ARTHRITIS & RHEUMATISM
Vol. 56, No. 3, March 2007, pp 882–891
Boundary Lubrication of Articular Cartilage
Role of Synovial Fluid Constituents
Tannin A. Schmidt, Nicholas S. Gastelum, Quynhhoa T. Nguyen,
Barbara L. Schumacher, and Robert L. Sah
Objective. To determine whether the synovial
fluid (SF) constituents hyaluronan (HA), proteoglycan 4
(PRG4), and surface-active phospholipids (SAPL) con-
tribute to boundary lubrication, either independently or
additively, at an articular cartilage–cartilage interface.
Methods. Cartilage boundary lubrication tests
were performed with fresh bovine osteochondral sam-
ples. Tests were performed using graded concentrations
of SF, HA, and PRG4 alone, a physiologic concentration
of SAPL, and various combinations of HA, PRG4, and
SAPL at physiologic concentrations. Static (␮static, Neq)
and kinetic (<␮kinetic, Neq>) friction coefficients were
calculated.
Results. Normal SF functioned as an effective
boundary lubricant both at a concentration of 100%
(<␮kinetic, Neq> ⴝ 0.025) and at a 3-fold dilution
(<␮kinetic, Neq> ⴝ 0.029). Both HA and PRG4 contrib-
uted independently to a low ␮ in a dose-dependent
manner. Values of <␮kinetic, Neq> decreased from ⬃0.24
in phosphate buffered saline to 0.12 in 3,300 ␮g/ml HA
and 0.11 in 450 ␮g/ml PRG4. HA and PRG4 in combi-
nation lowered ␮ further at the high concentrations,
attaining a <␮kinetic, Neq> value of 0.066. SAPL at 200
␮g/ml did not significantly lower ␮, either indepen-
dently or in combination with HA and PRG4.
Conclusion. The results described here indicate
that SF constituents contribute, individually and in
Supported by the NIH and NSF. Dr. Sah’s work was sup-
ported by a Howard Hughes Medical Institute Professors Program
award to the University of California, San Diego.
Tannin A. Schmidt, PhD, Nicholas S. Gastelum, Quynhhoa T.
Nguyen, BS, Barbara L. Schumacher, BS, Robert L. Sah, MD, ScD:
University of California, San Diego, La Jolla.
Address correspondence and reprint requests to Robert L.
Sah, MD, ScD, Department of Bioengineering, Mail Code 0412, 9500
Gilman Drive, University of California, San Diego, La Jolla, CA
92093-0412. E-mail: rsah@ucsd.edu.
Submitted for publication June 12, 2006; accepted in revised
form December 6, 2006.
882
DOI 10.1002/art.22446
© 2007, American College of Rheumatology
combination, both at physiologic and pathophysiologic
concentrations, to the boundary lubrication of apposing
articular cartilage surfaces. These results provide in-
sight into the nature of the boundary lubrication of
articular cartilage by SF and its constituents. They
therefore provide insight regarding both the homeo-
static maintenance of healthy joints and pathogenic
processes in arthritic disease.
Articular cartilage is the lubricious, load-bearing
tissue at the end of long bones in synovial joints that
normally facilitates low-friction and low-wear articula-
tion. When healthy, it provides low-friction properties to
the synovial joint through a combination of lubrication
mechanisms (1). Pressurized fluid, within the tissue and
between the surfaces, such as in a fluid film, can bear
significant portions of the load. Lubricant molecules
within a surface layer or film at the articular surface also
mediate load bearing, particularly the surface-to-surface
contact in the boundary mode of lubrication. This mode
of lubrication has been proposed to be important for the
protection and maintenance of articular surfaces since
the apposing cartilage layers within the joint make
contact over ⬃10% of the total area, where much of the
friction may occur (2). Synovial fluid (SF) contains
the molecules hyaluronan (HA) (3), proteoglycan 4
(PRG4) (the name assigned by the Human Genome
Organization Gene Nomenclature Committee for pro-
teins also known as lubricin, superficial zone protein,
and megakaryocyte-stimulating factor) (4,5), and
surface-active phospholipids (SAPL) (6), each of which
interacts with and adsorbs to the articular surface. Such
molecules are all ideally positioned to contribute to
boundary lubrication.
SF, as well as HA, PRG4, and SAPL, have each
demonstrated boundary-lubricating ability at various
test interfaces. SF was recently shown to function as an
BOUNDARY LUBRICATION OF ARTICULAR CARTILAGE
effective boundary lubricant at an interface between
apposed articular cartilage surfaces using an annulus-on-
disc configuration (7). These results were consistent with
findings of several previous studies and extended them
using native cartilage surfaces in similar (8,9) and dif-
ferent (10,11) test configurations, as well as using nonbio-
logic surfaces (8,12,13). The lubricating ability of HA
has been assessed at cartilage–cartilage (14–18),
cartilage–steel (19), and cartilage–glass interfaces (18,20),
as well as at a latex–glass interface under boundary lubri-
cation conditions (12,21), with variable conclusions, possi-
bly due to the different test surfaces and configurations
and the various resulting operative modes of lubrication.
Conversely, PRG4 proteins (22), which are synthesized
