Osteoarthritis and Cartilage 19 (2011) 1356e1362
The effect of molecular weight on hyaluronan’s cartilage boundary lubricating
ability e alone and in combination with proteoglycan 4
J.J. Kwiecinski y a, S.G. Dorosz y a, T.E. Ludwig z, S. Abubacker z, M.K. Cowman x, T.A. Schmidt yzk *
y Schulich School of Engineering: Centre for Bioengineering Research & Education, University of Calgary, Calgary, AB, Canada
z Biomedical Engineering, University of Calgary, Calgary, AB, Canada
x Polytechnic Institute of New York University, Brooklyn, NY, USA
k Faculty of Kinesiology, University of Calgary, Calgary, AB, Canada
a r t i c l e i n f o
s u m m a r y
Article history:
Received 11 April 2011
Accepted 29 July 2011
Objectives: (1) assess the molecular weight dependence of hyaluronan’s (HA) cartilage boundary lubricating ability, alone and in combination with proteoglycan 4 (PRG4), at physiological concentrations; (2)
determine if HA and PRG4 interact in solution via electrophoretic mobility shift assay (EMSA).
Methods: The cartilage boundary lubricating ability of a broad range of MW HA (20 kDa, 132 kDa,
780 kDa, 1.5 MDa, and 5 MDa) at 3.33 mg/ml, both alone and in combination with PRG4 at 450 mg/ml,
was assessed using a previously described cartilage-on-cartilage friction test. Static, mstatic, Neq, and
kinetic, <mkinetic, Neq>, were calculated. An EMSA was conducted with PRG4 and monodisperse 150 kDa
and 1,000 kDa HA.
Results: Friction coefficients were reduced by HA, in a MW-dependent manner. Values of <mkinetic, Neq> in
20 kDa HA, 0.098 (0.089, 0.108), were significantly greater compared to both 780 kDa, 0.080 (0.072,
0.088), and 5 MDa, 0.079 (0.070, 0.089). Linear regression showed a significant correlation between both
mstatic, Neq and <mkinetic, Neq>, and log HA MW. Friction coefficients were also reduced by PRG4, and with
subsequent addition of HA; however the synergistic effect was not dependent on HA MW. Values of
<mkinetic, Neq> in PRG4, 0.080 (0.047, 0.113), were significantly greater than values of PRG4 þ various MW
HA (similar in value, averaging 0.040 (0.033, 0.047)). EMSA indicated that migration of 150 kDa and
1,000 kDa HA was retarded when combined with PRG4 at high PRG4:HA ratios.
Conclusions: These results suggest alterations in HA MW could significantly affect synovial fluid’s cartilage boundary lubricating ability, yet this diminishment in function could be circumvented by physiological levels of PRG4 forming a complex, potentially in solution, with HA.
Ó 2011 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.
Keywords:
Cartilage boundary lubrication
Hyaluronan
Proteoglycan 4 (PRG4)
Introduction
Normal mechanical function of synovial joints is largely
dependent upon articular cartilage, synovial fluid (SF), and their
interaction. Specifically, articular cartilage serves as the
low-friction, wear resistant, load bearing tissue at the end of long
bones in synovial joints1. SF, and its constituents, serves to lubricate
the cartilage surface and function as a shock absorber2,3 as well as
provide nutrients to the cells and tissues within synovial joints4. As
a boundary lubricant, boundary mode lubrication being defined by
surface-to-surface contact of articular cartilage with frictional
*Address correspondence and reprint requests to: T.A. Schmidt, Faculty of
Kinesiology, 2500 University Dr NW, KNB 2232, University of Calgary, Calgary,
AB, Canada T2N 1N4. Tel: 1-403-220-7028; Fax: 1-403-284-3553.
E-mail address: [email protected] (T.A. Schmidt).
a
Authors contributed equally to the study.
characteristics dominated by surface molecules1, SF has been
shown to reduce friction in vitro at a cartilageecartilage interface to
extremely low levels5.
Hyaluronan (HA), a polymer of disaccharides composed of
D-glucuronic acid and N-acetyl D-glucosamine, is a primary constituent of SF6. In normal synovial joints, it is present in concentrations
from 1 to 4 mg/ml6,7, and its molecular weight (MW) ranges
continuously from 4 kDa to approximately 8 MDa8,9. High MW HA,
w800 kDa (SupArtz, Seikagaku), has been shown to function effectively as a cartilage boundary lubricant in a dose-dependant manner,
decreasing friction at a cartilageecartilage interface10. Similarly, high
MW of HA is thought to provide the viscoelastic component to SF11,
which may affect the cartilage lubricating ability of normal SF in
a fluid film mode of lubrication. It has been shown that injured/
diseased SF has reduced concentrations of HA w0.1e1.3 mg/ml in
diseased SF compared to 1e4 mg/ml in normal SF6,7,12. Some
1063-4584/$ e see front matter Ó 2011 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.joca.2011.07.019
J.J. Kwiecinski et al. / Osteoarthritis and Cartilage 19 (2011) 1356e1362
osteoarthritis (OA) SF has also been characterized by a decrease in
the proportion of high MW HA13,14. Because of these changes in HA
concentration and MW, the viscoelasticity of SF (and it’s ability to
protect the joint) can be reduced15e17. However, it remains unclear if
MW affects the ability of HA to interact with the articular surface of
cartilage, as well as other SF lubricant constituents, and function as
boundary lubricant at a physiological concentration.
