вход по аккаунту


Interleukin-1 -stimulated invasion of articular cartilage by rheumatoid synovial fibroblasts is inhibited by antibodies to specific integrin receptors and by collagenase inhibitors.

код для вставкиСкачать
Vol. 40, No. 7, July 1997, pp 1298-1307
0 1997, American College of Rheumatology
Objective. To study the role of integrin receptors
in the invasion of cartilage by rheumatoid synovial
fibroblasts (RSF).
Methods. RSF were cocultured with cartilage
slices alone or in the presence of various potential
activators or inhibitors. The penetration of the cartilage
surface by RSF was determined by live-cell imaging of
fluorescent-labeled cells.
Results. Interleukin-lp (IL-1P) and IL-8 stimulated the RSF invasion of cartilage. Invasion was specific for RSF and required a concentration gradient of
IL-1P. The IL-lp-activated invasion of cartilage was
inhibited by anti-IL-1 antibodies, IL-1 receptor antagonist, and collagenase inhibitors. RSF invasion was also
inhibited by antibodies to a4, a5, aV, and Pl integrins.
Conclusion. In this study, an IL-1P concentration
gradient was required for RSF invasion into cartilage,
raising the possibility that in vivo invasion may be
induced by IL-1P released by chondrocytes. The IL-lP
activation of RSF assayed in vitro may contribute to the
RSF invasion of cartilage in vivo. Cartilage invasion
requires the availability of pl and a4, a5, and aV
integrins and the presence of collagenase activity.
The invasion of cartilage by synovial pannus is a
characteristic feature of rheumatoid arthritis (RA) (1).
The processes involved are poorly understood. RheumaSupported by grants from the Western Pennsylvania Hospital
0682, and facilities in the National Science Foundation Science and
Technology Center at Carnegie Mellon University (DIR-8920118).
Allan 2. Wang, MD, PhD, Jane C. Wang, BS, Herbert S.
Diamond, MD: Western Pennsylvania Hospital, Pittsburgh; Gregory
W. Fisher, PhD: Carnegie Mellon University, Pittsburgh,
Address reprint requests to Allan Z. Wang, MD, PhD,
Department of Medicine, Western Pennsylvania Hospital, 4800
Friendship Avenue, Daly Building 233, Pittsburgh, PA 15224.
Submitted for publication August 22, 1996; accepted in
revised form February 14, 1997.
toid synovial fibroblasts (RSF) are a major component
of pannus tissue. RSF extend out over the articular
surface of cartilage, including the point of contact between the invading pannus and the cartilage surface (2).
Destruction of underlying cartilage matrix occurs at the
sites of cartilage-pannus contact (2). We hypothesized
that RSF that had been activated by mediators released
from mononuclear cells were capable of invading
The double-compartment transwell model has
been previously used to study tumor cell invasion ( 3 ) and
RSF chemotaxis of, adherence to, and invasion of
cartilage-associated matrix (4,5). We adapted this
double-compartment transwell system to create a model
suitable for in vitro studies of RSF invasion of cartilage.
The invasion of cartilage by RSF may require
binding of cells to the cartilage matrix, with mediation by
integrin receptors. RSF that are located at the cartilagepannus junction express the integrin subunits pl, 014, and
a5 (6,7). We hypothesized that the process of invasion of
cartilage by RSF requires specific interactions between
RSF membrane integrin receptors and cartilage matrix
proteins. We present evidence that RSF penetration of
cartilage can be activated by interleukin-1p (IL-lp), which
is a cellular mediator known to occur in rheumatoid
synovial lining. We also show that the invasion of cartilage
by RSF may require binding of cells to the cartilage matrix
mediated by pl integrins and the fibronectin-associated
cell surface receptor a subunits a4, a5, and aV.
Cell culture. Synovium was obtained, with consent,
from patients undergoing knee replacement surgery. Twelve
synovial membranes were obtained from patients with RA and
8 from patients with osteoarthritis (OA). The diagnoses were
based upon the criteria of the American College of Rheumatology (formerly, the American Rheumatism Association)
(8,9). Synovial tissues were washed, dissected free of fat,
fibrous, and elastic tissues, then cut into 1-2-mm pieces. The
pieces were plated onto 100-mm Petri dishes containing 10 ml
of Dulbecco's modified Eagle's medium (DMEM) plus 10%
fetal bovine serum (FBS) and antibiotics, including 100
unitsiml penicillin G, 100 pdml streptomycin sulfate, and 250
ngiml amphotericin B (Gibco BRL, Gaithersburg, MD), in
humidified air containing 5% CO, at 37°C. The media were
changed twice a week until the primary synovial cultures were
For subcultures, cells were released from Petri dishes
or flasks by brief trypsinization with a mixture of 0.05% trypsin
and 0.53 mM EDTA for 5-10 minutes at 37"C, then washed,
counted, and incubated with DMEM containing 10% FBS.