and secreted by cells lining the synovial cavity (4,5), have
consistently demonstrated boundary-lubricating ability
at both cartilage–glass (23) and latex–glass interfaces
(12,21,24,25). Studies examining the lubricating ability
of SAPL at a cartilage–cartilage interface (in combina-
tion with HA) (19) and at a cartilage–steel interface
(19,26), as well as at a latex–glass interface under
boundary lubrication conditions (13), suggest that SAPL
may also possess boundary-lubricating ability. Collec-
tively, these studies suggest that HA, PRG4, and SAPL
each may contribute to the boundary-lubricating ability
of SF at a cartilage–cartilage interface.
Consequently, the boundary-lubricating ability of
SF may be altered in joint injury and arthritis due to the
alteration in concentrations of HA, PRG4, and SAPL.
The concentration of HA in human SF ranges from 1–4
mg/ml in healthy individuals (27,28) and decreases after
effusive joint injury (29) and in arthritic disease to
⬃0.1–1.3 mg/ml (28,30). The concentration of PRG4 in
human SF ranges from 52 ␮g/ml to 350 ␮g/ml postmor-
tem and from 276 ␮g/ml to 762 ␮g/ml in SF obtained
from patients undergoing arthrocentesis procedures
(31). Conversely, using a rabbit knee injury model, the
concentration of PRG4 in SF decreased from 280 ␮g/ml
to 20–100 ␮g/ml 3 weeks after injury (32). The majority
of the lipids in human SF are phospholipids, the con-
centration of which ranges from ⬃0.1 mg/ml to ⬃0.2
mg/ml in normal individuals, increases in osteoarthritis
to ⬃0.2–0.3 mg/ml (28), and can decrease following
traumatic injury to ⬃0.02–0.08 mg/ml (33). While most
phospholipids are surface active, dipalmitoyl-phospha-
tidylcholine (DPPC) is particularly so and is the most
abundant form present in SF at ⬃45% (6,34).
The governing hypothesis of the present study
was that SF constituents contribute to the boundary
lubrication of articular cartilage. The specific objective
of this study was to determine whether the SF constitu-
883
ents HA, PRG4, and SAPL contribute to boundary
lubrication, either independently or additively, at an
articular cartilage–cartilage interface. To achieve this
objective, the effect of graded dilutions of SF on carti-
lage boundary lubrication was first determined. Then,
the independent effects of graded concentrations of HA
and PRG4, and of a physiologic concentration of SAPL,
on cartilage boundary lubrication were determined. Fi-
nally, the combined effect of physiologic concentrations
of HA, PRG4, and SAPL on cartilage boundary lubri-
cation was determined.
MATERIALS AND METHODS
Materials. Materials for lubrication testing were ob-
tained as described previously (7). In addition, high molecular
weight (MW) sodium hyaluronate (HA; SUPARTZ, 10 mg/ml,
620–1,170 kd) was obtained from Seikagaku Corporation
(Tokyo, Japan), and DPPC was obtained from Sigma-Aldrich
(St. Louis, MO).
Lubricant preparation and characterization. The con-
centrations of HA, PRG4, and phospholipids in the test
lubricants were determined by the carbazole reaction for
uronic acid (35), by enzyme-linked immunosorbent assay
(ELISA) (36), and by phospholipid assay using Phospholipids
B Standard Solution and Color Reagent (Wako, Richmond,
VA) (28), respectively.
HA. The concentration of HA in the SUPARTZ HA
stock solution was confirmed prior to storage at ⫺20°C.
PRG4. PRG4 was prepared from ⬃300 cartilage discs
(6 mm diameter and ⬃0.3 mm thick, including the articular
surface) harvested from 6 immature bovine stifle joints. Car-
tilage discs were incubated for 6–15 days in Dulbecco’s mod-
ified Eagle’s medium with 0.01% bovine serum albumin, 25
␮g/ml ascorbic acid, and 10 ng/ml recombinant human trans-
forming growth factor ␤1 (PeproTech, Rocky Hill, NJ). Cul-
ture medium was changed every 3 days and collected for
processing. To purify PRG4, the conditioned culture medium
was saved, pooled, and fractionated by anion-exchange chro-
matography essentially as described previously (4). Briefly, the
sample was applied onto DEAE-Sepharose, previously equili-
brated with 0.15M NaCl, 0.005M EDTA, and 0.05M sodium
acetate, pH 6.0. The 0.3–0.6M NaCl eluate was collected,
concentrated with a Centricon Plus 100-kd MW cutoff filter,
and then quantified by ELISA (36) using monoclonal antibody
(mAb) 3A4 (a gift from Dr. Bruce Caterson) (37) prior to
storage at ⫺20°C. Control studies indicated that the DEAE
buffer used for PRG4 purification did not alter boundary-
lubricating ability, since SF samples (described below) retained
lubricating ability after dialysis against the buffer.
The size distribution of immunoreactive PRG4 was
characterized by Western blot using mAb 3A4 after electro-
phoresis on a 4–20% sodium dodecyl sulfate–polyacrylamide
gel and transfer to a polyvinylidene difluoride membrane. A
single immunoreactive band at ⬃345 kd was visualized by
ECL-Plus detection and digital scanning with a STORM 840
Imaging System (Molecular Dynamics, Fairfield, CT). The
concentrations of HA (determined from uronic acid concen-
884 SCHMIDT ET AL