In addition to HA, proteoglycan 4 (PRG418, a mucin-like glycoprotein also known as lubricin19 and superficial protein20), is
another primary boundary lubricating constituent of SF10,21,22.
PRG4 has also been shown to contribute to the boundary lubrication of articular cartilage in a dose-dependent manner, and act
synergistically with HA at physiological concentrations to lower
friction to levels near that of SF10. Specifically, combining high MW
SupArtz HA at 3.33 mg/ml with PRG4 at 450 mg/ml in solution
significantly reduced kinetic friction in a boundary mode of lubrication at a cartilageecartilage biointerface, as compared to that in
phosphate buffered saline (PBS), and both HA or PRG4 alone at the
same respective concentrations10. PRG4 and HA were first reported
to act synergistically at a latexeglass interface under boundary
lubricating conditions, with HA enabling PRG4 to lubricate at
higher contact pressures23. It is currently unknown if HA’s MW
affects its interaction with PRG4, either in solution or at the surface
of articular cartilage, and ability to synergistically reduce friction.
The molecular basis for a HA-PRG4 interaction remains to be
fully elucidated. PRG4 has been shown to bind to HA coated plastic
in a preliminary study24. Conversely, HA was reported to not bind to
other model surfaces coated in PRG425, but a PRG4eHA solution did
adsorb to these surfaces to a greater extent than PRG4 alone. These
studies support the notion that molecular interactions at surfaces
can be different in solution. A recent study using surface forces
apparatus with cartilage suggested the HA-PRG4 complex functions
as effective boundary lubricant by becoming trapped at the cartilage surface under compression26. Conversely, no HA-PRG4 synergistic effect on friction behavior was observed between hydrophilic
and hydrophobic model surfaces studied using colloidal force
microscopy25. While indirect biophysical evidence for a functional
PRG4eHA interaction in SF has been reported27, it remains to be
clearly demonstrated if a PRG4eHA complex can form in solution.
Electrophoretic mobility shift assay (EMSA) is the core technology underlying a wide range of analyses for the characterization
of interacting systems28. EMSA is a rapid and sensitive method
generally used to detect proteinenucleic acid interactions, and is
based on observations that the electrophoretic mobility of a proteinenucleic acid complex is typically less than that of the free
nucleic acid. EMSA has also been used to detect proteineprotein
and proteineoligosaccharide interactions, both with and without
prelabeling of proteins (reviewed in28), demonstrating its versatility. As such, EMSA is a simple, ideal, and previously unused
method to study the interactions of well-defined preparations of
PRG4 with HA in solution.
The objectives of this study were therefore to (1) assess the MW
dependence of HA’s cartilage boundary lubricating ability, alone
and in combination with PRG4, at physiological concentrations, and
(2) determine if HA and PRG4 interact in solution via EMSA.
1357
described previously29. The bicinchoninic acid (BCA) assay kit was
obtained from Thermo Firsher Scientific (Rockford, IL, USA).
Osteochondral samples were prepared for lubrication testing
from fresh, skeletally mature bovine stifle joints, obtained from
a local abattoir (Calgary, AB, Canada). Fresh bovine SF and materials
for PRG4 preparation and lubrication testing were obtained as
described previously10. In addition, 20 kDa, 132 kDa, 780 kDa, and
1.5 MDa sodium hyaluronate (HA) was obtained from Lifecore
Biomedical (Chaska, MN, USA), 5 MDa (Healon GV) from Abbott
Medical Optics (Abbott Park, IL, USA). Monodisperse 150 kDa and
1,000 kDa HA, and Mega, Hi and LoLadder HA, were obtained from
Hyalose, L.L.C. (Oklahoma City, OK, USA).
Lubricant preparation and characterization
HA
Five HA solutions of differing MWs (20 kDa, 132 kDa, 780 kDa,
1.5 MDa, and 5 MDa) were prepared in PBS at a physiological
concentration of 3.33 mg/ml6. Solutions were stored at 20 C. MW
of the HA solutions was qualitatively confirmed by 1% AGE under
non-reducing conditions (Fig. 1), as described previously30. Briefly,
sample (2.5 mg) and Mega, Hi and LoLadder HA (1 mg) were
prepared to a final concentration of 0.005% bromophenol blue and
333 mM sucrose in TriseAcetateeEDTA (TAE) buffer, then subject to
1% AGE in TAE buffer for 3 h at 50 V. The gel was then stained with
Stains-All and destained with 10% ethanol. Migration of the bands
was assessed by densitometric scanning of the gel followed by
analysis with Image J (NIH, Bethesda, MD, USA).
PRG4
PRG4 was prepared from culture media conditioned by mature
bovine cartilage explants, essentially as described previously10. The
purity of the purified solution was confirmed by 3e8% TriseAcetate
SDS-PAGE followed by protein stain and Western blotting with
anti-PRG4 Ab LPN29 using Invitrogen’s NuPAGE system (data not
shown). The concentration was then determined by BCA assay, which
was unaffected by the glycosylations on PRG4 (Supplementary Data).