Cells harvested from the second passage were used for experiments, except where specifically noted. The normal synovial
cell line (HIG-82) was supplied by American Type Culture
Collection (Bethesda, MD). Human dermal fibroblasts were
provided by the Department of Pathology, University of
Pittsburgh School of Medicine.
Preparation of cartilage slices. Fresh bovine cartilage
was obtained from knee joints of newborn calves (10 days old),
which were provided by Rendalic Packing Co. (a slaughterhouse in Mckeesport, PA). The cartilage surface was trimmed
to round pieces with a diameter of 5 mm using a cork borer.
The cartilage was sliced with a Tissue-TED microtome (Ames,
Elkhart, IN) to 30-pm thin slices. The integrity of the cartilage
slices was checked by inverted phase-contrast microscopy (20X
We chose 30-pm sections based upon preliminary
experiments. We found that this thickness of cartilage enabled
us to reproducibly obtain high-quality sections of similar size
that did not leak and had no perforations or defects. We chose
to extract these slices from within 150 pm of the upper
cartilage surface, since matrix components such as type I1
collagen fibrils and fibronectin are distributed more densely in
the upper layer of the cartilage than in the deeper layer, which
is composed largely of glycosaminoglycans (2,lO). Furthermore, the superficial layer of human articular cartilage is more
susceptible to IL-1-induced damage than is the deeper layer (11).
Cartilage slices were sterilized for 4 hours with DMEM
containing 200 unitsiml penicillin G, 200 pdml streptomycin
sulfate, and 500 ng/ml amphotericin B (Gibco BRL). Some
cartilage specimens were suspended in DMEM containing
10% DMSO, and stored frozen in liquid nitrogen. Most studies
were performed using bovine cartilage. In addition, we compared the results of IL-lp-induced RSF invasion of cartilage
using bovine cartilage with the results obtained using human
cartilage (from patients undergoing knee replacement surgery). The preparation of human cartilage was identical to that
described for bovine cartilage. Only cartilage slices that appeared normal by both macro- and microscopic examination
were used.
Measurement of invasion of cartilage by RSF. A
30-pm slice of cartilage, prepared as described above, was
placed on the top of a filter with an 8-pm pore size in the upper
chamber of a 2-compartment transwell (Costar, Cambridge,
MA). The edges of the cartilage slice were immobilized with a
ring made from parafilm with a central hole (diameter of 3.1
mm), and further sealed with 3% agar (Figure 1). The agar was
dissolved in DMEM by boiling with a microwave device, and
then cooled down to 45°C by immersion in a water bath preset
at a constant temperature of 45°C. A 3 mm-diameter plastic
column was placed at the center of the cartilage slice, just
touching the inner rim of parafilm. Then, 60 pl of agar solution
was carefully pipetted to the upper well in the space surrounding the column. The column was withdrawn when the agar was
semisolidified, leaving a reproducible center area of cartilage
slice clear of agar for the examination of RSF invasion in
RSF (5 X lo4)were prestained with 3-5 pA4 of the vital
fluorogenic dye, Calcein-fluorescein isothiocyanate (FITC)
(Molecular Probe, Eugene, OR), for 30 minutes. Then, RSF
suspended in 0.2 ml of DMEM containing 1% bovine serum
albumin (BSA) were seeded onto the surface of each cartilage
slice, and 0.5 ml of media, with or without potential activators
of invasion, was added to the lower chamber. IL-1P, IL-2, IL-6,
IL-8, granulocyte-macrophage colony-stimulating factor (GMCSF), transforming growth factor P l (TGFpl), or tumor
necrosis factor a (TNFa) (UBI, Lake Placid, NY) were added
to test their potential as activators of synoviocyte invasion of
cartilage. Potential inhibitors were added to both upper and
lower chambers. The anti-human antibodies to IL-16 and the
integrin receptors, including al, a2,a3,a4, a5, a6, aV, and pl,
were obtained from Boehringer Mannheim (Indianapolis, IN)
or Gibco BRL, and added at dilutions of 1:200-1:400. The
purified RGD derivatives (Gibco BRL), GRGDSP (Gly-ArgGly-Asp-Ser-Pro) or GRGNSP (Gly-Arg-Gly-Asp-Asn-Pro),
were added at 0.1-1.0 mM. Individual protease inhibitors,
including anti-pain dichloride, aprotinin, phenylmethylsulfonyl
fluoride (PMSF), E-64, leupeptin, 1,lO-o-phenanthroline
(PNT), and phosphoramidon, were added at 10-150 pl4 in
both chambers to test their effect on RSF invasion. Inhibitors
were added to both upper and lower chambers to increase
opportunities to observe a response. We were concerned that
inhibitors added to the upper chamber might not reach the
RSF-matrix interface. These experiments could not be designed to test the effectiveness of chondrocytes.