tration) and phospholipids in the PRG4 preparation (at 450

␮g/ml) were 30 ␮g/ml and ⬍0.5 ␮g/ml, respectively.
SAPL. A 10⫻ solution of DPPC at 2,000 ␮g/ml (⬃106
times the concentration of phospholipids at which micelles can
be formed [38]) was sonicated (Sonics & Materials, Danbury,
CT) (13) in phosphate buffered saline (PBS) for 15 minutes to
solubilize the DPPC, a standard technique used to produce
liposomes (39) resulting in an opaque solution for which the
concentration of SAPL was confirmed. The SAPL solution was
stored at ⫺20°C prior to use.
SF. Normal bovine SF pooled from 5 animals was
obtained as described previously (7) and clarified by centrifu-
gation (10,000g for 60 minutes at 4°C) prior to storage at
⫺80°C. The concentrations of HA and phospholipid were
⬃1,000 ␮g/ml and ⬃100 ␮g/ml, respectively. The concentra-
tion of the major immunoreactive PRG4 band at ⬃345 kd,
visualized by Western blot (described above) after hyaluroni-
dase treatment, was calculated to be ⬃450 ␮g/ml by quantita-
tive comparison to a similar molecular weight band of a known
amount of purified bovine PRG4 (4), as determined by ELISA
(36) (described above).
Sample preparation. Fresh osteochondral samples
(n ⫽ 40) were prepared for friction testing from the patel-
lofemoral groove of 10 skeletally mature bovine stifle joints
(⬃1 year old, ⬃0.86 moles pyridinoline/mole collagen [40]), as Figure 1. Boundary lubrication test protocol. A, Osteochondral annu-
described previously (7). Briefly, each sample consisted of an lus and core samples were compressed axially to a compression level
osteochondral core (radius 6 mm) and an apposed osteochon- (1 ⫺ ⌳Z, where ⌳Z is the stretch ratio) equal to 18% of the total
dral annulus (outer radius [Ro] 3.2 mm, inner radius [Ri] 1.5 cartilage thickness. B, A rotational test protocol with preconditioning,
mm), both with central holes (radius 0.5 mm) drilled down into a stress relaxation duration (Tsr) of 3,600 seconds, and an effective
and exiting the bone to facilitate fluid depressurization. In sliding velocity (veff) of 0.3 mm/second was then used to determine the
addition, samples were first rinsed vigorously overnight in ⬃40 effects of test lubricants and presliding durations (Tps) of 1,200, 120,
ml of PBS to deplete the articular surface of residual SF prior 12, and 1.2 seconds on the boundary lubrication of articular cartilage.
to lubrication testing in PBS. (Pilot studies confirmed that the
glycosaminoglycan content within the articular cartilage [41] of
rinsed samples was similar [within 1%] to that of nonrinsed
samples [P ⫽ 0.83] [n ⫽ 4].) Samples were then bathed in ⬃0.5 Experimental design. To determine whether HA,
ml of the subsequent test lubricants, completely immersing the PRG4, and SAPL contribute to cartilage boundary lubrication
cartilage, at 4°C for 24 hours prior to lubrication testing. either independently or additively, test lubricants were pre-
Lubrication test. Cartilage boundary lubrication tests pared in PBS. In all experiments, samples of articulating
(Figure 1) were performed on an ELF 3200 (Bose Endur- cartilage substrate were tested 3–5 times, in PBS (serving as
aTEC, Minnetonka, MN) essentially as described previously the negative control test lubricant) on the first day of lubrica-
(7). Briefly, samples of articulating cartilage were precondi- tion testing, in various test lubricant(s), and then in SF (serving
tioned by compressing at a constant rate of 0.002 mm/second as the positive control test lubricant) on the last day of
to a compression level (1 ⫺ ⌳Z, where ⌳Z is the stretch ratio lubrication testing.
[42]) equal to 18% of the total cartilage thickness, rotated ⫹2 Graded dilutions of SF. To determine the effect of
revolutions and then ⫺2 revolutions at an effective sliding graded dilutions of SF on cartilage boundary lubrication, tests
velocity (veff; equal to ␻Reff, where ␻ is the angular frequency were performed in PBS, 3.3% SF, 10% SF, 33% SF, and then
in radians/second and Reff is the effective radius, calculated to 100% SF.
be 2/3 ⫻ [(Ro3 – Ri3)/(Ro2 – Ri2)] ⫽ 2.4 mm [9]) equal to 3 SF constituents alone. To determine the independent
mm/second, and then unloaded to 0%. This sequence was then effects of graded concentrations of HA and PRG4 and of a
repeated twice more. Samples were then tested by first com- physiologic concentration of SAPL on cartilage boundary
pressing to 1 ⫺ ⌳Z ⫽ 18% and allowing a 60-minute stress lubrication, 3 sequences of tests were performed. For HA, tests
relaxation duration (Tsr) for interstitial fluid depressurization. were performed in PBS, 110 ␮g/ml HA, 1,100 ␮g/ml HA, 3,300
Then, samples were rotated ⫹2 revolutions and then ⫺2 ␮g/ml HA, and then SF. For PRG4, tests were performed in
revolutions at a veff of 0.3 mm/second (which is slow enough to PBS, 4.5 ␮g/ml PRG4, 45 ␮g/ml PRG4, 450 ␮g/ml PRG4, and
maintain a boundary mode of lubrication at a depressurized then SF. For SAPL, tests were performed in PBS, 200 ␮g/ml
articular cartilage–cartilage interface [7] and is on the order of SAPL, and then SF.
that used in other test configurations [43]) with presliding SF constituents in combination. To determine the ad-
durations (Tps; the duration the sample is stationary prior to ditive effect of physiologic concentrations of HA (3,300 ␮g/
rotation) of 1,200, 120, 12, and 1.2 seconds. The test sequence ml), PRG4 (450 ␮g/ml), and SAPL (200 ␮g/ml) in various
was then repeated in the opposite direction of rotation. combinations on cartilage boundary lubrication, tests were
BOUNDARY LUBRICATION OF ARTICULAR CARTILAGE 885

performed in PBS, HA or PRG4, HA plus PRG4, HA plus

PRG4 plus SAPL, and then SF.
Statistical analysis. To evaluate the boundary lubrica-
tion properties of test lubricants, 2 friction coefficients (␮)
were determined, as described previously (7). These para-
meters were calculated from the expression ␮ ⫽ ␶/(Reff ⫻
⌵eq), where ␶ is torque, ⌵eq is the equilibrium axial load after
the 60-minute Tsr, and Reff is the effective radius of the
annulus sample described above (9). Briefly, a static friction
coefficient, ␮static, Neq, was calculated using the peak | ␶ |,
measured just after (within 10° of) the start of rotation, and
the equilibrium axial load at the end of the 60-minute
stress relaxation period, ⌵eq. A kinetic friction coefficient,
⬍␮kinetic, Neq⬎ (the brackets indicate that the value is an
average), was calculated using the | ␶ | averaged during the
second complete revolution of the test sample and ⌵eq. Then,
␮static, Neq and ⬍␮kinetic, Neq⬎ were averaged for the ⫹ and ⫺
revolutions in each test.
Unless indicated otherwise, data are presented as the
mean ⫾ SEM. The effects of test lubricant and Tps (as a
repeated factor) on each of the 2 friction coefficients,
␮static, Neq and ⬍␮kinetic, Neq⬎, were assessed by analysis of
variance (ANOVA). In all test lubricants, ⬍␮kinetic, Neq⬎
increased slightly with increasing Tps, with mean ⫾ SD values
at Tps ⫽ 1.2 seconds being on average within 9 ⫾ 11%, or
0.009 ⫾ 0.015, of values at Tps ⫽ 1,200 seconds (which ranged
from ⬃0.02 to ⬃0.3) (P ⬍ 0.001). Therefore, for brevity and
clarity, ⬍␮kinetic, Neq⬎ data are presented at Tps ⫽ 1.2 seconds
only. Accordingly, the effect of test lubricant on ⬍␮kinetic, Neq⬎
at Tps ⫽ 1.2 seconds was assessed by ANOVA, with Tukey post
hoc testing. Statistical analysis was implemented with Systat
10.2 (Systat, Richmond, CA).