Boundary lubrication tests
Sample preparation
Fresh osteochondral samples (n ¼ 20) were prepared for friction
testing from the patellofemoral groove of eight skeletally mature
bovine stifle joints, as described previously5,10. Note that samples
were first rinsed vigorously overnight in w35 ml of PBS at 4 C to rid
the articular surface of residual SF (confirmed by lubrication
testing5,10) prior to lubrication testing in PBS. Samples were then
bathed in w0.3 ml of the subsequent test lubricants (core bathed in
w0.2 ml, annulus bathed in w0.1 ml), completely immersing the
cartilage, at 4 C overnight prior to lubrication testing.
Methods
Materials
The majority of materials for agarose gel electrophoresis (AGE)
were obtained as described previously29. In addition, sucrose and
Stains-All were obtained from SigmaeAldrich (St. Louis, MO, USA).
Materials and equipment for SDS-PAGE Western blotting and protein
staining were obtained from Invitrogen (Carlsbad, CA, USA), and as
Fig. 1. MW characterization & confirmation of HA used in test lubricants (20 kDa,
132 kDa, 780 kDa, 1.5 MDa, 5 MDa) via AGE and staining, as described in the methods.
1358
J.J. Kwiecinski et al. / Osteoarthritis and Cartilage 19 (2011) 1356e1362
Lubrication test
Lubrication testing was performed on a Bose ELF 3200, using the
previously described cartilage-on-cartilage friction test5,10. Briefly, all
samples were compressed at a constant rate of 0.002 mm/s to 18% of
the total cartilage thickness, and were allowed a 40-min stress
relaxation period for interstitial fluid depressurization. Then, the
samples were rotated þ2 revolutions and then 2 revolutions at an
effective velocity of 0.3 mm/s (shown to maintain boundary mode
lubrication at a depressurized cartilageecartilage interface) with
pre-sliding durations of 1,200, 120, 12 and 1.2 seconds (s) (Tps; time
the sample is stationary prior to rotation, which physically represents
the time for which the two solid surfaces are in contact, also referred
to as dwell time31). The test sequence was then repeated in the
opposite direction of rotation.
Experimental design
To determine the cartilage boundary lubricating ability of a broad
range of MW HA, both alone and in combination with PRG4, four
sequential test sequences were performed. In all experiments, n ¼ 5
samples of articulating cartilage (typically harvested from two joints,
not necessarily from the same animal) were tested sequentially in five
test lubricants: PBS (serving as the negative control test lubricant), in
three test lubricants, and then in SF (serving as the positive control
test lubricant). Lubricants were prepared in PBS, with HA at a physiological concentration of 3.33 mg/ml, and PRG4 at 450 mg/ml.
HA. To determine the effect of MW on HA’s cartilage lubricating
ability alone at a physiological concentration, two tests were performed: HA Test 1: PBS, 132 kDa HA, 780 kDa HA, 1.5 MDa HA, then
SF; and HA Test 2: PBS, 20 kDa HA, 780 kDa HA, 5 MDa HA, then SF.
analysis of variance (ANOVA). To compare test lubricants containing HA and/or PRG4 within each test sequence, the effect of test
lubricant on mstatic, Neq and <mkinetic, Neq> at specific Tps was
assessed by ANOVA, with Fisher’s LSD post-hoc testing. Additionally, the dependencies of mstatic, Neq and <mkinetic, Neq> on the log of
HA MW were assessed by linear regression. Statistical analysis was
implemented with Systat12 (Systat Software, Inc., Richmond, CA).
Results
Lubrication test characterization
In all experiments, friction was modulated by test lubricant and
Tps. In all test lubricants, mstatic, Neq increased markedly with
increasing Tps, with mean SD values at Tps ¼ 1,200 s being on
average 68 13% greater than those at Tps ¼ 1.2 s, and appeared to
approach <mkinetic, Neq> asymptotically as Tps decreased from
1,200 s toward 0 s (to 1.2 s). Conversely, <mkinetic, Neq> increased
only slightly with mean SD values at Tps ¼ 1.2 s being on average
within 15 10% of values at Tps ¼ 1,200 s. Therefore, for brevity and
clarity, <mkinetic, Neq> data is presented at Tps ¼ 1.2 s only. In all test
sequences, values of mstatic, Neq were consistently greatest in PBS,
ranging from 0.245 0.027 to 0.488 0.031 (mean SD) with
increasing Tps; values of mstatic, Neq were consistently lowest in SF,
ranging from 0.035 0.006 to 0.204 0.021, (mean SD) with
increasing Tps. Similarly with <mkinetic, Neq>, values were greatest in
PBS, 0.185 0.020, and lowest in SF, 0.033 0.004 (mean SD).
The average equilibrium compressive stress for tests was
0.099 0.008 MPa (mean SD).
Effect of HA MW alone
HA þ PRG4. To determine the effect of MW on HA’s cartilage
lubricating ability in combination with PRG4, both at physiological
concentrations, two tests were performed: HA+PRG4 Test 1: PBS,
PBS þ PRG4, 132 kDa HA þ PRG4, 1.5 MDa HA þ PRG4, then SF.
HA + PRG4 Test 2: PBS, PBS þ PRG4, 20 kDa HA þ PRG4, 5 MDa
HA þ PRG4, then SF.+
EMSA
To examine the HA-PRG4 interaction in solution, EMSA was
used. PRG4 (0.3, 1, 3, and 10 mg), monodisperse HA (1 mg 150 kDa or
1,000 kDa), and monodisperse HA þ PRG4 were subject to 1% AGE
under non-reducing conditions for 3 h at 50 V, as described above.
