 100-600mg wet weight) and the
synovial fluid taken ftom the same knee, at arthroplasty. In ail, 37 OA hees were
examined to see whether there was any correlation between the cartilage content of
the 846 epitope and that in the synovial fluids of the same joint.
Fig. 23 shows a weak but signincant correlation @ = 0.0308, r = 0.36) between the
amount of 846 epitope present in the cartilage and the synovial fluid of patients. In
Figure 22a. Spearman Rank Correlation Analysis of 846 Epitope Content of Cartilage and Culture Media From Cartilage Cultures of 10 OA Patients. Cartilage pieces were cultured for 48 hrs pnor to labelling with 50 pCi/mi 35S-sulphate for 24 hrs. Cartilage and media fkom 66 wells were analysed. ('Outliers' are identified as a)
Figure 22b. Spearman Rank Correlation Analysis of 846 Epitope Content of Cartilage and Culture Media Ftym Cartilage Cultures of 10 OA Patients. Data fkom Fi,we 22a are analysed by omitting 6 samples (a).
pg 846lrng cartilage
Figure 23. Spearman Rank Correlation Analysis of 846 Epitope Content of Articular Cartilage and Synovial Fluids from 37 OA Patients.
light of our in vitro data fiom bovine chondrocyte cultures, which show that a smdl
proportion of newly synthesised aggrecan molecules are released into culture
medium, it thus appears that the level of epitope which is measured in die synovial
fiuids of OA patients is representative of some of the newly synthesised aggrecan
molecules ffom the cartilage of these patients. However, it is also important to note
that the 846 epitope levels in the synovial fluids could also reflect the level of
degradation of resident as well as newly synthesised aggrecan molecules, if these
molecules are large enough so as to provide a high epitope density for detection by
the 846 radioimmunoassay (see 3.6.). Furthermore, the molecules detected in the
synovial fluid may only represent those aggrecan molecules which have a low affinity
for retention within the cartilage, since we have previously shown that the structure of
newly synthesised aggrecan moleniles in the cartilage may be different fiom those
released fkom the tissue (3.2.), however no analyses were performed on the sizes of
the aggrecan moleniles released into synovial fluid.
C. THE STRUCTURE AND LOCATION OF THE 846 EPITOPE
3.5. Chondroitinase ABC and AC11 time course meriment
In order to investigate the location of the 846 epitope on the CS chahs, i.e to
determine whether the epitope was present near the reducing (core protein attachment
region) or non-reducîng end of the CS chai., partial digestions were performed with
chondroitinases. This technique has previously been used by Hardingham et al.
(1994b) to map the positions of the 3B3, 3D5 and 7D4 epitopes on CS chahs. The
DMMB assay is used to quantitate aay ranainllig GAG on the PG molecuie, since the
dye does not react with the smaller products of digestion. Digestion with
chondroitinase ABC and AC11 results in a rapid loss of the 846 epitope and GAG
fkom the PG molecules Within the nrst 30 mins of digestion with chondroitinase
ABC, about 60% of the epitope and 30% of total GAG are lost (Fig. 24), whereas
with chondroitinase AC& approxhately 50% of epitope and 20% of GAG are lost
(Fig. 25). But by the end of 120 mins, 40% of GAG and 90% of epitope is lost with
chondroitinase ACII (Fig. 25), whereas with chondroitinase ABC, 60% of GAG and
70% of the epitope are lost (Fig. 24). Chondroitinase ABC cm act both as an endo-
and exo- glycosidase @ut preferentidy acts an endo-glycosidase), whereas
chondroitinase ACII is an exoglycosidase, sequentially removing disaccharides from
the non-reducing end of CS chains (Hardingham et al., 1994b). Thus upon digestion
with chondroitliase ACT& a non-reducing terminal epitope would be lost
immediately, and at a fater rate than the loss of GAG. Hence, these expaiments
have proven firstly, that the 846 epitope is located on CS chains and is lost upon
digestion with chondroitinase enzymes. Furthemore, the immediate loss of the
epitope upon digestion with chondroitinase AC& as weli as its faster rate of loss
compared to GAG loss, suggests bat the epitope is located near the non-reducing end
of the CS chains ratha than being randomly dismbuted or located near the reducing
terminus.
3.6, Dot blot anaIvsis of CS chains ~ r e ~ a r e d fkom foetal and aduIt PG
It has previously been demonstrated that the ability to detect 846 epitope by
radioimmunoassay is much reduced in foetal aggrecan treated with pro teinases such
O 60 120 180 240 300 digestion tirne (mins)
Figure 24. The Effect of Chondroitinase ABC Treatment of Foetal Bovine PG, on the Loss of GAG and of the 846 Epitope. A total of 50 pg GAG (determined by DMMB assay) was digested with 0.005 U enzyrne/mg GAG for 0,2,5, 15,30,60, 120 and 240 minutes. The % of total GAG @y DMMB assay) and 846 epitope (by 846 RIA) remaining is illustrated here.
O 60 120 180 240 300
Digestion time (mins)
Figure 25. The Effect of Chondroitinase ACII Treatment of Foetal Bovine PG, on the Loss of GAG and of the 846 Epitope. A total of 50 pg GAG (determined by DMMB assay) was digested with 0.005 U enzyme/mg GAG for 0,2,5, 15,30,60, 120 and 240 minutes. The % of total GAG @y DMMB assay) and 846 epitope @y 846 RIA) remaining is illustrated here.
as pronase and papain, which cleave the aggrecan core protein (Glant et al., 1986).