RESULTS Figure 2. Effect of graded concentrations of synovial fluid (SF) on

Lubrication test characterization. In all experi-
ments, friction was modulated by test lubricant (e.g.,
differing at least between SF, the positive control, and
PBS, the negative control) and Tps. In all test lubricants,
␮static, Neq increased with increasing Tps and appeared
to approach ⬍␮kinetic, Neq⬎ asymptotically as Tps de-
creased from 1,200 seconds toward 0 seconds (to 1.2
seconds). Mean ⫾ SD values of ␮static, Neq were consis-
tently highest in PBS, ranging from 0.34 ⫾ 0.06 to 0.58 ⫾
0.03 with increasing Tps; mean ⫾ SD values of ␮static, Neq
were consistently lowest in SF, ranging from 0.037 ⫾
0.008 to 0.22 ⫾ 0.02 with increasing Tps. Similarly, with
⬍␮kinetic, Neq⬎, mean ⫾ SD values were highest in PBS
(0.24 ⫾ 0.04) and lowest in SF (0.028 ⫾ 0.006). Equilib-
rium compressive load was not affected by test lubricants
(P ⫽ 0.15) with mean ⫾ SD ⌵eq ⫽ 2.7 ⫾ 0.6 N.
Effect of graded dilutions of SF. Friction coeffi-
cients were reduced by SF in a dose-dependent manner
(Figure 2). ␮static, Neq varied with test lubricant (P ⬍
0.001) and Tps (P ⬍ 0.001), with an interaction effect
(P ⬍ 0.001) (Figure 2A). Values of ␮static, Neq decreased
the boundary lubrication of articular cartilage. Shown are static
(␮static, Neq) (A) and kinetic (⬍␮kinetic, Neq⬎ [the brackets indicate that
the value is an average]) (B) friction coefficients in phosphate buffered
saline (PBS) and various concentrations of SF. In B, the presliding
duration was 1.2 seconds. Values are the mean and SEM (n ⫽ 4–8).
with increasing concentrations of SF at all Tps. At Tps ⫽
120 seconds, values of ␮static, Neq decreased from 0.40 ⫾
0.03 in PBS to 0.117 ⫾ 0.006 in 33% SF and 0.120 ⫾
0.006 in 100% SF. ⬍␮kinetic, Neq⬎ at Tps ⫽ 1.2 seconds
also varied with test lubricant (P ⬍ 0.001) (Figure 2B).
Values of ⬍␮kinetic, Neq⬎ decreased significantly with
increasing concentrations of SF, from 0.20 ⫾ 0.02 in PBS
to 0.11 ⫾ 0.02 in 10% SF (P ⬍ 0.01), 0.029 ⫾ 0.002 in
33% SF (P ⬍ 0.01), and 0.025 ⫾ 0.005 in 100% SF (P ⬍
0.01).
Effect of SF constituents alone. HA. Friction
coefficients were reduced by HA in a dose-dependent
manner (Figure 3). ␮static, Neq varied with test lubricant
(P ⬍ 0.001) and Tps (P ⬍ 0.001), with an interaction
effect (P ⬍ 0.01) (Figure 3A). Values of ␮static, Neq
886 SCHMIDT ET AL

of PRG4 at all Tps. At Tps ⫽ 120 seconds, values of

␮static, Neq decreased from 0.48 ⫾ 0.02 in PBS to 0.27 ⫾
0.03 in 450 ␮g/ml PRG4, with the value in SF being the
lowest at 0.12 ⫾ 0.02. ⬍␮kinetic, Neq⬎ at Tps ⫽ 1.2
seconds also varied with test lubricant (P ⬍ 0.001)
(Figure 4B). Values of ⬍␮kinetic, Neq⬎ decreased signif-
icantly with increasing concentrations of PRG4, from
0.23 ⫾ 0.02 in PBS and 0.20 ⫾ 0.01 in 4.5 ␮g/ml PRG4
to 0.10 ⫾ 0.02 in 450 ␮g/ml PRG4 (both P ⬍ 0.001),
which was greater than the value of 0.04 ⫾ 0.01 in SF
(P ⬍ 0.05).
SAPL. Friction coefficients were not affected by
SAPL (Figure 5). ␮static, Neq varied with test lubricant
(P ⬍ 0.001) and Tps (P ⬍ 0.001), with an interaction
effect (P ⬍ 0.001) (Figure 5A). Values of ␮static, Neq
appeared to decrease slightly with the addition of SAPL

Figure 3. Effect of graded concentrations of hyaluronan (HA) on the

boundary lubrication of articular cartilage. Shown are static (A) and
kinetic (B) friction coefficients in PBS, various concentrations of HA,
and SF. In B, the presliding duration was 1.2 seconds. Values are the
mean and SEM (n ⫽ 4–8). See Figure 2 for other definitions.

decreased with increasing concentrations of HA at all

Tps. At Tps ⫽ 120 seconds, values of ␮static, Neq decreased
from 0.46 ⫾ 0.02 in PBS to 0.22 ⫾ 0.02 in 3,300 ␮g/ml
HA, with the value in SF being the lowest at 0.11 ⫾ 0.01.
⬍␮kinetic, Neq⬎ at Tps ⫽ 1.2 seconds also varied with
test lubricant (P ⬍ 0.001) (Figure 3B). Values of
⬍␮kinetic, Neq⬎ decreased significantly with increasing
concentrations of HA, from 0.26 ⫾ 0.01 in PBS to
0.21 ⫾ 0.02 in 110 ␮g/ml HA (P ⬍ 0.05) and to 0.118 ⫾
0.009 in 3,300 ␮g/ml HA (P ⬍ 0.01), which was greater
than the value of 0.031 ⫾ 0.004 in SF (P ⬍ 0.01).
PRG4. Friction coefficients were reduced by PRG4
in a dose-dependent manner (Figure 4). ␮static, Neq varied Figure 4. Effect of graded concentrations of proteoglycan 4 (PRG4)
on the boundary lubrication of articular cartilage. Shown are static (A)
with test lubricant (P ⬍ 0.001) and Tps (P ⬍ 0.001), with and kinetic (B) friction coefficients in PBS, various concentrations of
no interaction effect (P ⫽ 0.72) (Figure 4A). Values PRG4, and SF. In B, the presliding duration was 1.2 seconds. Values
of ␮static, Neq decreased with increasing concentrations are the mean and SEM (n ⫽ 8). See Figure 2 for other definitions.
BOUNDARY LUBRICATION OF ARTICULAR CARTILAGE 887

with the addition of HA and/or PRG4 at all Tps. At

Tps ⫽ 120 seconds, values of ␮static, Neq decreased from
0.51 ⫾ 0.02 in PBS to 0.22 ⫾ 0.02 in HA and 0.23 ⫾
0.02 in PRG4 and to 0.18 ⫾ 0.01 in HA plus PRG4,
with the value in SF being the lowest at 0.130 ⫾
0.008. ⬍␮kinetic, Neq⬎ at Tps ⫽ 1.2 seconds also varied
with test lubricant (P ⬍ 0.001) (Figure 6B). Values of
⬍␮kinetic, Neq⬎ decreased significantly with the addition
of HA and/or PRG4, from 0.29 ⫾ 0.01 in PBS to 0.12 ⫾
0.01 in HA (P ⬍ 0.01) and 0.11 ⫾ 0.01 in PRG4 (P ⬍
0.01) and to 0.066 ⫾ 0.003 in HA plus PRG4 (P ⬍ 0.05
and P ⬍ 0.01, respectively), which was greater than the

Figure 5. Effect of surface-active phospholipids (SAPL) on the

boundary lubrication of articular cartilage. Shown are static (A) and
kinetic (B) friction coefficients in PBS, 200 ␮g/ml SAPL, and SF. In B,
the presliding duration was 1.2 seconds. Values are the mean and SEM
(n ⫽ 8). See Figure 2 for other definitions.

