Rheumatology 1999;38:1108–1112
Joint immobilization reduces synovial fluid
hyaluronan concentration and is accompanied by
changes in the synovial intimal cell populations
A. A. Pitsillides, T. M. Skerry1 and J. C. W. Edwards2
Department of Veterinary Basic Sciences, The Royal Veterinary College, London
NW1 0TU, 1Department of Biology, University of York, York YO1 5YW and
2Rheumatology Unit, University College London, London W1P 9PG, UK
Abstract
Objectives. Synovial fluid (SF ) of normal joints contains high hyaluronan (HA)
concentrations. However, the mechanism by which these are controlled and how they are
influenced by articulation and loading are not established. In this study, we have examined
whether immobilization influences SF HA concentration and whether this is associated with
alterations in the synovial lining.
Methods. Hock joints of five adult sheep were immobilized by external fixation. Twelve
weeks later, SF and synovium samples were collected. The HA concentration in SF was
assayed using an ELISA-based method. Non-specific esterase (NSE ) and uridine
diphosphoglucose dehydrogenase ( UDPGD) activities were assessed in cryostat sections of
snap-chilled synovial samples using cytochemical techniques, and UDPGD activity per cell
was measured in synovial lining cells by scanning and integrating microdensitometry.
Results. We found that the SF HA concentration was decreased from 1.65 ± 0.25 mg/ml in
control joints to 0.68 ± 0.16 mg/ml in immobilized joints. Synovial intimal cell UDPGD
activity decreased from 18.0 ± 2.7 U/cell in control joints to 12.2 ± 1.5 after immobilization.
There was also a decrease in UDPGD-positive intimal cell numbers. Intimal surfaces in
controls contained numerous NSE-positive cells, which were rarely observed in intima from
immobilized joints.
Conclusions. These results suggest that immobilization decreases SF HA levels and that this
is associated with reduced intimal cell UDPGD activity (essential for HA formation).
Immobilization also decreased the prevalence of (NSE-positive) intimal macrophages. These
findings suggest that mechanosensitive homeostatic mechanisms exist within the synovial
intima.
K : Synovial lining cells, Hyaluronan, Immobilization, Joint, Uridine diphosphoglucose
dehydrogenase, Type B cells, Synovial fluid.
The view that connective tissue structure adapts dynamically to its prevailing mechanical milieu is supported
by many studies. Thus, a diarthrodial joint’s capacity
to provide mobility without damage or failure may also
be achieved, at least partly, by such functionally related
mechanisms. The response of joints to changes in loading, and the mechanisms by which immobilization may
alter the synovial envelope, are not understood. In this
study, we examine whether immobilization-mediated
removal of such functional stimuli influences synovial
intimal structure and function, as indicated by synovial
fluid (SF ) hyaluronan (HA) concentration and intimal
cell number and phenotype.
Submitted 16 February 1999; revised version accepted 14 May 1999.
Correspondence to: A. A. Pitsillides, Department of Veterinary
Basic Sciences, The Royal Veterinary College, Royal College Street,
London NW1 0TU, UK.
The SF of normal diarthrodial joints contains high
concentrations of HA [1, 2], suggesting that high levels
are required for joint function to be preserved.
Experimental findings support the notion that these high
HA concentrations may contribute to the joint’s functional competence. For example, the functional improvement observed in rheumatoid arthritis (RA) joints
(where HA concentrations are low whilst the total
amount of HA may be high; [2–3]) after intra-articular
corticosteroid treatment is associated with a restoration
in these SF HA concentrations [4]. Nevertheless, the
mechanisms by which joint HA concentrations are controlled, and how they are influenced by articulation and
loading, are not fully elucidated.
Whilst the changes in SF HA content in disease are
well established, their relationship with the cellular
changes occurring in the intimal lining of synovium
remains largely unaddressed [5]. Synovial intima con-
1108
© 1999 British Society for Rheumatology
Joint immobilization and SF hyaluronan concentration
tains two major cell populations: macrophages (type A
cell; [6, 7]) and specialized fibroblasts (type B cell ). One
specialized feature of intimal fibroblasts relates to HA
synthesis. We have shown that uridine diphosphoglucose
dehydrogenase ( UDPGD) activity, which is essential for
HA production and rate limiting in UDP-glucuronate
(one of HA’s monosaccharide subunits) formation, is
6-fold higher in normal synovial intimal fibroblasts than
other fibroblasts, and that this decreases in chronically
inflamed rabbit and human RA joints [8–10]. Cells with
high UDPGD activity also appear in tissue lining a
cavity containing a moving implant, but not in an airfilled cavity [5], suggesting that mechanical stimuli control cellular HA-synthetic characteristics.