This observation suggests that in order for the 846 epitope to be detected by the IgM
antibody '846', a high epitope density is required In other words multiple CS chains
have to be presented to the antibody for epitope detection. This may be due to the
pentavalent nature of IgM antibodies, resulting in lower affinity of binding of a single
epitope. It has also been observed that the 846 epitope is barely detectable in normal
adult cartilage (Glant et al., 1986). This was true also for the 383 epitope (Caterson et
(II., 1990b; Caterson et al., 1995). Howeva recent studies have shown that the 3B3
epitope is indeed present in normal aduit cartilage, but that the manner in which the
CS chains are presented to the antibody 3B3 is critical for epitope detection (Plaas et
al., 1997).
We wanted to investigate whetha or not the reason for the lack of 846 epitope
detection in adult cadage was due to the same reasons. In order to answer this
question, we developed a blotting technique which would enable the detection of the
846 epitope on single CS chains, since a solution phase radioimmunoassay cannot be
used for this purpose. In Fig. 26, we demonstrate that the 846 epitope cm be
recognised if high concentrations of foetal bovine aggrecan CS chains are applied to
the PVDF membrane and that this detection is concentration dependent. Fig. 27
Uustrates that in cornparison to CS chains prepared fiom foetal PG, the 846 epitope
is barely detectable on CS chains prepared fiom normal aduit human PO using even
this technigue and therefore the epitope is indeed absent fiom CS chains of addt
aggrecan.
Figure 26. Dot Blot of CS Chains Prepared from Foetal Bovine PGAfter Papain Digestion and CPC Precipitation. A total of 500, 250, 125 and 62 pg CS (detemiined by DMMB assay) was applied to PVDF membranes. The 846 ascitic fluid was used at 1:200 dilution, The control blot was performed using the same dilution of 846 ascitic fluid preabsorbed with foetal bovine PG.
Figure 27. Dot Blot of CS Chahs Prepared from Adult Buman PGAfter Papain Digestion and CPC Precipitation. A total of 500, 250, 125 and 62 pg CS (determined by DMMB assay) was applied to PVDF membranes. The 846 ascitic fluid was used at 1:200 dilution. The control blot was performed using the same dilution of 846 ascitic fluid preabsorbed with foetal bovine PG.
3.7. Effect of CS chain leneth on 846 e ~ i t o ~ e egpression
The aim of this experiment was to investigate whether or not CS ch& length had an
effect on the 846 epitope distribution. In order to achieve this, CS chah prepared
fiom OA human cartilage and foetal bovine PG were separated by gel filtration
chromatography on a Sepharose CL-€33 column in order to separate CS chains of
different lengths nom these PG preparations. Figs. 28 and 29 illustrate the size
distribution of CS chains fiom foetal bovine and OA human PG, respectively. CS
chains &om foetal bovine PG eluted with Kavs in the range 0.24 to 0.5, whereas those
fkom OA cartilage eluted in the range 0.0 to 0.3 1. Fractions representing the various
chain sizes were assayed for the 846 epitope on dot blots. Unfortunately, difficdties
in blotting samples with high viscosity are iliustrated in the blots of the samples kom
the fiactions of foetal bovine PG (Fig. 30). Nevertheless, both Figs. 30 and 31 show
that the 846 epitope can be detected on all CS chains, in both human OA and foetal
bovine PG, irrespective of chah length.
3.8. Non-reducinp terminal residue analvsis of 0-ducuronidase treated foetal
bovine PG.
CS chains can have 4 possible non-reducing taminal stmctures (Table 3), containhg
either a GlcA or a GaNAc as the non-reducing terminal residue. Since our
experiments suggested that the 846 epitope was located near the non-reducing end of
CS chahs, we wanted to investigate whether or not it was in fact present on the non-
reducing end, and if so, which terminal residue was involved in its structure. W e
therefore treated foetal bovine PG with a mammalian source of P-glucuronidase
0.0 0.2 0.4 O .6 Kav
Figure 28. Sepharose CLdB Chromatography of CS Chains Prepared From Foetal Bovine PG, by Papain Digestion and CPC Precipitation. Between 5-10 mg GAG were chromatographed and the GAG content in the fiactions deterrnined by the DMMB assay.
-0.2 0.0 0.2 0.4 0 -6 0.8 1 .O Kav
Figure 29. Sepharose CL-6B Chromatography of CS Chains Prepared from Extracts of OA Cartilage, by Papain Digestion and CPC Precipitation. Between 5-10 mg GAG were chrornatographed and the GAG content in the fractions determined by the DMMB assay.
0.29 0.31 0.33 0.36 0.38 0.40 0.43 0.45 Kav
Figure 30. Dot Blot of Sepharose CL-6B Fractions of CS Chains from Foetal Bovine PG. A total of 150 pg CS (by DMMB assay) from fractions representative of the elution profile, was applied to PVDF membranes and blotted using 846 ascitic fluid at a 1:200 dilution. The control blot was performed using the same dilution of 846 ascitic fluid preabsorbed with foetal bovine PG.