(Samples were mixed and allowed to sit at room temperature for
30 min prior to expedient loading and beginning the electrophoresis.) The gel was then stained with Stains-All and destained30,
scanned, and the migration of the bands assessed by densitometric
analysis with Image J. Specifically, integrated pixel intensity of
PRG4 lanes were subtracted from the corresponding HA þ PRG4
lane and plotted vs migration distance to assess the effect of PRG4
on the migration distance of the HA.
Statistical analysis
To evaluate the boundary lubrication properties of each test
lubricant, two friction coefficients (m) were calculated, averaged for
the þ and e revolutions, as described previously10. Briefly, mstatic,
Neq represents the friction coefficient at startup from a static
condition, whereas <mkinetic, Neq> represents the friction coefficient
at steady sliding, or a kinetic condition.
Unless otherwise indicated, data is presented as mean with
a 95% confidence interval (CI) (lower limit, upper limit). The effects
of test lubricant and Tps (as a repeated factor) on each of the two
friction coefficients, mstatic, Neq and <mkinetic, Neq>, were assessed by
Friction coefficients were reduced by HA, in a MW-dependent
manner in the range tested here. HA Test 1 (132kDa, 780kDa,
1.5MDa): mstatic, Neq varied with test lubricant and Tps with an interaction (all P < 0.0001) [Fig. 2(A)]. Values of mstatic, Neq at Tps ¼ 1,200 s
were significantly greater in 132 kDa HA compared to both 780 kDa
and 1.5 MDa (P ¼ 0.021 and P ¼ 0.044 respectively), which were
similar to each other (P ¼ 0.70). <mkinetic, Neq> at Tps ¼ 1.2 s also varied
with test lubricant (P < 0.0001) [Fig. 2(B)]. Values of <mkinetic, Neq> in
132 kDa, 780 kDa, and 1.5 MDa HA were intermediate and not
significantly different from each other (P ¼ 0.41e0.91). HA Test 2
(20kDa, 780kDa, 5MDa): mstatic, Neq varied with test lubricant and Tps
(both P < 0.0001) with an interaction (P ¼ 0.001) [Fig. 2(C)]. Values of
mstatic, Neq at Tps ¼ 1,200 s were significantly greater in 20 kDa HA
compared to both 780 kDa and 5 MDa (P ¼ 0.003 and P < 0.0001,
respectively), which were also significantly different from each other
(P ¼ 0.049). <mkinetic, Neq> at Tps ¼ 1.2 s also varied with test lubricant
(P ¼ 0.02) [Fig. 2(D)]. Values of <mkinetic, Neq> in 20 kDa, 0.098 (0.089,
0.108), were significantly greater compared to 780 kDa, 0.080 (0.072,
0.088), and 5 MDa, 0.079 (0.070, 0.089), (P ¼ 0.012 and P ¼ 0.017
respectively), which were similar to each other (P ¼ 0.87).
Values of mstatic, Neq at Tps ¼ 1,200 s and <mkinetic, Neq> at
Tps ¼ 1.2 s in HA test lubricants, from both HA Tests 1 and 2, were
correlated to log HA MW. Linear regression showed a significant
correlation between mstatic, Neq at Tps ¼ 1,200 s (r2 ¼ 0.99, P < 0.0001,
slope ¼ 0.099) and log HA MW [Fig. 3(A)], as well as <mkinetic, Neq>
at Tps ¼ 1.2 s (r2 ¼ 0.77, P ¼ 0.049, slope ¼ 0.0086) and log HA MW
[Fig. 3(B)].
Effect of HA MW in combination with PRG4
Friction coefficients were reduced by PRG4, and with subsequent addition of HA; however the synergistic effect was not
dependent on the MW of the HA tested here. HA+PRG4 Test 1
J.J. Kwiecinski et al. / Osteoarthritis and Cartilage 19 (2011) 1356e1362
Fig. 2. Effect of MW of HA alone on the boundary lubrication of articular cartilage. Test
1 (n ¼ 5): 132 kDa, 780 kDa, 1.5 MDa (A, B); Test 2 (n ¼ 5): 20 kDa, 780 kDa, 5 MDa (C,
D). Shown are static mstatic, Neq (A, C), and kinetic <mkinetic, Neq> (the brackets indicate
that the value is an average) (B, D) friction coefficients in PBS, PBS þ various MW HA at
3.33 mg/ml, and SF. In B, D, the pre-sliding duration, Tps, was 1.2 s. Data are
mean 95% CI.
(PRG4, PRG4+132kDa, PRG4+1.5MDa): mstatic, Neq varied with test
lubricant and Tps (P < 0.0001) with an interaction (P ¼ 0.049)
[Fig. 4(A)]. Values of mstatic, Neq in PRG4 at Tps ¼ 1,200 s were
significantly greater compared to both PRG4 þ 132 kDa (P ¼ 0.012)
and PRG4 þ 1.5 MDa (P ¼ 0.006), which were similar to each other
1359
Fig. 4. Effect of MW of HA when combined with PRG4 on the boundary lubrication of
articular cartilage. Test 1 (n ¼ 5): PRG4, PRG4 þ 132 kDa, PRG4 þ 1.5 MDa (A, B) Test 2
(n ¼ 5): PRG4, PRG4 þ 20 kDa, PRG4 þ 5MDa (C, D). Shown are static mstatic, Neq (A, C),
and kinetic <mkinetic, Neq> (B, D) friction coefficients in PBS, PBS þ PRG4 at 450 mg/ml,
PBS þ PRG4 various MW HA at 3.33 mg/ml, and SF. In B, D the pre-sliding duration,
Tps, was 1.2 s. Data are mean 95% CI.