Synovial intima appears to have a low rate of cell
replacement and, although direct evidence is lacking,
intimal fibroblasts are probably derived locally from
subintimal fibroblasts, whilst intimal macrophages are
likely to arrive in the joint as circulating monocytes,
undergoing their macrophagic transition in the subintimal perivascular stroma. However, the effect of
normal joint function on the local differentiation of
these apparently distinct cell types has not been established. The aim of this study was to determine the effect
of an extended period of external fixator-induced
immobilization on SF HA levels and to establish whether
immobilization is also associated with alterations in the
cellular characteristics of the synovial intima.
Methods
Joint immobilization
Five adult female Welsh mountain sheep were acclimatized to laboratory housing for 2 weeks. Thereafter,
sheep were anaesthetized by i.v. thiopentane (20 mg/kg),
intubated and maintained using 2.5% halothane in 60%
oxygen/40% nitrous oxide. In each animal, the medial
aspect of the left hind leg was exposed. Six skin incisions
were made and three 4.5-mm-diameter Schantz screws
inserted into the metatarsal and three into the tibia in a
medio-lateral direction. As previously described [11], a
customized angular fixator bar was clamped to the
screws and wounds dressed with iodine-soaked swabs.
All the sheep were then kept in laboratory pens for
12 weeks.
In vivo strain gauge recordings in other sheep showed
that in addition to reducing the peak strain magnitude
in the calcaneus during walking, the fixator also abolished the high-magnitude transient strains experienced
when the sheep made sudden movements. Thus, application of the fixator protects the hock joint from loading
and also prevented hock joint articulation [11].
Collection of SF and lining samples
Twelve weeks after surgery, the sheep were killed by
lethal pentobarbitone injection and synovial cavities of
left and right hock joints were sampled by needle
insertion between the digital extensor tendons on the
cranial aspect. After collection of SF, the skin was
removed and the tibiotarsal joint disarticulated. Samples
1109
of synovial tissue were collected from equivalent sites in
the anterior and posterior compartments of all joints.
The synovial tissue was cut into pieces not greater than
0.5 cm3, immersed for 1 min in a 5% solution of polyvinyl alcohol containing 5% CaCl ·2H O [8] and then
2
2
chilled to −70°C in n-hexane (BDH; low in aromatic
hydrocarbons). The tissue was stored in dry tubes at
−70°C and examined within 2 weeks.
Assessment of HA concentration by ELISA
Biotinylated HA binding region (HABr) prepared from
pig laryngeal cartilage was a kind gift from Professor
M. T. Bayliss ( The Royal Veterinary College, London).
Aliquots (0.5 ml ) of SF were incubated with 0.5 ml of
20 ml/ml papain (Boehringer Mannheim) in 0.2  acetate
buffer (pH 5.8) containing 10 m EDTA and 10 m
cysteine (added just prior to use), at 60°C for 16 h. The
papain was inactivated by addition of 5 m iodoacetic
acid. Experiments in which HA of different, precisely
defined, molecular weights (a gift from Pharmacia LKB,
Sweden) were treated as above indicated that assay
sensitivity was unaffected by HA’s molecular mass (not
shown).
After centrifugation, HA concentration in the supernatant was assayed, in duplicate, using an ELISA-based
method [12]. Each sample was assayed at least twice,
and samples which did not yield a dilution profile
gradient corresponding to that obtained for standard
HA were discarded and the assay repeated (no sample
was excluded more than once). This method relies on
competition between unknown concentrations of sample
HA and a fixed quantity of plate-bound HA, for binding
to a fixed concentration of biotinylated HABr. The
biotinylated HABr’s capacity to bind plate-bound HA
is inversely related to the HA concentration of that
sample. Sample HA concentration (mg/ml ) was determined by comparison with the competition produced by
known concentrations of exogenous HA, and results
were evaluated by paired t-test.