compared with PBS at all Tps. At Tps ⫽ 120 seconds,

values of ␮static, Neq ranged from 0.39 ⫾ 0.02 in PBS to
0.34 ⫾ 0.02 in 200 ␮g/ml SAPL, with the value in SF
being the lowest at 0.110 ⫾ 0.009. ⬍␮kinetic, Neq⬎ at
Tps ⫽ 1.2 seconds also varied with test lubricant (P ⬍
0.001) (Figure 5B). Values of ⬍␮kinetic, Neq⬎ did not
decrease significantly from 0.21 ⫾ 0.02 in PBS to 0.17 ⫾
0.01 in 200 ␮g/ml SAPL (P ⫽ 0.17), which was greater
than the value of 0.031 ⫾ 0.002 in SF (P ⬍ 0.001).
Effect of SF constituents in combination. Friction Figure 6. Effect of hyaluronan (HA), proteoglycan 4 (PRG4), and
coefficients were reduced by certain combinations of surface-active phospholipids (SAPL) in combination on the boundary
HA, PRG4, and SAPL at physiologic concentrations of lubrication of articular cartilage. Shown are static (A) and kinetic (B)
3,300 ␮g/ml, 450 ␮g/ml, and 200 ␮g/ml, respectively friction coefficients in PBS, 3,300 ␮g/ml HA, 450 ␮g/ml PRG4, 3,300
␮g/ml HA plus 450 ␮g/ml PRG4, 3,300 ␮g/ml HA plus 450 ␮g/ml
(Figure 6). ␮static, Neq varied with test lubricant (P ⬍ PRG4 plus 200 ␮g/ml SAPL, and SF. In B, the presliding duration was
0.001) and Tps (P ⬍ 0.001), with an interaction effect 1.2 seconds. Values are the mean and SEM (n ⫽ 4–8). See Figure 2 for
(P ⬍ 0.001) (Figure 6A). Values of ␮static, Neq decreased other definitions.
888
value of 0.024 ⫾ 0.003 in SF (P ⬍ 0.05). The addition
of SAPL did not significantly lower the value of
⬍␮kinetic, Neq⬎ (0.062 ⫾ 0.002) from its value in
HA plus PRG4 (P ⫽ 0.99).
DISCUSSION
The results described here indicate that SF con-
stituents contribute, individually and in combination,
both at physiologic and pathophysiologic concentra-
tions, to the boundary lubrication of apposing articular
cartilage surfaces. Normal SF functioned as an effective
boundary lubricant at the articular cartilage–cartilage
interface tested here (⬍␮kinetic, Neq⬎ ⫽ 0.025), even with
a 3-fold decrease in constituent concentration
(⬍␮kinetic, Neq⬎ ⫽ 0.029) (Figure 2B). Both HA and the
PRG4 preparation used here contributed independently
to a low ␮ value in a dose-dependent manner. Values of
⬍␮kinetic, Neq⬎ decreased from ⬃0.24 in PBS to 0.12 in
3,300 ␮g/ml HA (Figure 3B) and 0.11 in 450 ␮g/ml
PRG4 (Figure 4B). HA and PRG4 in combination
lowered the ␮ value further at these concentrations,
attaining a ⬍␮kinetic, Neq⬎ value of 0.066 (Figure 6B).
SAPL at 200 ␮g/ml did not significantly lower the ␮
value, either independently or in combination with HA
and PRG4. Collectively, these results suggest that the SF
constituents PRG4 and HA contribute individually and
in combination to the effective boundary lubrication of
articular cartilage.
The SF constituents used in the present study
were representative of those in native SF. SF is com-
posed of HA ranging from 2,000 kd to 10,000 kd (30,44),
PRG4 proteins ranging from ⬃14 kd to ⬃345 kd (4,45),
and SAPL of different types with DPPC being a major
form (33). HA has been shown to lubricate at a
cartilage–cartilage interface on a joint scale equally well
at 1,030 kd and 1,930 kd (14), and certain other prop-
erties of HA do not depend on molecular weight either
(in the range of 500–6,000 kd) (46). Therefore, the use
of SUPARTZ HA (average MW 800 kd) in the present
study was reasonable for studying the boundary-
lubricating ability of HA. The boundary-lubricating abil-
ity of PRG4 is associated with its large central mucin-
like domain (43), which is present in various forms of
PRG4 with MW ⬎⬃220 kd (25); therefore, an ⬃345-kd
form of PRG4 was prepared from conditioned media
and used. Finally, DPPC was chosen since it is the major
component of SAPL in SF (34).
Similar SF constituents have been used in several
previous studies examining their lubricating function
(12,18,21,23–26), and future studies may examine the
SCHMIDT ET AL
friction-lowering effects of other specific forms of SF
constituents. Additional studies examining the
structure–function relationship of SF constituents con-
tributing to boundary lubrication, both alone and in
combination, may provide the framework for the poten-
tial complete recapitulation of the boundary-lubricating
ability of whole SF. In the present study, the observed
friction-lowering effect of the constituents, used at phys-
iologic concentrations, suggests that they are sufficient
for much of the boundary lubrication of articular carti-
lage that is naturally mediated by SF. The friction
coefficient values, and their variation, were determined
here to focus on boundary lubrication mechanism, and
thus may not represent those values for human articular
cartilage in normal SF. Both normal human cartilage
and normal SF are difficult to obtain in a sufficient
quantity and controlled manner for the types of exper-
iments performed here.
The dose-dependent boundary-lubricating abili-
ties of SF, as well as those of PRG4 individually, are
consistent with (and extend) the findings of several
previous studies. Swann et al demonstrated a dose-
dependent effect of SF at a cartilage–glass interface
(23). Using the test configuration and protocol used in
the present study (7), SF was previously demonstrated to
be an effective boundary lubricant with a similar value of
⬍␮kinetic, Neq⬎ (⬃0.02). In the present study, the values
of ⬍␮kinetic, Neq⬎ in PBS (⬃0.24) were considerably
higher than those reported previously (⬃0.07) (7). As
has been noted (17), this can be attributed to the rinsing
of samples in PBS after harvest to remove residual SF
from the articular surface prior to testing in PBS. The
low values of ⬍␮kinetic, Neq⬎ in SF (⬃0.025) indicate that
the rinsing did not affect the ability of SF to effectively
lubricate the cartilage samples.
The dose-dependent effect of PRG4 is consistent
with findings in a previous study by Jay (24), using a
similar test configuration with a latex–glass interface,
in which lubrication function occurred at concentra-
tions ⬎200 ␮g/ml. However, the absolute value of
⬍␮kinetic, Neq⬎ in 450 ␮g/ml PRG4 observed here
(⬃0.10) (Figure 4B) is slightly more than the range of
␮ values reported in several other studies by Jay et al
(⬃0.047–0.018 in 250–400 ␮g/ml PRG4) (12,25,43).
This difference may be due to the way in which PRG4
interacts with native articular cartilage surfaces, used
in the present study, compared with other test sur-
faces. Nevertheless, the results of the present study
indicate that PRG4 contributes to the boundary lubri-