(P ¼ 0.71). Similarly, <mkinetic, Neq> at Tps ¼ 1.2 s also varied with test
lubricant (P < 0.0001) [Fig. 4(B)]. Values of <mkinetic, Neq> in PRG4,
0.089 (0.052, 0.125), were significantly greater than both
PRG4 þ 132 kDa, 0.051 (0.042, 0.061), and PRG4 þ 1.5 MDa, 0.046
(0.038, 0.053) (both P ¼ 0.037 and P ¼ 0.020 respectively), which
were similar to each other (P ¼ 0.74). HA+PRG4 Test 2 (PRG4,
PRG4+20kDa, PRG4+5MDa): mstatic, Neq varied with test lubricant
and Tps with an interaction (all P < 0.0001) [Fig. 4(C)]. Values of
mstatic, Neq in PRG4 at Tps ¼ 1,200 s were significantly greater
compared to both PRG4þ20 kDa and PRG4 þ 5MDa (both
P ¼ 0.004), which were similar to each other (P ¼ 0.99). Similarly,
<mkinetic, Neq> at Tps ¼ 1.2 s also varied with test lubricant
(P < 0.001) [Fig. 4(D)]. Values of <mkinetic, Neq> in PRG4 were
significantly greater than both PRG4 þ 20 kDa, 0.030 (0.025, 0.036),
and PRG4 þ 5 MDa, 0.032 (0.027, 0.037), (P ¼ 0.002 and P ¼ 0.003
respectively), which were similar to each other (P ¼ 0.90).
Values of mstatic, Neq at Tps ¼ 1,200 s and <mkinetic, Neq> at
Tps ¼ 1.2 s in HA þ PRG4 test lubricants, from both HA þ PRG4 Tests 1
and 2, were correlated to log HA MW. Linear regression showed no
significant correlation between mstatic, Neq at Tps ¼ 1,200 s
(r2 ¼ 0.0014, P ¼ 0.96) and log HA MW [Fig. 5(A)], as well as <mkinetic,
2
Neq> at Tps ¼ 1.2 s (r ¼ 0.011, P ¼ 0.89) and log HA MW [Fig. 5(B)].
EMSA
Fig. 3. Dependence of the cartilage boundary lubricating properties of HA alone at
3.33 mg/ml on MW. Regression lines are shown for mean static mstatic, Neq friction
values at pre-sliding duration, Tps ¼ 1,200 s (A), and mean kinetic <mkinetic, Neq> friction
values at Tps ¼ 1.2 s (B) obtained from HA Test 1 & 2 (Fig. 2) vs log MW HA. Mean values
in PBS and SF are shown for reference.
EMSA data indicated that migration of both 150 kDa and
1,000 kDa HA were retarded when combined with PRG4 at high
PRG4:HA ratios. PRG4 alone appeared predominantly as broad
migrating purple bands, and both 150 kDa and 1,000 kDa HA
appeared as tight blue bands whose migration appeared slightly
retarded when electrophoresed with increasing amounts of PRG4
[Fig. 6(A and C)]. Densitometric analysis confirmed the migration of
both 150 kDa and 1,000 kDa HA was indeed retarded when
combined with PRG4. The greatest apparent effect was observed in
both cases with the 10 mg of PRG4, although there was evidence of
a subtle shift with 1 mg PRG4 as well [Fig. 6(B and D)].
1360
J.J. Kwiecinski et al. / Osteoarthritis and Cartilage 19 (2011) 1356e1362
Fig. 5. Dependence of the cartilage boundary lubricating properties of PRG4 þ HA, at
450 mg/ml and 3.33 mg/ml respectively, on HA MW. Regression lines are shown for
mean static mstatic, Neq friction values at pre-sliding duration, Tps ¼ 1,200 s (A), and
mean kinetic <mkinetic, Neq> friction values at Tps ¼ 1.2 s (B) obtained from HA þ PRG4
Test 1 & 2 (Fig. 4) vs log MW HA. Mean values in PBS, PRG4 at 450 mg/ml, and SF are
shown for reference.
Discussion
The results described here indicate that the cartilage boundary
lubricating ability of HA alone at physiological levels varies with
MW; however when combined with PRG4 at physiological levels
the synergistic friction-reducing ability of HA þ PRG4 did not vary
with the MW of HA tested here. In test lubricants containing HA
alone at 3.33 mg/ml, values of both mstatic, Neq and <mkinetic, Neq>
significantly correlated with log MW; increasing with decreasing
MW of HA. mstatic, Neq at Tps ¼ 1,200 s varied markedly with HA MW
[Fig. 3(A)], while <mkinetic, Neq> at Tps ¼ 1.2 s varied less so
[Fig. 3(B)], yet still in a significant manner. Conversely, when HA at
3.33 mg/ml was combined with PRG4 at 450 mg/ml, values of both
mstatic, Neq and <mkinetic, Neq> did not vary significantly with HA MW
(Fig. 5). Furthermore, the retardation of 150 kDa and 1,000 kDa HA
when combined with PRG4 observed via EMSA suggests HA-PRG4
complexes of various MW HA can form in solution (Fig. 6). Collectively, these results suggest that alterations in HA MW in SF could
significantly affect SF’s cartilage boundary lubricating ability.