Assessment of UDPGD activity
To assay UDPGD activity, we used the method
described by Mehdizadeh et al. [13]. Briefly, fresh 10 mm
sections were incubated at 37°C in an atmosphere of
nitrogen, in medium containing 5.3 m UDP-glucose
(3 mg/ml; Sigma), 0.45 m NAD (0.3 mg/ml; Boehringer) in 30% (w/v) polyvinyl alcohol (Grade G04/140;
Wacker Chemicals Ltd, Surrey) in 0.05  glycylglycine
buffer (pH 7.8). Just prior to use, medium was saturated
with nitrogen, adjusted to pH 7.8 and 3.7 m nitroblue
tetrazolium (3 mg/ml; Sigma) was added. In all cases,
reaction time was 20 min. Serial sections were incubated
in medium lacking substrate (no UDP-glucose) or
in full medium including 0.2 m UDP-xylose
(0.21 mg/ml; Sigma), a specific inhibitor of UDPGD
activity.
After reaction, sections were washed in water, dried
and mounted in Aquamount (BDH ). The amount of
precipitated formazan was measured using a Vickers
M85A scanning and integrating microdensitometer [8],
A. A. Pitsillides et al.
1110
at a wavelength of 560 nm, in individual cells within the
synovial lining, with a ×40 objective and a scanning
spot of 0.5 mm in the plane of the section. At least 20
cells were measured in histologically defined intima in
each of two duplicate sections in each assay, and up to
three samples were examined from each joint. Results,
expressed as units of mean integrated extinction (MIE
×100) per cell, for a 20 min reaction time were evaluated
by paired t-test.
Assessment of non-specific esterase (NSE) activity
NSE activity was assessed in fresh 10 mm serial sections
using a-naphthyl acetate as described previously [10].
In all instances, serial sections were also stained with
0.1% toluidine blue in 0.1  acetate buffer (pH 6.1),
washed, air-dried and mounted in DePeX (BDH ).
Results
Effect of immobilization on SF HA concentration
HA concentration in the SF of each immobilized joint
was decreased when compared to contralateral control
joints ( Fig. 1A). Such decreases in SF HA concentration, from 1.65 ± 0.25 mg/ml (mean ± ...) in control
joints to 0.68 ± 0.16 mg/ml in immobilized joints,
Control
Immobilized
Control
Immobilized
Control
Immobilized
indicate a statistically significant (P < 0.04) 59 ± 13%
decrease in HA concentration (Fig. 1B).
Effect of immobilization on UDPGD activity in synovial
lining cells
UDPGD activity/cell in the synovial intima of immobilized joints was decreased when compared to control
joints from the same animal (Fig. 1C ). Immobilization produced a small (29.6 ± 6.3%) but significant
(P < 0.03) reduction in UDPGD activity/cell in the
synovial intima: from 18.0 ± 2.7 U/cell (mean ± ...)
in control joints to 12.2 ± 1.5 U/cell in immobilized
joints (Fig. 1D).
In accord with previous findings, the majority of
intimal cells in control synovium showed high UDPGD
activity, whilst subintimal synovial cells showed little, if
any, UDPGD activity (Fig. 2a and c). Intimal cells with
high UDPGD activity formed an almost continuous
band of cells which followed the contours of the intimal
layer ( Fig. 2c). In immobilized joints, the thickness of
the UDPGD-containing intimal layer was frequently
reduced; in control joints, it was often 2–3 cells thick,
whilst in immobilized joints it was rarely more than one
cell thick ( Fig. 2b and d).
In addition to decreases in UDPGD activity/cell,
synovium from immobilized joints showed less uniform
intimal cell UDPGD activity, with only a minority of
these cells showing high UDPGD activity and large
areas devoid of UDPGD-positive cells ( Fig. 2g and h).
Synovial lining cells in sections of synovium from both
control ( Fig. 2e and f ) and immobilized ( Fig. 2g–i)
joints showed negligible levels of activity when reacted
in medium lacking substrate (minus UDP-glucose).
Effect of immobilization on NSE activity in synovial
lining cells
It has previously been found that macrophages in synovial intima show consistently high NSE activity [10].
The intimal lining of synovium from control joints
contained distinct regions in which a significant proportion of cells were NSE positive (Fig. 2j). Commonly,
these NSE-positive cells were superficial to cells in the
same region showing high UDPGD activity. In contrast,
regions of NSE-positive rich intima were not evident in
synovium from immobilized joints, and single NSEpositive cells were only rarely observed ( Fig. 2k).
Discussion
Control
Immobilized
F. 1. (A) and (B) Alterations in hyaluronan concentration
(mg/ml ) in synovial fluid of immobilized joints [(A) individual
paired joints; (B) mean changes, mean ± ...; **denotes
statistical significance at P < 0.04]. (C ) and (D) Alterations
in UDPGD activity (MIE × 100/cell ) in synovial lining cells
in immobilized joints [(C ) paired joints; (D) mean changes,
mean ± ...; **denotes statistical significance at P < 0.03].