cation of articular cartilage, as previously concluded
from studies at a latex–glass interface. These contri-
BOUNDARY LUBRICATION OF ARTICULAR CARTILAGE
butions appeared specific to PRG4, since the HA and
SAPL content in the preparation was low, and control
proteins (albumins and globulins) at a physiologic
concentration did not independently lower the ␮ value
(data not shown).
The significant contribution of HA to the bound-
ary lubrication of apposed articular cartilage surfaces
reported here extends and clarifies the findings of
previous studies examining the lubricating ability of HA
with test protocols and/or configurations, particularly
where a boundary mode of lubrication was dominant.
Bell et al (16) demonstrated that Arthrease, a
fermentation-derived sodium hyaluronate with an MW
of 3,000 kd, functioned as an effective lubricant at a
cartilage–cartilage interface, but only under static con-
ditions in which the intrinsic biphasic lubrication was
depleted. Despite the absolute values of ␮static for both
HA and PBS being ⬃3-fold less than those reported
here, which may be attributable to differences in the test
configuration and protocols, the study by Bell et al and
the present study both show that HA contributes to
boundary lubrication. Similarly, Jay et al (12) demon-
strated that Healon, an uncrosslinked form of HA,
lowered the ␮ value from ⬃0.14 in PBS to ⬃0.07 at 3,340
␮g/ml, but not to the level in SF (⬃0.02), at a latex–glass
interface under boundary-lubricating conditions. This
trend is similar to that observed in the present study,
although again, the absolute value of ␮ is somewhat
different from the ⬍␮kinetic, Neq⬎ in 3,300 ␮g/ml HA
reported here (⬃0.12) (Figure 3B). Such differences
may be due to interactions of HA with test surfaces, as
postulated for PRG4 above, since the boundary lubrica-
tion function of HA is facilitated by binding to the test
surfaces (47).
Investigators in several other studies have re-
ported HA to be both effective (14,18) and ineffective
(15,17) as a boundary lubricant, using different whole-
joint test apparatuses in which several modes of lubrica-
tion were likely operative. The conflicting results of
those studies support the disposition that characteriza-
tion of a test configuration, surfaces, and mode of
lubrication is important when analyzing the mechanism
of boundary lubrication of articular cartilage. Accord-
ingly, the test configuration and protocol used in the
present study were characterized previously to achieve a
boundary mode of lubrication (7). Thus, HA does
appear to contribute, in a dose-dependent manner, to
the boundary lubrication of articular cartilage. Addi-
tional pilot studies indicated that HA adsorbed to the
articular surface of samples was able to contribute to
boundary lubrication even without HA in the test bath,
889
since samples soaked in HA, then rinsed and tested in
PBS, still had a low ␮ value. These studies suggest that
HA may function by being retained at or between the
articular cartilage surfaces under relative motion during
testing. Such adsorbed layers of HA at the articular
surface may have facilitated sliding (16), due to their
inherent slipperiness and ease of disentangling (48), and
therefore reduced friction between asperities in contact.
The results indicating that SAPL, in the form of
DPPC, do not significantly contribute to the boundary
lubrication of articular cartilage, either alone or in
combination with other SF constituents tested here,
provide additional insight into the controversial role of
SAPL as physiologic boundary lubricants of articular
cartilage. DPPC at ⬃0.35 mg/ml has been shown to
slightly lower friction at a cartilage–steel interface (26),
and phosphatidylcholine at 10 mg/ml dramatically re-
duced the ␮ value to ⬃0.016 (compared with a ␮ value
of ⬃0.028 in bovine SF) at a latex–glass interface under
boundary-lubricating conditions (13). In the present
study, SAPL in the form of DPPC at a physiologic
concentration of 200 ␮g/ml did not significantly lower
⬍␮kinetic, Neq⬎ alone (Figure 5B), with ⬍␮kinetic, Neq⬎
for DPPC remaining ⬃5-fold greater than that for SF, or
in combination with HA and PRG4 (Figure 6B). Addi-
tionally, because the PRG4 preparation tested in the
present study was free of SAPL in appreciable quantities
(⬍0.5 ␮g/ml), SAPL were not indirectly contributing to
the boundary-lubricating ability of PRG4, as has been
postulated as a complicating consideration (49).
However, the boundary-lubricating ability of ad-
ditional forms of SAPL (phosphatidylcholines, phos-
phatidylethanolamines, and sphingomyelins [6]) at a
cartilage–cartilage interface, in various combinations,
and the potential effect of the mode of delivery of the
SAPL, still remain to be determined. The state in which
endogenous lipids predominantly exist in SF (lamellae,
micelles, or vesicles) also remains to be determined,
although lamellar bodies have been detected by electron
microscopy (39).
HA and PRG4 synergistically lowered friction at
the cartilage–cartilage interface tested here, presumably
due to molecular interactions facilitating a molecular
distribution of shear at the articular surfaces. Jay et al
(21) previously reported that HA and PRG4 acted
synergistically, with HA enabling PRG4 to lubricate
under higher contact pressures at a latex–glass interface
under boundary-lubricating conditions. In that study,
hydrophobic attraction of PRG4 molecules along the
length of HA was suggested as a possible mechanism for
this interaction, although PRG4 does contain other
890
putative binding domains as well (25). In the present
study, the interaction between HA and PRG4 appeared
specific, since the subsequent addition of control pro-
teins (albumins and globulins) at a physiologic concen-
tration did not further lower the ␮ value (data not
shown). Collectively, these results suggest that the re-
pulsive force generated at the individual contact asper-
ities, on the articular cartilage surfaces, coated in HA
and PRG4 was greater than that at asperities coated in
either HA or PRG4 alone. This repulsive force may have
been provided by the structuring of water at the articular
surface by the aggregation and/or interaction of PRG4
and HA (5). Regardless of how the repulsive force is
generated, it may indirectly provide protection to chon-
drocytes from wear and mechanical disturbances in vivo