However, this diminishment in function could be circumvented by
the presence of physiological levels of PRG4 in SF that can form
a complex and function synergistically to maintain SF’s effective
lubricating ability.
The HA and PRG4 used here were representative of those in
native SF and have been used in other studies. The range of MW
used spanned the majority of the physiological range of HA present
in normal SF8,9. The various MW HA used for lubrication testing
were polydisperse, as indicated by the MW characterization via
AGE (Fig. 1), therefore the MW reported here were that provided by
the suppliers. The design of lubricant test sequences was guided by
the number of sequential tests that could be conducted on a single
osteochondral sample pair; additional testing indicated the order
of testing used here did not affect the reported results
(Supplementary Data). Monodisperse HA was used for EMSA to
visualize any potential shift in electrophoretic migration due to an
interaction with PRG4. 150 kDa and 1,000 kDa were chosen as
relatively low and high MW HA that would not co-migrate with the
PRG4. The PRG4 prepared essentially as described previously10,
using skeletally mature bovine cartilage for the explant culture, and
contained high MW species of PRG432 known to be present in
bovine SF29.
The results of this study agree with and extend a previous study
examining the boundary lubricating abilities of HA and PRG4 at
Fig. 6. EMSA of 1 mg 150 kDa (A) and 1,000 kDa (C) monodisperse HA with graded amounts of PRG4 (0, 1, 10 mg shown), and densitometric analysis of migration distance of 150 kDa
(B) and 1,000 kDa (D) HA, as described in the methods. PRG4:HA molar ratios are calculated using an approximate average MW of PRG4 ¼ 400 kDa.
J.J. Kwiecinski et al. / Osteoarthritis and Cartilage 19 (2011) 1356e1362
a cartilageecartilage biointerface. Values of mstatic, Neq in 780 kDa
HA at Tps ¼ 120 s and <mkinetic, Neq> at Tps ¼ 1.2 s are similar to those
reported in a recent preliminary study examining the effects of HA
concentration and MW33, as well as for w800 kDa SupArtz HA in
another study10, at the same concentration of 3.33 mg/ml. Similarly, the values reported here for PRG4 at 450 mg/ml, purified from
media conditioned by mature bovine cartilage explants, are similar
to those reported for PRG4 at the same concentration, although
purified from media conditioned by immature bovine explants10.
This suggests that both immature and mature cartilage secretes
functional, friction-reducing, PRG4. The relatively small variation of
<mkinetic, Neq> at Tps ¼ 1.2 s with the MW of HA tested here, yet
correlation with log MW [Fig. 3(B)], suggests that even relatively
low MW HA (e.g., 20 kDa) maintains the ability to interact with the
surface of articular cartilage and function as a boundary lubricant
reasonably well. The marked variation of mstatic, Neq at Tps ¼ 1,200 s
with MW, and again significant correlation [Fig. 3(A)], may be
related to excluded volume effects34 (i.e., macromolecular crowding, which increases the effective concentration of the HA molecules, and which depends on the hydrodynamic volume of the HA
molecules). mstatic, Neq significantly correlating with log MW at all
other earlier Tps as well (Tps ¼ 120 s: r2 ¼ 0.96, P ¼ 0.004; Tps ¼ 12 s:
r2 ¼ 0.91, P ¼ 0.013; Tps ¼ 1.2 s: r2 ¼ 0.94, P ¼ 0.016) with decreasing
slopes (0.069, 0.040, 0.034, respectively) suggests there is
a potential temporal dependency, in terms of time the apposed
surfaces are allowed to interact (i.e., Tps, which affects the adhesion
between the contacting surfaces and therefore static friction31), of
the MW effect as well. The previously reported PRG4 þ HA synergism in 3.33 mg/ml SupArtz HA þ 450 mg/ml PRG4 prepared from
immature bovine cartilage explants10 was observed in the present
study over a range of MW HA (20 kDa, 132 kDa, 1.5 MDa, 5 MDa)
combined with PRG4 prepared from mature bovine cartilage
explants. The reason values of <mkinetic, Neq> at Tps ¼ 1.2 s appeared
to be slightly lower, and approaching more closely the SF controls,
may be due an enhanced ability of PRG4 prepared from mature
bovine explants to interact with HA. However, more studies are
required to fully elucidate any potential effect the age of cartilage
from which PRG4 is produced has on the synergistic effect of
PRG4 þ HA.