0.09 0.14 0.19 0.24 0.29 0.33 Kav
Figure 31. Dot Blot of Sepharose CL-6B Fractions of CS Chains from Extracts of OA Cartilage. A total of 150 pg CS (by DMMB assay) from fiactions representative of the elution profile, was applied to PVDF membranes and blotted using 846 ascitic fluid at a 1:200 dilution. The çontrol blot was performed using thc same dilution of 846 ascitic fluid preabsorbed with foetal bovine PG.
(which removes the non-reducing terminal GlcA) to see if this would result in the loss
of epitope detection.
Table 5 surnmarises the effect of P-glucuronidase treatment on foetal bovine PG after
analysis of the samples for changes in their non-reducing taminal componmts by
ion-exchange HPLC, in collaboration with Dr Anna Plaas, Shriners Hospital, Tampa,
FL. These samples were also chromatographed on a Superose-6 column to check for
any changes in hydrodynamic size of the PG moledes upon digestion with P-
glucuronidase (due to possible contaminating proteases). Fortunat ely , no proteoly sis
was observed (data not shown). The P-glucuronidase treatment shows that up to 69%
of non-reducing terminal GlcA-GdNAc4S residues are lost, afia digestion with 10
and 1OOU of enzyme. (Note that this is the only type of terminus containhg GlcA in
these samples as non-reducing taminal GlcA-GaNAc6S (Di6S) residues are not
detectable). However, none of the concentrations of enzyme used result in any loss of
the 846 epitope, indicating that the non-reducing terminal GlcA residue is not
involved in the 846 epitope. The loss of non-reducing terminal GalNAc4,6S residues
with increasing P-glucuronidase concentration suggests that the enzyme preparation
is contarninated with a terminal 6-sulphatase. Ho wever, this fornittous observation
indicates that taminal GalNac4,oS residues are also not involved in the epitope
structure, since the 846 epitope content ranains constant. It is reassuring to know that
no contaminating hexosaminidases are present since the number of disaccharides per
chah, and therefore CS chah length, remains constant. Mammalian hexosaminidases
would cleave the non-reducing taminal GaNAc residues on CS chains, thus
1 O 200 113 21.2 Not detected 29.92 952 I 2 10 271 63 6.6 Not detected 29.36 962
3 100 260 25 6.4 Not detected 34.32 975
Table 5. Unsaturated Disaccharide and Non-reducing Terminal Residue Analysis of Foetal Bovine PG Treated with a-glucuronidase. A total of I mg PG was digested with 0, 10 and 100 U of enzyme. The 846 epitope content of the samples was detennined by radioimmunoassay alter digestion with B-glucuronidase. The samples were ethanol precipitated to separate the PG molecules from the products of digestion. The non-reducing terminal structures of the precipitated material were analysed (in collaboration with Dr. Anna Plaas, Shriners Hospital, Tampa, FL) by ion-exchange HPLC.
decreasing CS chah length. These observations suggest therefore that after B-
glucuronidase treatmait, the ody non-reducing terminal residue that is not reduced in
content is GalNAc4S. The content of this residue in fact increases sornewhat, due to
the contamhating activity of the 6-sulphatase. These data indicate that if the 846
epitope resides on the non-reducing terminus of the CS chains, then only terminal
GaINAc4S residues could be hvolved in its structure.
4. DISCUSSION.
The need for biochemical markers of joint damage in arthrîtis is becoming
increasingly important in ordex to understand changes in the metabolism of joint
tissues such as articular cartilage, in vitro and in vivo. Traditional imaging techniques
such as radiographic analysis and more r e c d y , magnetic resonance imaging (MRI),
have their limitations, in that they only provide information on the consequences of
the disease process, by which t h e irreversible damage may have occurred. These
changes may include the destruction of cartilage and hcreased joint space narrowing,
bone destruction and bone ranodelling. Such imaging techniques are unable to
intapret signals regarding specific molecular changes in the matrix molecules. In
contrast, the use of tissue specific biochemical markers of joint metabolism (synthesis
andfor degradation) may provide valuable information on early changes in tissue
metabolism in pathology and enable the monitoring of disease activity. In addition,
they may be of use prognostically and help in evaluating the effects of different
therapies in controhg disease activity and promoting tissue repair.
The development of sensitive immunochemical assays for the measurement of a
panel of tissue specific macromolecules and their biosynthetic and degradation
products is beginning to provide some insight into the pathogenesis of arthritis
(Poole, 1994; 1997; Poole and Dieppe, 1994). The chondroitin sulphate 846 epitope
is one such biosynthetic markn of cartilage aggrecan which has been measured in the
synovial fluid and s a u m of -tic patients, where it is found in elwated amounts
compared to non-arthritic individuals (Poole et al., 1994; Mdnsson et al., 1995).
In this thesis, 1 have tried to bnter understand aggrecan synthesis, by pdorming in
vitro experiments on isolated cells and cartilage explants in order to establish whether
or not the 846 epitope is ûuly associated with the synthesis of aggrecan molecules,
and not a product of modification of the existing aggrecan in the extracellular rnatrix
of cartilage. Such studies are crucial in the field of markers, for one should be M y
aware of what is being measured and what the markers represent in vivo. 1 have also
attempted to characterise, using various enzymatic treatments, the structure and
location of the 846 epitope on the aggrecan molecule.