Our findings indicate that removing a joint’s normal
functional stimuli, by external fixator-mediated immobilization, results in decreased SF HA concentration. This
is accompanied by decreases in intimal cell UDPGD
activity, and by decreases in the number of both
UDPGD-positive cells and the prevalence of NSEpositive macrophages within synovial lining. These cellular changes and reduced HA levels may not necessarily
be related, but their association raises several questions
regarding the role of movement in maintaining joint
function.
Joint immobilization and SF hyaluronan concentration
F. 2. (a–d) Sections of synovial lining from control (a and
d) and immobilized (b and c) joints reacted for UDPGD
activity (in all cases incubation time = 20 min). (e–i) Substrate
specificity for UDPGD activity; sections of synovial lining
from control (e and f ) and immobilized (g–i) joints stained
with 0.1% toluidine blue (g), and either reacted for UDPGD
activity in full medium (containing substrate; e and h) or in
medium lacking substrate (control; f and i). ( j and k) Sections
of synovial lining from control ( j) and immobilized (k) joints
reacted for non-specific esterase (NSE) activity and briefly
counterstained with 0.1% toluidine blue. Arrowheads indicate
the position of NSE-positive cells.
The immobilization-induced reduction in SF HA concentration could be contributed to by decreased synthesis, increased clearance or an increase in volume of
distribution. Decreases in intimal cell UDPGD activity
indicate that immobilization may reduce the formation
of one of HA’s essential monosaccharides, and until the
mechanisms regulating HA synthesis are determined, it
is possible that this alone may decrease rates of HA
synthesis. Since the volume of SF retained in a normal,
mobile joint is probably determined by its HA content,
a reduced rate of HA synthesis might be expected to
give rise to a reduced SF volume with an unchanged
1111
concentration of HA. However, immobility may also be
expected to reduce SF clearance and thereby increase
the volume of distribution directly [14]. Although the
precise estimation of SF volume was not possible in this
study, no major change was apparent in immobilized
joints. Thus, immobilization may have a dual action,
such that the concentration of HA falls and SF volume
is not increased because any tendency to retain fluid is
offset by the diminished water-retaining capacity of a
reduced intra-articular HA content.
It is possible that the immobilization-induced changes
observed were mediated by alterations in fluid flux
consequent on muscle wastage and change in blood
flow. However, the level of use of the limb was not
significantly reduced and there was neither significant
muscle wastage detectable by palpation nor oedema
formation observed. Although our study has not
addressed the role of movement in controlling HA’s
molecular mass, the immobilization-induced decreases
in HA’s concentration are consistent with an adaptive
change to the removal of the joint’s normal functionally
related stimuli, suggesting also that high SF HA levels
are essential in this regard.
The mechanisms by which intimal cell numbers are
regulated in the synovium of normal joints and how
they are modified by inflammation, in which intimal cell
number increases significantly, are ill-defined. As well
as decreasing UDPGD activity, we have found that
immobilization reduces the number of both UDPGDpositive cells and NSE-positive macrophages in the
intima, suggesting that their recruitment is modulated
by mobility. Movement may influence intimal cell numbers by modifying local cytokine release, may involve
nerve endings or even a paracrine influence from resident
intimal fibroblast-like cells; the latter may explain why
macrophages preferentially accumulate in the intima,
rather than subintima, but this remains to be addressed.
Intimal macrophage function in normal synovium is
not established. Several lines of evidence suggest that
these cells are employed in remodelling their extracellular matrix [15]. If this is the case, it is tempting to speculate that fibroblast-derived HA synthesis is decreased
F. 3. Diagram summarizing a model by which HA concentrations in SF and synovial intimal cell populations might be
influenced by functionally related mechanical stimuli.
1112
A. A. Pitsillides et al.
primarily by movement’s withdrawal and that this has
secondary effects on macrophage recruitment (see
Fig. 3). By corollary, findings based on the analysis of
SF after episodes of exercise are consistent with movement increasing local degradation [16, 17]. Further,
Culty et al. [18] showed that alveolar macrophages are
capable of uptake and lysosome-mediated degradation
of HA, suggesting that decreases in macrophage numbers would point to a diminished HA-degradatory
potential in the intima. This would be in harmony with
coincident decreases in HA synthesis, and the possibility
that changes in local (fibroblast-derived ) HA concentration modify macrophage recruitment.