While indirect biophysical evidence for a functional PRG4eHA
interaction in SF has been reported using a particle tracking
microrheology technique27, the retardation of both 150 kDa and
1,000 kDa HA when combined with PRG4 observed here [Fig. 6(A
and C)] directly suggests an HAePRG4 complex of various MW HA
can form in solution. This shift in electrophoretic mobility was small
compared to typical EMSAs28, which suggests the HAePRG4
complex is not a strong one that moves as a specific species, but
more as a reversible interaction that forms and breaks apart, slowing
the HA migration slightly. However, the shift was consistent and
repeatable, was clearly observed without any prelabeling, and was
not dependent upon the location of sample loading or due to
a ‘smiling’ effect of the migration front (data not shown). The 10:1
weight ratio of PRG4:HA mg required to observe a shift is greater then
that approximately present in SF, w1:7 for the physiological
concentrations used here, which may be a limitation of the technique’s resolution. However, the PRG4:HA molar ratios for the
10 mg þ 150 kDa and 1 mg þ 1,000 kDa, assuming an approximate
average MW of PRG4 ¼ 400 kDa (based on reported values of
w345 kDa20 and w445 kDa32), are w3.8 and w2.5 respectively,
which are at least on the same order of magnitude of that present in
native SF w 0.75, (assuming an average MW HA ¼ 3 MDa8,9). The
PRG4 dose-dependency of the observed shift may be due to multiple
PRG4 molecules being able to interact with a single HA molecule.
These collective results provide insight into the cartilage
boundary lubricating properties of HA, alone and in combination
1361
with PRG4. The decline in lubricating ability associated with lower
MW HA alone suggests that alterations in SF MW HA composition
could significantly affect normal joint lubrication function. The
observed PRG4 þ HA functional synergism with various MW HA
further suggests that normal levels of PRG4 may be able to maintain
normal lubricating function of SF even when HA MW composition
is altered. Therefore, in situations where PRG4 levels are decreased
in SF, the molecular mechanism for diminished lubricating ability
could involve two aspects: (1) absence of PRG4 to contribute both
alone (in a dose-dependent manner10) and synergistically with HA
to reduce friction; (2) altered MW of HA whose effects on diminished lubricating ability becomes more apparent with absence of
PRG4 and its synergistic function with HA to reduce friction. More
sophisticated biophysical experimental techniques, such as microrheology or confocal fluorescence recovery after photobleaching35,
could be useful to further examine the PRG4 þ HA complex in
native SF. Likewise, the potential effect of PRG4’s monomeric/
multimeric structure29 on this complex formation is currently
unclear. As such, the underlying mechanism of HA þ PRG4 interactions in solution and at the articular surface, which need not be
the same, are both important in terms of cartilage lubrication and
remain to be determined. Elucidation of such mechanism(s) could
contribute to the development of SF biotherapeutics for the treatment or potentially prevention of OA in situations where lubricant
molecule composition is altered and/or SF lubricating function is
diminished.
Author contributions
All authors contributed to the conception and design of the
original study and approved the final submitted manuscript. JK, SD,
and SA were responsible of the data acquisition and analysis. The
article was first drafted by JK, and critically reviewed by SD, TL, SA,
MC, and TS. TS obtained funding for the study, and takes full
responsibility for the integrity of the work as a whole.
Conflict of interest
The authors have no potential conflicts of interest.
Acknowledgments
This work was supported by funding from the Natural Sciences
& Engineering Research Council of Canada, the Canadian Arthritis
Network, as well as the Faculty of Kinesiology and the Schulich
School of Engineering’s Centre for Bioengineering Research and
Education at the University of Calgary.
Supplementary material
Supplementary data related to this article can be found online at
doi:10.1016/j.joca.2011.07.019.
References
1. Ateshian GA, Mow VC. Friction, lubrication, and wear of articular
cartilage and diarthrodial joints. In: Mow VC, Huiskes R, Eds.
Basic Orthopaedic Biomechanics and Mechano-Biology. Philadelphia: Lippincott Williams & Wilkins; 2005:447e94.
2. Balazs EA, GIbbs DA. The rheological properties and biological
function and hyaluronic acid. In: Balazs EA, Ed. Chemistry and
Molecular Biology of the Intercellular Matrix. London, UK:
Academic Press; 1970:1241e54.
3. Gomez JE, Thurston GB. Comparisons of the oscillatory shear
viscoelasticity and composition of pathological synovial fluids.
Biorheology 1993;30:409e27.
1362
J.J. Kwiecinski et al. / Osteoarthritis and Cartilage 19 (2011) 1356e1362
4. Whiting W, Zernicke R. Biomechanics of Musculoskeletal
Injury. Champaign, IL: Human Kinetics; 2008.
5. Schmidt TA, Sah RL. Effect of synovial fluid on boundary
lubrication of articular cartilage. Osteoarthritis Cartilage
2007;15:35e47.
6. Watterson JR, Esdaile JM. Viscosupplementation: therapeutic
mechanisms and clinical potential in osteoarthritis of the knee.
J Am Acad Orthop Surg 2000;8:277e84.
7. Mazzucco D, Scott R, Spector M. Composition of joint fluid in
patients undergoing total knee replacement and revision
arthroplasty: correlation with flow properties. Biomaterials
2004;25:4433e45.
8. Laurent TC, Ryan M, Pietruszkiewicz A. Fractionation of hyaluronic acid. The polydispersity of hyaluronic acid from the
bovine vitreous body. Biochim Biophys Acta 1960;42:476e85.
9. Ghosh P. The role of hyaluronic acid (hyaluronan) in health
and disease: interactions with cells, cartilage and components
of synovial fluid. Clin Exp Rheumatol 1994;12:75e82.