Initial experiments to study the synthesis of aggrecan and of the 846 epitope were
perfonned on foetal bovine cartilage explant.. Although this system has certain
advantages by allowing chondrocytes to be contained within their onginal
extracellular matrix, therefore mimicking their in situ environment, experiments
using this culture system were discontinued fot several rasons. The high levels of
endogenous 846 epitope in foetal cartilage made the Uiterpretation of any data very
difticult, since newiy synthesised 846 epitope could not be distinguished fkom the
existing epitope in the matrix. No relationship was observed between aggrecan
synthesis and the levels of the 846 epitope both in the cartilage and in culture
medium. Furthmore, there was much variation in the data, with Little reproducibilty
of the expaiments. The immunoprecipitation of US-sulphate-labelled 846 epitope-
containing aggrecan molecules fkom the cartilage extracts and media would have
answered the question of whetber or not the 846 epitope was present on newly
synthesised aggrecan molecules. It would also have enabled the quantitation of the
proportion of newly synthesised molecules that contain the epitope and the
determination of the haf-life of these molecules in the tissue. However, repeated
attempts at imrnunoprecipitating 846-bea~ing 35S-sulphate-labelled aggrecan
molecules were unsuccessful due to the nature of the antibody 846, which belongs to
the IgM class of monoclonal antibodies. This class of antibodies has sevaal
properties which pose limitations to their practical applications. These will be
addressed Iater on in the discussion.
Experiments using foetal bovine chondrocyte cultures eliminated the problems of
having a resident matrix containing hi& levels of the 846 epitope. Any epitope
detected using this culture system would thus be a produa of new synthesis of
aggrecan. The data f?om these experiments indicate that the 846 epitope is present on
newly synthesised aggrecan molecules and that these molecules accumulate in the
cell layer, as the cell layer develops with tirne in culture. The observation that the
majority of these epitope-bearing moledes are retained within the cell layer,
together with the fact that the moleniles in the cell layer have a higher epitope
density compared to those that are released into the culture medium, suggests that the
846 epitope-bearing molecules may have a role in the formation of the extracellular
matrix of cartilage. The higha epitope density on the aggrecan molecules
preferentidy retained within the cell laya may be due either to the synthesis of more
epitope per CS chain, or the synthesis of a greater nurnber of epitope-containing CS
chahs. This epitope may thus be generated by chondrocytes when thae is a
requirement for an increase in aggrecan synthesis, such as that occuring in a repair
process (Mankh and Lipiello, 1971; Thompson and Oegema, 1979; Sandy et al.,
1984).
Analysis of the media fiom these cultures shows that the 846 epitope is released
together with the newly synthesised aggrecan molecules, and is therefore a marka of
synthesis. However, the molecules in the medium have a lower epitope density
compared to those retained in the cell layer. This is an indication that ail newly
synthesised aggrecan molecules are not necessarily the same in structure and do not
contain the same arnount of @tope. However, irrespective of this structural
difference, the fact that the 846 epitope content of the medium correlates with the
arnount of newly synthesised PG in the medium, and that this amount is always a
fked proportion of the total amount of newly synthesised PG, indicates that the 846
epitope cm be used as a marker of total aggrecan synthesis in such a culture system.
Using gel filtration chromatography, it was shown that in the cell layer, the 846
epitope is present on the larger population of newly synthesised aggrecan molecules
which aggregate M y with HA. Analysis of proteoglycans extracted fiom
osteoarthritic cartilage has also shown that the 846 epitope is present on the largest of
these proteoglycan molecules (Rizkalia et ai.,1992) which aggregate with HA. In
contrast, the present data have shown that the 846-bearing aggrecan molecules
released into cell culture media are slightly smaiier in size, but neveaheless still
aggregate with HA, showing that they are not released into the medium as a result of
proteolytic processing between the G142 globular domains. Further evidence
supporting the observation that aggrecan molecules nom the medium of chondrocyte
cultures retâin their abrlity to bind HA, has also been provided by Sandy et al. (1989).
It is possible that the 846-epitope beaiing molecules released into culture medium
rnay have a lower &ty for retention within the matrix. One explmation could be
the presence of insufncient binding sites on the HA molecules in the matnu. It is also
possible that the molecules reieased into culture medium exhibit delayed aggregation
to HA, a process which has been identified as occuring following PG synthesis and
secretion from the chondrocyte (Sandy et al., 1989; Melching and Roughley, 1 990).
The smaller size of 846-bearing molecules in the culture media rnay either indicate
that these molecules contain fewer CS chains or that they result fiom proteolytic
cleavage within the c a r b o x y - t d a l region of the core protein prior to release into
culture medium. The proteolytic removal of the carboxy-taminal G3 globular
domain has been O bserved during the extracellular processing of aggrecan in newbom
calf chondrocyte cultures (Flannery et al., 1992). It has been shown, in these calf
chondrocyte cultures, that the medium contains a higher G3 content relative to the
cell layer, suggating that the retention of aggrecan molecules in the cell laya is
accompanied by rapid proteolytic tanoval of the G3 domain, which enables some of
the molecules to diffuse into the medium of these cultures.