In conclusion, these findings show that immobilization
results in decreases in SF HA concentration. They also
highlight the possibility that changes in the synovial
intima may reflect this alteration, or perhaps that
fibroblast- and macrophage-like synovial lining cells
directly coordinate and control SF HA content by some
mechanodependent mechanism.
Acknowledgements
We wish to thank Prof. Mike T. Bayliss of the
Department of Veterinary Basic Sciences, The Royal
Veterinary College, London, for the provision of
biotinylated hyaluronan-binding region and Ms Angie
Poole for her help with illustrations. This work was
supported by the Arthritis Research Campaign, UK.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
References
1. Ogston AG, Stanier JE. The physiological functions of
hyaluronic acid in synovial fluid; viscous, elastic and
lubricant properties. J Physiol (London) 1953;199:244–52.
2. Balazs EA, Watson D, Duff IF, Roseman S. Hyaluronic
acid in synovial fluid: 1. Molecular parameters of hyaluronic acid in normal and arthritic human fluids. Arthritis
Rheum 1967;10:357–75.
3. Sundbland L. Studies on hyaluronic acid in synovial fluid.
Acta Soc Med Ups 1953;58:113–9.
4. Pitsillides AA, Will RK, Bayliss MT, Edwards JCW.
Circulating and synovial fluid hyaluronan levels: effects of
intra-articular corticosteroid on the concentration and the
rate of turnover. Arthritis Rheum 1993;37:1030–8.
5. Pitsillides AA. Role of hyaluronan in joint cavitation. In:
Archer CW, Caterson B, Benjamin M, Ralphs J, eds.
16.
17.
18.
Biology of the synovial joint. Singapore: Harwood
Academic Press, 1999.
Barland P, Novikoff AB, Hamerman D. Electron microscopy of the human synovial membrane. J Cell Biol
1962;14:207–16.
Dreher R. Origins of synovial type A cells during inflammation. Immunobiology 1982;161:232.
Pitsillides AA, Blake SM. Uridine diphosphoglucose dehydrogenase activity in synovial lining cells in experimental
antigen induced model of rheumatoid arthritis: an indication of synovial lining cell function. Ann Rheum Dis
1992;51:992–5.
Pitsillides AA, Wilkinson LS, Mehdizadeh S, Bayliss MT,
Edwards JCW. Uridine diphosphoglucose dehydrogenase
activity in normal and rheumatoid synovium: the description of a specialized synovial lining cell. Int J Exp Path
1993;74:27–34.
Wilkinson LS, Pitsillides AA, Worrall JG, Edwards JCW.
Light microscopic characterisation of the fibroblast-like
synovial intimal cell (synoviocyte). Arthritis Rheum
1992;35:1179–84.
Skerry TM, Lanyon LE. Short duration walking exercise
does not prevent disuse bone loss in the functionally
protected sheep calcaneus. Bone 1997;16:269–75.
Fosang AJ, Hey NJ, Carney SL, Hardingham TE. An
ELISA plate-based assay for hyaluronan using biotinylated proteoglycan G1 domain. Matrix 1990;10:306–10.
Mehdizadeh S, Bitensky L, Chayen J. The assay of uridine
diphosphoglucose dehydrogenase activity. Cell Biochem
Funct 1991;9:103–10.
Levick JR, Thompson PW. Intra-articular volume as an
important factor governing macro-molecular half-life in
synovial fluid. Ann Rheum Dis 1988;47:701–2.
Edwards JCW, Wilkinson LS, Jones HM, Soothill P,
Henderson KJ, Worrall JG et al. The formation of human
synovial joint cavities: a possible role for hyaluronan
and CD44 in altered interzone cohesion. J Anat 1994;
185:355–67.
Roos H, Dahlberg L, Hoerrner LA, Lark MW, Thonar
EJ-MA, Shinmei M et al. Markers of cartilage matrix
metabolism in human joint fluid and serum: The effect of
exercise. Osteoarthritis Cartil 1995;31:7–14.
Grootveld M, Henderson EB, Farrell A, Blake DR, Parkes
HG, Haycock P. Oxidative damage to hyaluronate and
glucose in synovial fluid during exercise of the inflamed
rheumatoid joint. Detection of abnormal low-molecularmass metabolites by proton-n.m.r. spectroscopy. Biochem
J 1991;273:459–67.
Culty M, Nguyen HA, Underhill CB. The hyaluronan
receptor (CD44) participates in the uptake and degradation of hyaluronan. J Cell Biol 1992;116:1055–62.