10. Schmidt TA, Gastelum NS, Nguyen QT, Schumacher BL, Sah RL.
Boundary lubrication of articular cartilage: role of synovial
fluid constituents. Arthritis Rheum 2007;56:882e91.
11. Fam H, Bryant JT, Kontopoulou M. Rheological properties of
synovial fluids. Biorheology 2007;44:59e74.
12. Dahl LB, Dahl IM, Engstrom-Laurent A, Granath K. Concentration and molecular weight of sodium hyaluronate in synovial
fluid from patients with rheumatoid arthritis and other
arthropathies. Ann Rheum Dis 1985;44:817e22.
13. Balazs EA, Watson D, Duff IF, Roseman S. Hyaluronic acid in
synovial fluid. I. Molecular parameters of hyaluronic acid in
normal and arthritis human fluids. Arthritis Rheum 1967;
10:357e76.
14. Waddell DD, Marino AA. Chronic knee effusions in patients with
advanced osteoarthritis: implications for functional outcome of
viscosupplementation. J Knee Surg 2007;20:181e4.
15. Balazs EA. Analgesic effect of elastoviscous hyaluronan solutions and the treatment of arthritic pain. Cells Tissues Organs
2003;174:49e62.
16. Balazs EA. Viscosupplementation for treatment of osteoarthritis: from initial discovery to current status and results.
Surg Technol Int 2004;12:278e89.
17. Goldberg VM, Buckwalter JA. Hyaluronans in the treatment of
osteoarthritis of the knee: evidence for disease-modifying
activity. Osteoarthritis Cartilage 2005;13:216e24.
18. Ikegawa S, Sano M, Koshizuka Y, Nakamura Y. Isolation,
characterization and mapping of the mouse and human PRG4
(proteoglycan 4) genes. Cytogenet Cell Genet 2000;90:291e7.
19. Swann DA, Silver FH, Slayter HS, Stafford W, Shore E. The
molecular structure and lubricating activity of lubricin isolated
from bovine and human synovial fluids. Biochem J 1985;
225:195e201.
20. Schumacher BL, Block JA, Schmid TM, Aydelotte MB,
Kuettner KE. A novel proteoglycan synthesized and secreted
by chondrocytes of the superficial zone of articular cartilage.
Arch Biochem Biophys 1994;311:144e52.
21. Jay GD. Characterization of a bovine synovial fluid lubricating
factor. I. Chemical, surface activity and lubricating properties.
Connect Tissue Res 1992;28:71e88.
22. Jay GD. Lubricin and surfacing of articular joints. Curr Opin
Orthop 2004;15:355e9.
23. Jay GD, Lane BP, Sokoloff L. Characterization of a bovine
synovial fluid lubricating factor III. The interaction with hyaluronic acid. Connect Tissue Res 1992;28:245e55.
24. Schmid T, Homandberg G, Madsen L, Su J, Kuettner K. Superficial zone protein (SZP) binds to macromolecules in the
lamina splendens of articular cartilage. Trans Orthop Res Soc
2002;27:359.
25. Chang DP, Abu-Lail NI, Guilak F, Jay GD, Zauscher S. Conformational mechanics, adsorption, and normal force interactions
of lubricin and hyaluronic acid on model surfaces. Langmuir
2008;24:1183e93.
26. Greene GW, Banquy X, Lee DW, Lowrey DD, Yu J,
Israelachvili JN. Adaptive mechanically controlled lubrication
mechanism found in articular joints. Proc Natl Acad Sci U S A
2011;108:5255e9.
27. Jay GD, Torres JR, Warman ML, Laderer MC, Breuer KS. The role
of lubricin in the mechanical behavior of synovial fluid. Proc
Natl Acad Sci U S A 2007;104:6194e9.
28. Hellman LM, Fried MG. Electrophoretic mobility shift assay
(EMSA) for detecting protein-nucleic acid interactions. Nat
Protoc 2007;2:1849e61.
29. Schmidt TA, Plaas AH, Sandy JD. Disulfide-bonded multimers
of proteoglycan 4 (PRG4) are present in normal synovial fluids.
Biochim Biophys Acta 2009;1790:375e84.
30. Lee HG, Cowman MK. An agarose gel electrophoretic method
for analysis of hyaluronan molecular weight distribution. Anal
Biochem 1994;219:278e87.
31. Bhushan B. Introduction to Tribology. Wiley; 2002.
32. Alvarez MC, Kooyman J, Schmidt TA. Synthesis of proteoglycan
4 (PRG4) disulfide-bonded multimers by chondrocytes in
cartilage explants. Trans Orthop Res Soc 2010;56:850.
33. Antonacci JM, Cai MZ, Beeman SM, Schumacher BL, Chen AC,
McIlwraith CW, et al. Mechanisms of articular cartilage lubrication by hyaluronan: effects of concentration and molecular
weight. Trans Orthop Res Soc 2011;36:2138.
34. Cowman MK, Matsuoka S. Experimental approaches to hyaluronan structure. Carbohydr Res 2005;340:791e809.
35. Raynal BD, Hardingham TE, Thornton DJ, Sheehan JK.
Concentrated solutions of salivary MUC5B mucin do not
replicate the gel-forming properties of saliva. Biochem J 2002;
362:289e96.