Since the 846 epitope is present on the Iargest, newly synthesised aggrecan molecules
in the ceil layer, it is possible that it may be prefaentially located on CS chains near
the C-terminal G3 globular domain of the aggrecan molecule. The cleavage of the
aggrecan core protein near this domain may also remove part of the CS-2 domain (see
Fig. 2), resulting in the loss of the epitope fkom a proportion of molecules in the
culture medium. The lectin-like domain in the G3 dornain of aggrecan has been
shown to have the capacity to bind carbohydrates, such as fucose and galactose
(Halberg et al., 1988) as w d as HA (Yang et al., 1995), indicating that this domain
may be involved in binding to cellular or matrix ligands. If the 846 epitope is indeed
located near the G3 domain, it may explain why, in the chondrocyte cultures, the
prefaential retention of 846-bearing aggrecan molecules was obsmed within the cell
layer. Thus it can be envisaged that the 846 epitope is a marker for the formation and
repair of the matrix. Evidence h m the work of Cheng and Kimura (1994) shows that
CS structure can indeed Vary between the K and C-tamini of the aggrecan core
protein. It is therefore possible that the 846 epitope is located on selected CS chains
near the G3 globular domain of aggrecan.
Adult articular cartilage explant cultures were established to investigate M e r the
expression of the 846 epitope in a repair process. Since the 846 epitope is barely
detectable in nomal adult cartilage, a study was conducted to detamine whether
treatment of the cartilage wîth trypsin (to mimic the proteolysis that occm in arthritic
cartilage) would induce the chondrocytes to undergo a phenotypic change, such as
that which is suggested to occur in OA (Byers et al., 1992; Walker et ai., 1995),
resulting in the synthesis of more foetal-type aggrecan molecules as part of an
attempted repair process. Treatment of the cartilage with trypsh would not only
damage and remove the resident matrix proteoglycans, but it may also cleave off
some of the cell surface receptors on the chondrocytes, such as integrins or heparan
sulphate proteoglycans, which would affect cell-matrix interactions. This could
subsequently alter the phenotype of the cens, stimulahg them to produce 'repair-
like' rnatrix molecules. Hence one could study the repair process in an in vitro 'OA
model' of cartilage. The data show that PG synthesis is reduced &er trypsin
treatment, possibly because the cells are unable to respond Mly to their culture
conditions as a result of cell damage and/or the removal of cell surface receptors.
Nevertheless, as in the foetal chondrocyte cultures, there is an increase in the 846
epitope to GAG density in the articular cartilage &a trypsin treatment. This is
further evidence that the 846 epitope is generated as part of a 'repair process', and
may explain its elevated expression in OA cartilage.
The 846 epitope has been found in elevated amounts in the degenerate articular
cartiiage (Rizkalla et al., 1992) and body fluids of OA patients (Poole et al., 1994).
Its measurement in the body fluids of artbritic patients is thought to be indicative of
aggrecan synthesis as part of the repair process in the disease. In spite of the problems
encountered with foetal bovine explant cultures, we therefore thought it important to
study the synthesis of aggrecan and investigate its relationship to the expression of
the 846 epitope in explant cultures of samples of OA cartilage obtained following
arthroplasty. Our data show that there is no conelation between the levels of "S-
Iabelled PG and of the 846 epitope in the cartilage fiom these patients. There may be
several reasons for this, including variations in patient age, their disease duration, the
types of therapy they are undergoing, as w d as the stage of the disease at the time of
aahroplasty. These parameters would affect the level of the 846 epitope present in
their joints at the t h e of surgery, so that the endogenous levels of epitope in the
cultures could be v q variable, thus masking any newly synthesised epitope.
Howeva, in the culture medium, there is a statktically significant correlation
between the levels of newly synthesised aggrecan and the epitope content,
connrming our previous observations, fkom the foetal bovine chondrocyte cultures,
that the 846 epitope cm be used as a marker for aggrecan synthesis. It is important to
remember, however, that since the structure and epitope density of the newly
synthesised aggrecan molecules released into tbe medium may be merent f?om
those in the tissue (as was previously denonstrated in the chondrocyte cultures), the
amount of the epitope in the medium rnay not necessarily be representative of dl
newly synthesised aggrecaa moledes in the tissue. It is also no? known what
proportion of newly synthesised aggrecan molecules contain tiiis epitope (due to the
limitations of the immunoprecipitation technique). The above limitations need to be
considered when measuring the epitope in the body fluids of patients. In the OA
cartilage culhues it was also observed that the content of the 846 epitope in cattiiage
is also reflected by the release of epitope-bearing molecules into the culture medium.
These may represent the newly synthesised 846-bearing molecdes that have a lower
affinity for matrix retention, such as those seen in the chondrocyte cultures. On the
other hand, one cannot discount that the moledes in the medium rnay dso arise
fÏom the degradation and release of the resident cartïîage 846-bearing molecules.
Whai meauring 846 epitope levels in OA patients, a weak correlation is observed
between the cartilage and synovial fluid levels of the epitope, indicating that the
content of the epitope in the cartilage is reflected in the release of 846-bearing
molecules into the synovial fluid. Even though there is a correlation between levels of
aggrecan synthesis and of the 846 epitope in the medium fiom cultures of OA
cartilage, synovial fluid levels of the epitope could be influenced by several other
factors in addition to patient variation and resident 846 epitope content variation
described above. An increase in synovial fluid levels of the epitope could arise fkom
the degradation and release of existing as well as newly synthesised 846-bearing
molecules nom the cartilage. In other words, increased detection of the 846 epitope
in synovial fluid could mean either increased synthesis of aggrecan by chondrocytes,
or increased degradation of the 846-bearing molecules in the cartilage matrix. The
synovial fiuid epitope levels could also reflect newly synthesised aggrecan molecules
that have not undergone aggregation with HA in the matrix and are consequently
released ftom the tissue. In contrast, decreased detection of the epitope in synovial
fluid could be a consequaice of elevated clearance rates of the molecules fiom the
joint or increased retention of 846-bearing molecules within the cartilage.