by reducing surface tissue shear.
The role of boundary lubrication relative to other
operative modes of lubrication mechanisms in vivo
remains to be fully elucidated. The lubrication mecha-
nisms associated with pressurized fluid, within cartilage
and between its surfaces, likely contribute substantially
to the low friction and low wear articulation within
synovial joints. Extremely low ␮ values, ⬃0.004–0.024,
have been reported (for natural articular cartilage with
SF as a lubricant) under test conditions in which bulk
fluid pressurization is significant (11,17) compared with
those presented here. However, with increasing loading
time and dissipation of hydrostatic pressure, the
lubricant-coated surfaces of articular cartilage bear an
increasingly higher portion of the load relative to pres-
surized fluid, and, consequently, ␮ values can become
increasingly dominated by the boundary mode of lubri-
cation (50,51). Indeed, ␮static, Neq increasing with increasing
Tps (as observed previously [7]) is suggestive of a time-
dependent interdigitation of surface molecules (52) at the
articular cartilage–cartilage interface. Early signs of artic-
ular cartilage wear have recently been associated with a
loss of the boundary lubrication function of SF postin-
jury (32). Accordingly, boundary lubrication has been
postulated to be critical to cartilage homeostasis by
facilitating low friction and low wear (2). Future studies
of the postulated role of boundary and other modes of
lubrication in arthritic disease are needed.
The collective results of this study provide insight
into the nature of the boundary lubrication of articular
cartilage by SF and its constituents. The maintenance of
the boundary-lubricating ability of SF, at the cartilage–
cartilage interface tested here, even at a 3-fold dilution,
suggests that the lubricant molecules in SF are normally
present in excess. However, the rapid decline in boundary-
lubricating ability with a further decrease in constituent
SCHMIDT ET AL
concentration suggests that such an alteration, which can
occur in the settings of both injury and disease (28–
30,32,33,53), can impair lubricating function. The combi-
nation of the SF constituents HA, PRG4, and SAPL at
physiologic concentrations approaching, but not fully rep-
licating, the boundary-lubricating ability of SF suggests that
additional lubricant molecules and/or complexes remain to
be identified. More than one specific molecule contributing
to the boundary lubrication of articular cartilage is not
particularly surprising given the variety of interactions that
can occur between the many molecules present in SF and
at the articular surface (for example, see refs. 5,21,54,55).
AUTHOR CONTRIBUTIONS
Dr. Sah had full access to all of the data in the study and takes
responsibility for the integrity of the data and the accuracy of the data
analysis.
Study design. Schmidt, Nguyen, Sah.
Acquisition of data. Schmidt, Gastelum, Nguyen, Schumacher.
Analysis and interpretation of data. Schmidt, Gastelum, Nguyen, Sah.
Manuscript preparation. Schmidt, Sah.
Statistical analysis. Schmidt, Sah.
REFERENCES
1. Ateshian GA, Mow VC. Friction, lubrication, and wear of articular
cartilage and diarthrodial joints. In: Mow VC, Huiskes R, editors.
Basic orthopaedic biomechanics and mechano-biology. 3rd ed.
Philadelphia: Lippincott Williams & Wilkins; 2005. p. 447–94.
2. Morrell KC, Hodge WA, Krebs DE, Mann RW. Corroboration of
in vivo cartilage pressures with implications for synovial joint
tribology and osteoarthritis causation. Proc Natl Acad Sci U S A
2005;102:14819–24.
3. Noyori K, Takagi T, Jasin HE. Characterization of the macromo-
lecular components of the articular cartilage surface. Rheumatol
Int 1998;18:71–7.
4. Schumacher BL, Block JA, Schmid TM, Aydelotte MB, Kuettner
KE. A novel proteoglycan synthesized and secreted by chondro-
cytes of the superficial zone of articular cartilage. Arch Biochem
Biophys 1994;311:144–52.
5. Jay GD. Lubricin and surfacing of articular joints. Curr Opin
Orthop 2004;15:355–9.
6. Sarma AV, Powell GL, LaBerge M. Phospholipid composition of
articular cartilage boundary lubricant. J Orthop Res 2001;19:671–6.
7. Schmidt TA, Sah RL. Effect of synovial fluid on boundary lubrication
of articular cartilage. Osteoarthritis Cartilage 2007; 15:35–47.
8. Davis WH, Lee SL, Sokoloff L. A proposed model of boundary
lubrication by synovial fluid: structure of boundary water. J Bio-
mech Eng 1979;101:185–92.
9. Malcom LL. An experimental investigation of the frictional and
deformational responses of articular cartilage interfaces to static
and dynamic loading [dissertation]. San Diego (CA): University of
California San Diego; 1976.
10. Jones ES. Joint lubrication. Lancet 1934;223:1426–7.
11. Charnley J. Lubrication of animal joints. In: Proceedings of the
Symposium on Biomechanics. London: Institute of Mechanical
Engineers; 1959. p. 12–22.
12. Jay GD, Haberstroh K, Cha CJ. Comparison of the boundary-
lubricating ability of bovine synovial fluid, lubricin, and Healon.
J Biomed Mater Res 1998;40:414–8.
13. Jay GD, Cha CJ. The effect of phospholipase digestion upon the
BOUNDARY LUBRICATION OF ARTICULAR CARTILAGE
boundary lubricating activity of synovial fluid. J Rheumatol 1999;
26:2454–7.
14. Mabuchi K, Obara T, Ikegami K, Yamaguchi T, Kanayama T.
Molecular weight independence of the effect of additive hyaluronic
acid on the lubricating characteristics in synovial joints with experi-
mental deterioration. Clin Biomech (Bristol, Avon) 1999;14:352–6.
15. Radin EL, Swann DA, Weisser PA. Separation of a hyaluronate-
free lubricating fraction from synovial fluid. Nature 1970;228:377–8.
16. Bell CJ, Ingham E, Fisher J. Influence of hyaluronic acid on the
time-dependent friction response of articular cartilage under
different conditions. Proc Inst Mech Eng [H] 2006;220:23–31.
17. Linn FC. Lubrication of animal joints. II. The mechanism. J Bio-
mech 1968;1:193–205.
18. Higaki H, Murakami T, Nakanishi Y, Miura H, Mawatari T,
Iwamoto Y. The lubricating ability of biomembrane models with
dipalmitoyl phosphatidylcholine and gamma-globulin. Proc Inst
Mech Eng [H] 1998;212:337–46.
19. Murakami T, Higaki H, Sawae Y, Ohtsuki N, Moriyama S, Nakanishi
Y. Adaptive and multimode lubrication in natural synovial joints and
artificial joints. Proc Inst Mech Eng [H] 1998;212:23–35.
20. Kumar P, Oka M, Toguchida J, Kobayashi M, Uchida E, Naka-
mura T, et al. Role of uppermost superficial surface layer of