Furthermore, smaller degradation products containhg the epitope may nevcr be
detected due to the limitations of the 846 IgM antibody, which will be discussed
shorlly. In summary therefore, the measurement of the 846 epitope in the synovial
fluid of patients would be indicative of a repair pmcess only if no epitope was present
to begin with, for instance in the v a y early phases of the diseases. But in well
established OA, with an elevated content of the 846 epitope in the tissue, it would be
dinicult to teil whether the epitope detected had arisen from synthesis and repair or
from the degradation of existing 846-bearing molecules. Hence the use of 846
analysis in isolation may be of limited value and it may be necessary to analyse a
variety of biochemical markers of both spthesis and degradation, in individual
patients, in order to better understand the dynamics of tissue turnover in pathology.
As rnentioned already, antibody 846 which has been used in this work has an IgM
isotype. Such antibodies have several limitations regarding their use due to their weak
affinity for their ligands, which, in this case, is a carbohydrate epitope. Consequently,
it has provm very difncult to immunoprecipitate metabolicaliy labelled PG molecules
bearing the 846 epitope fiom cartilage extracts in the foetal bovine explant cultures
described above. It should be noted that due to the pentameric nature of the antibody,
epitope detection requires interactions with multiple CS chahs. Initial evidence for
this property came from the work of Glant et al. (1986), who showed that treatment
of foetal proteoglycan with proteases such as papain, that cleave tbe core protein at
numerous sites, resulted in loss of epitope detection by the 846 antibody in a solution
phase radioimmunoassay. Hence the 846 epitope was originaily describeci as a core
protein-related epitope (Glant et al., 1986). This feanire is, howeva, indicative of the
requirement for a high epitope density for the detection of this epitope. A similar
property has aiso been demonstrated by the anti-KS IgG antibody, AN9PI. The
univalent Fab preparation of this antibody binds l a s favourabiy to PG compared with
the IgG (Poole et al., 1989). This phenornenon is particularly relevant when
measuring the 846 epitope in body fluids, where s m d degradation products
containing the epitope may not be detected by the solution phase cornpetition
radioimmunoassay used to quantitate this epitope. AIthough digestion of foetal PG
with papain results in Ioss of epitope detection in the solution phase assay, direct
binding analyses (dot blots) on irnmobilised single CS chains prepared by papain
digestion of the PG demonstrated that the epitope can be detected on single CS chains
if they are applied in high concentrations onto PVDF membranes. This shows the
need for high epitope density for recognition.
This dependence of epitope detection on presentation is however not unique to
antibody 846, but has also been observed for monoclonal antibody 3B3. The epitope
recognised by this antibody bas also been proposed as being a marker for aggrecan
synthesis in OA (Caterson et al., 1 WOa; 1 WOb; l992), but recent evidence has shown
that the reason why the 3B3 epitope was not previously detected in healthy adult
cartilage extracts by techniques such as Western bloning and immunohistochernistry,
was merely because of the concentration and presentation of the CS chains bearing
this epitope in such systems. Upon preparation of single CS chah followed by their
immobilisation onto membranes, it is possible to detect the 3B3 epitope from extracts
of normal adult cartilage by immunoblotting (Plaas et al., 1997). In contrast, we have
shown that the 846 epitope is absent from nomal adult cartilage using even this
technique. The absence of this epitope from normal adult cartilage codd be a resdt
of proteolysis of the core protein, for instance if the epitope were located near the G3
globular domain, as discussed above. This domain is removed fiom aggrecan in
mature cartilage (Paulsson et al., 1987; Vilim and Fosang, 1994). It is also possible,
however, that the epitope is never synthesised by adult chondrocytes in the h t place,
in which case, its appearance in arthritis may mise as a result of a phenotypic change
by adult chondrocytes to a more foetal form.
Difficulties in the development of technologies for the selective immunoprecipitation
and enrichment of molecules bearing the 846 epitope have made the structural
characterisation of this epitope difficult. These problems are related to the propaties
of the antibody, as aiready discussed. Therefore, a different approach was used
whereby a variety of enzymes were employed to digest a monomeric (AlDl)
preparation of foetal bovine aggrecan and thereby partially characterise the structure
of this epitope. This approach has limitations of its own due to the lack of
commercially available glycosidases and sulphatases with defined specificities and
minimum contamination fiom other enzymes. Initidy, t h e course experiments
using chondroitinase ABC and AClI wae performed to map the position of the 846
epitope dong the CS chah. These types of analyses have previously been performed
for the 3B3 and 3D5 epitopes (Hardingham et al., 1994b). It has bem shown that
chondroitinase AC11 acts preferentially as an exo-gly cosidase, sequentially releasing
disaccharides fkom the non-reducing end of CS chahs, whereas chondroitinase ABC
acts preferentially as an endo-glycosidase (Hardingham et ai., 1994b). Thus the faster
rate of loss of the epitope compared to GAG loss, upon digestion of the PG with
chondroitinase ACII, suggests that the 846 epitope is located near the free non-
reducing end of the CS chahs.
To study the role of CS chah termination Fable 3) in 846 epitope structure, PG
samples were digested with bovine P-glucuronidase to specifically remove the non-
reducing tenninal GlcA, to see if this would result in loss of epitope detection. At
enzyme concentrations where more than 65% of the non-reducing terminai GlcA had
been lost, there was no loss in the 846 epitope content. This indicated that non-
reducing terminai GlcA was not demonstrably involved in the 846 epitope structure.
Analysis of the non-reducing terminal components before and after P-gIucuronidase
treatment revealed that the only residue that was not reduced in content was the
terminai GaNAc4S, suggesting that this is a candidate for involvement in the epitope
stnictwe. The use of mammalian sources of tesminal hexosaminidases or sulphatases
would diredy address this question, by specifically removing the non-reducing
terminai GalNAc and 4-sulphate, respectively. Howeva, such enzymes are
commercidy mavailable.
Although these techniques have not given a direct solution to our question, they
suggest that the non-reducing terminal component of the 846 epitope is most likely a
GalNAc4S. This may not be so surprising since newly synthesised CS chahs from
other species have been shown to preferentially contain 4-sulphated and 4,6-
disulphated hexosamine residues at their non-reducing tenninus (Chu et al., 1985). In
humans too, GaNAc4S is the predominant terminal structure in the foetus through to
15 years of age, thereafter dropping in content with increasing age, whilst
GaNAc4,6S increases in content to about 50% in adults (Plaas et aLJ997).
However, although non-reducing terminal W A c 4 S residues are presmt in the adult
(at about 50% of the foetal lwels), it was demonstrated, using dot blots, that the 846
epitope is undetectable in the adult. One possible explanation for this apparent
paradox may be that other residues downstream fkom the non-reducing terminal
GaNAc4S are also involved in the epitope. For instance, the second hexosamine
residue downstream fiom the terminal GalNAc4S may play a role in epitope
recognition. Whereas in the foetus this residue would predominantly be another
GalNAc4S, in the adult it is more like1y to be a GalNAc6S, thus showing no
reactivity to the 846 antibody. However, it is also possible that the dot blots have
certain limitations; the increase in the relative abundance of CS chains taminating in
GaNAc4,6S in adult PG, may prevent epitope detection, by not permithg a high
enough epitope density for recognition. On the other han& the longer 846-bearing CS
chains of foetal PG (Bayliss et al., 1978; Roughley and White, 1980) may be more
flexible on the PVDF
antibody 846, making
PG.
membrane and therefore more accessible and exposed to the
the detection of the epitope easia on CS chains âom foetal
The 3B3 and 846 epitopes were initially thought to be the same because they were
both detectable in foetal and OA cartilage and not detected in noxmal adult cartilage.
Severai h e s of evidence now indicate that the 846 epitope is distinct fiom the 3B3
epitope. The proposed sequence for the 3B3 epitope consists of the residues
GlcAP 1,3GalNAc6S (Caterson et al., 1985) (Table 4). Our data suggests that GlcA at
the non-reducing terminus of CS chains is not part of the epitope reactive to the 846
antibody. Moreover, there is indirect evidence for the presence of GaINAc4S as the
terminal residue on chains bearing the 846 epitope. Previous work Eom our
laboratory ushg cornpetition ELISA has shown that binding of antibody 3B3 to foetal
bovine PG that has been immobilised onto the ELISA plates, cannot be competed by
the addition of biotinylated 846 IgM as a second antibody and vice versa (Reiner and
Poole, unpublished observations). This indicates that the two antibodies bind at
diflkent sites on the PG molecule. FinaIly, as discussed above, it has recently been
shown that 3B3 is, in fact, present on aggrecan rnolecdes f?om normal aduit cartilage
(Plaas et aZ.,1997), whereas our data show that this is not the case for the 846
epitope.
The functional significance of the changes in the sulphation patterns of CS chains still
remains to be understood One could speculate that these changes rnay innuence the
spacial arrangement of the CS ch& on the aggrecan core protein, that they could
affect the capacity of the chahs to bind to water molecules, or that they could affect
the ability of the CS chains to interact with other extracellular ma& molecules. The
functional relevance of the distinct chah termination sequenca on CS chains h m
cartilages of clifferait ages or pathologies is also not known. They may merely reflect
differences in the sulphotransferase and glycosyltramfesase activities in these tissues,
as has beai proposed by Plaas et al. (1997). In this respect, the 846 epitope may thus
be a fortuitous marker of a phenotypic change in the cartilage during disease,
uidicating a reparative process. Whether the 846 epitope plays a fundional role in
cartilage repair or is rnaely an indicator of such repair, its analysis in tissue and body
fluids does provide an indication of disease presence, and changes in its level will
indicate that a change in the metabolic status of the tissue has occured, albeit one that
is difficult to intqret with respect to the processes of synthesis and degradation
without the use of additional markers. Thus antibody 846 represents one tool by
which disease progression and treatment may be monitored in &tic patients. Since
the epitope is also found in human chondrosarcomas, its measment in body fluids
may be a valuable indicator of disease presence.