articular cartilage in the lubrication mechanism of joints. J Anat
2001;199:241–50.
21. Jay GD, Lane BP, Sokoloff L. Characterization of a bovine
synovial fluid lubricating factor. III. The interaction with hyal-
uronic acid. Connect Tissue Res 1992;28:245–55.
22. Ikegawa S, Sano M, Koshizuka Y, Nakamura Y. Isolation, char-
acterization and mapping of the mouse and human PRG4 (pro-
teoglycan 4) genes. Cytogenet Cell Genet 2000;90:291–7.
23. Swann DA, Hendren RB, Radin EL, Sotman SL, Duda EA. The
lubricating activity of synovial fluid glycoproteins. Arthritis Rheum
1981;24:22–30.
24. Jay GD. Characterization of a bovine synovial fluid lubricating
factor. I. Chemical, surface activity and lubricating properties.
Connect Tissue Res 1992;28:71–88.
25. Jay GD, Tantravahi U, Britt DE, Barrach HJ, Cha CJ. Homology
of lubricin and superficial zone protein (SZP): products of
megakaryocyte stimulating factor (MSF) gene expression by hu-
man synovial fibroblasts and articular chondrocytes localized to
chromosome 1q25. J Orthop Res 2001;19:677–87.
26. Ozturk HE, Stoffel K, Jones CF, Stachowiak GW. The effect of
surface-active phospholipids on the lubrication of osteoarthritic
sheep knee joints: friction. Tribal Lett 2004;16:283–9.
27. Watterson JR, Esdaile JM. Viscosupplementation: therapeutic
mechanisms and clinical potential in osteoarthritis of the knee.
J Am Acad Orthop Surg 2000;8:277–84.
28. Mazzucco D, Scott R, Spector M. Composition of joint fluid in
patients undergoing total knee replacement and revision arthro-
plasty: correlation with flow properties. Biomaterials 2004;25:
4433–45.
29. Asari A, Miyauchi S, Sekiguchi T, Machida A, Kuriyama S,
Miyazaki K, et al. Hyaluronan, cartilage destruction and hydrar-
throsis in traumatic arthritis. Osteoarthritis Cartilage 1994;2:79–89.
30. Dahl LB, Dahl IM, Engstrom-Laurent A, Granath K. Concentra-
tion and molecular weight of sodium hyaluronate in synovial fluid
from patients with rheumatoid arthritis and other arthropathies.
Ann Rheum Dis 1985;44:817–22.
31. Schmid T, Lindley K, Su J, Soloveychik V, Block J, Kuettner K, et al.
Superficial zone protein (SZP) is an abundant glycoprotein in human
synovial fluid and serum. Trans Orthop Res Soc 2001;26:82.
32. Elsaid KA, Jay GD, Warman ML, Rhee DK, Chichester CO.
Association of articular cartilage degradation and loss of bound-
ary-lubricating ability of synovial fluid following injury and inflam-
matory arthritis. Arthritis Rheum 2005;52:1746–55.
33. Rabinowitz JL, Gregg JR, Nixon JE. Lipid composition of the
891
tissues of human knee joints. II. Synovial fluid in trauma. Clin
Orthop Relat Res 1984;(190):292–8.
34. Hills BA, Butler BD. Surfactants identified in synovial fluid and their
ability to act as boundary lubricants. Ann Rheum Dis 1984;43:641–8.
35. Bitter T, Muir HM. A modified uronic acid carbazole reaction.
Anal Biochem 1962;4:330–4.
36. Schmidt TA, Schumacher BL, Klein TJ, Voegtline MS, Sah RL.
Synthesis of proteoglycan 4 by chondrocyte subpopulations in
cartilage explants, monolayer cultures, and resurfaced cartilage
cultures. Arthritis Rheum 2004;50:2849–57.
37. Schumacher BL, Hughes CE, Kuettner KE, Caterson B, Aydelotte
MB. Immunodetection and partial cDNA sequence of the proteo-
glycan, superficial zone protein, synthesized by cells lining synovial
joints. J Orthop Res 1999;17:110–20.
38. Tanford C. The hydrophobic effect and the organization of living
matter. Science 1978;200:1012–8.
39. Hills BA. Oligolamellar lubrication of joints by surface active
phospholipid. J Rheumatol 1989;16:82–91.
40. Williamson AK, Chen AC, Masuda K, Thonar EJ, Sah RL. Tensile
mechanical properties of bovine articular cartilage: variations with
growth and relationships to collagen network components. J Or-
thop Res 2003;21:872–80.
41. Farndale RW, Buttle DJ, Barrett AJ. Improved quantitation and
discrimination of sulphated glycosaminoglycans by use of dimeth-
ylmethylene blue. Biochim Biophys Acta 1986;883:173–7.
42. Fung YC. Biomechanics: mechanical properties of living tissues.
New York: Springer-Verlag; 1981.
43. Jay GD, Harris DA, Cha CJ. Boundary lubrication by lubricin is
mediated by O-linked ␤(1-3)Gal-GalNAc oligosaccharides. Glyco-
conj J 2001;18:807–15.
44. Fraser JR, Laurent TC, Laurent UB. Hyaluronan: its nature,
distribution, functions and turnover. J Intern Med 1997;242:27–33.
45. Rhee DK, Marcelino J, Al-Mayouf S, Schelling DK, Bartels CF,
Cui Y, et al. Consequences of disease-causing mutations on
lubricin protein synthesis, secretion, and post-translational pro-
cessing. J Biol Chem 2005;280:31325–32.
46. Goldberg VM, Buckwalter JA. Hyaluronans in the treatment of
osteoarthritis of the knee: evidence for disease-modifying activity.
Osteoarthritis Cartilage 2005;13:216–24.
47. Tadmor R, Chen N, Israelachvili JN. Thin film rheology and
lubricity of hyaluronic acid solutions at a normal physiological
concentration. J Biomed Mater Res 2002;61:514–23.
48. Benz M, Chen N, Israelachvili J. Lubrication and wear properties
of grafted polyelectrolytes, hyaluronan and hylan, measured in the
surface forces apparatus. J Biomed Mater Res A 2004;71A:6–15.
49. Schwarz IM, Hills BA. Surface-active phospholipids as the lubri-
cating component of lubricin. Br J Rheumatol 1998;37:21–6.
50. Meyer E, Overney RM, Dransfeld K, Gyalog T. Nanoscience:
friction and rheology on the nanometer scale. River Edge (NJ):
World Scientific Publishing; 2002.
51. McCutchen CW. Boundary lubrication by synovial fluid: demon-
stration and possible osmotic explanation. Fed Proc 1966;25:1061–8.
52. Maeda N, Chen N, Tirrell M, Israelachvili JN. Adhesion and
friction mechanisms of polymer-on-polymer surfaces. Science
2002;297:379–82.
53. Young AA, McLennan S, Smith MM, Smith SM, Cake MA, Read
RA, et al. Proteoglycan 4 downregulation in a sheep meniscectomy
model of early osteoarthritis. Arthritis Res Ther 2006;8:R41.
54. Yingsung W, Zhuo L, Morgelin M, Yoneda M, Kida D, Watanabe H,
et al. Molecular heterogeneity of the SHAP-hyaluronan complex:
isolation and characterization of the complex in synovial fluid from
patients with rheumatoid arthritis. J Biol Chem 2003;278:32710–8.
55. Ghosh P, Hutadilok N, Adam N, Lentini A. Interactions of
hyaluronan (hyaluronic acid) with phospholipids as determined by
gel permeation chromatography, multi-angle laser-light-scattering
photometry and 1H-NMR spectroscopy. Int J Biol Macromol
1994;16:237–44.