5. CONCLUSIONS
The experiments described in this thesis have shown that the 846 epitope is present
on newly synthesised aggrecan molecules which show a preferential retention within
the extracellular matrix synthesised by foetal chondrocytes in culture. The molecules
in the matrix contain a higher epitope density compared to those released into culture
medium, suggesting a role for these molecules in matrix formation and repair such as
that which occurs in OA. This conclusion is finther supported by observations fkom
the normal adult cartüage cultures treated with -sin, which indicate a higher
epitope density on the aggrecan molecules generated as part of the repais process. The
media fkom the chondrocyte cultures and cultures of OA cartilage show that the
amount of 846 epitope released fkom the tissue is reflective of the release of a
proportion of newly synthesised aggrecan moledes, and is directly conelated with
aggrecan synthesis. Analysis of synovial fluid and cartilage fkom OA patients
indicates that the content of the epitope in the cartilage is reflected in the release of
the epitope-containing molecules into the synovial fluid. While this supports the use
of 846 as a marker of aggrecan synthesis, the use of other biochemical markers of
cartilage turnover, such as a marker for aggrecan degradation, is required in addition,
in order to fully appreciate changes in aggrecan metabolism occuring in arthritis.
Structural analyses of the 846 epitope suggest that it may be present on the non-
reducing taminal ends of the CS chains of aii sizes and involves a terminal
GaNAc4S and probably adjacent residues. The analyses also reveal the requirement
for a hi& epitope density for the detection of this epitope, and iiiustrate that lack of
epitope detection does not necessdy mean lack of i ts presence.
6. STATEMENT OF ORIGINAL CONTRIBUTIONS
These studies are the h t to demonstrate through in viho experiments on aggrecan
biosynthesis, that
the 846 epitope is pnsent on newly synthesised aggrecan molecules.
the 846 epitope-bearing molecules are preferentidy retained in the matrix.
the 846 epitope may be involved in the repair process of cartilage.
the release of the 846 epitope-bearing molecules from OA cartilage may be
indicative of the release of a proportion of newly synthesised aggrecan molecules
bearing the epitope.
the release of the 8 46 epitope-bearing molecules is correlated with the synthesis of
aggrecan,
In addition, these studies are also the first to show that
the 846 epitope content of the synovial fluid of OA joints may be reflective of the
epitope content of the cardage fiom that joint.
the 846 epitope is not present on CS chains of aggrecan molecules f?om adult
cartilage.
the 846 epitope may be present at the non-reducing taminal end of CS chains.
the structure of the 846 epitope may involve a non-reducing terminal GalNAc4S.
the 846 epitope is present on CS chains of ail sizes.
the 846 epitope can be detected on single CS c h a b if hi& concentrations of the
epitope-bearing chains are applied to PVDF membranes.
7. PUBLICATIONS
The following abstracts and manuscripts were derived fiom the work presented in this
thesis :
Abstracts:
1. H. Jugessw, P.J. Roughley and A l . Poole.
The Chondroitin Sulphate epitope 846 of aggrecan: Evidence for its synthesis in
tissue repair and its partial characterisation.
The Canadian Connective Tissue Confaence, Kingston, ON, 1997.
2. H. Jugessur and A.R. Poole.
The Chondroitin Sulphate epitope 846 is a putative rnarker for Aggrecan
Biosynthesis.
The Canadian Connective Tissue Conference, Toronto, ON, 1996.
The XVth Meeting of the Federation of the European Connective Tissue Societies,
Muaich, Germany, 1996.
3. H. Jugesur and A.R. Poole.
Aggrecan Turnover. The Chondroitin Sulphate epitope 846 as a putative marker for
aggrecan biosynthesis.
The Inaugural Confêrace of the Canadian Musculoskeletal (Connective Tissue)
Society, Montreal, QC, 1995.
Manuscripts:
1. H. Jugessur, P.J. Roughley, A.H.K. Plam, M. Tanzer, D. Zucker and A.R. Poole.
The Chondroitin Sulphate 846 epitope of aggrecan. Its relationship to aggrecan
synthesis and its partial charactexisation. (In preparation).
2. L.S. Lohmander, M. lunesnr, H. Jugessur and A.R. Poole.
Changes in Joint Cardage Aggrecan metabolism after knee injury and Osteoarthritis.
(In press, Arthritis & Rheumatism).
3. N. Ishiguro, T. Ito, H. Ito, H. Iwutu, M. Ionem, H. Jugessur and A R . Poole.
Relationships of Matrix Metdoproteinases and their Inhibitors to Cartilage
Roteoglycan and Collagen Turnover reveded by Analyses of Synovial Fluids from
Patients with Osteoarthntis. (submitted to Arhritis & Rheumatism).
4. X: Chevalier, P. Claudepierre, A&. Poole, B. Puscal, M. lonescu, H Jugessur,
J. Rymer and J. Piette.
Evidence for changes in Cartilage Matrix Turnover in Relapshg Polychondntis.
(submitted to Arthritis & Rheumatism).
8. REFERENCES
h m , E.C., Decicco, C.P., Chwey, R and Tortorella, MD. 1997. Cleavage of
Native Cartilage Aggrecan by Neutrophil Collagenase (MMP-8) 1s Distinct fiom
Endogenous CIeavage by Aggrecanase. J. Biol. Chem. 272: 92949299.
M o , A., Stamenkovic, L, Melnick, M., Underbill, C.B. and Seed, B. 1990. CD44
is the principal ceU surface receptor for hyaluronate. Ce22 6 1 : 1303- 13 13.
Aviezer, D., Hecht, D., Safhn, M., Eisinger, M., David, G. and Yayon, A. 1994.