The optimal concentrations of the tested activators and
inhibitors were predetermined for each assay. As negative
controls, 1% BSA and mouse or rabbit IgG were used at the
same dilutions. RSF and cartilage slices were cocultured in a
37°C humidified incubator containing 5% CO,. The transwells
were rocked during the first 10 minutes of incubation. After
incubation of the RSF-cartilage coculture for 24 hours or
longer, the numbers of fluorescent-stained RSF appearing on
the lower surface of the transwell filter were counted as
invading cells using live-cell imaging microscopy. The invading
cells were observed and quantified by an extensively remodeled Zeiss Axiovert inverted fluorescent microscope (Carl
Zeiss, Jena, Germany), equipped with a constant-stage temperature controller, a robot-operated stage, a high-resolution
cooled charge coupled device camera, and image processing
software. Fluorescent images were taken using epifluorescence by excitation at a wavelength of 480 nm and
emission at a wavelength of 515 nm. Sixty-four to 121 randomly
selected microscopic fields (0.05-0.2 mm2 per field) of each
cartilage slice were scanned, and the number of vital-stained
RSF present was visually counted. The cell images were shown
on both video and computer monitors operated by a software
program from Detective Biological System (Pittsburgh, PA).
Of the vital fluorescent stains tested, Calcein-
Vital fluorogenic dye (Calcein-FITC)
pre-stained RSF 0 7
! (30
Figure 1. A model system for the in vitro study of invasion of cartilage
by rheumatoid synovial fibroblasts (RSF). A 30-pm thin section of
cartilage was placed on top of a filter in the upper chamber of a
double-compartment transwell. The cartilage section was immobilized
with a layer of parafilm ring, which was covered and sealed with 3%
agar. RSF prestained with Calcein-fluorescein isothiocyanate (FITC),
a vital fluorogenic dye, were placed on the cartilage surface and
cocultured in serum-free media, and 0.5 ml of media, with or without
potential activators of invasion, was added to the lower chamber.
Potential inhibitors were added to both chambers. After coincubation
of the RSF and cartilage slice for 24 hours, the number of fluorescent
RSF appearing on the lower surface of the filter was counted visually
using live-cell imaging microscopy. Sixty-four to 121 randomly selected
microscopic fields of 0.05-0.2 mmz were scanned at intervals of 0.5
mm. IL-IP = interleukin-lp.
acetoxymethyl, an FITC-conjugated vital fluorogenic dye, was
the best for this purpose. Preloading of 3-5 pA4 Calcein-AM
for 30-60 minutes at 37°C provided optimal labeling of the
RSF. The fluorescent labeling could be maintained for up to 72
hours, and was retained for another 48 hours when the samples
were fixed with 3.8% freshly prepared paraformaldehyde.
Under the conditions described above, Calcein-AM did not
affect cell growth, viability, or plating efficiency, and did not
label cartilage matrix or chondrocytes.
To study the effect of an IL-1p gradient, 4 ng/ml IL-lp
was added to either the upper chamber or the lower chamber
of the transwell for 12 hours, and then the chamber was
washed free of IL-lp. Alternatively, the cartilage slice
mounted on the transwell filter was pretreated with 4 ngiml
IL-10 added either to the lower or to the upper chamber for 1
hour, and then washed free of IL-1p. Assays for the RSF
invasion of cartilage slices were then performed as described
Demonstration of invasion of cartilage by RSF. To
directly observe the penetration of RSF into cartilage, chondrocytes were counterstained with 1 pA4 Hoechst dye (Molecular Probe), and 0.5-pm (in diameter) rhodamine latex beads
(Sigma, St. Louis, MO) were spread on the cartilage to locate
its surface. The invasion of cartilage slices by RSF was
demonstrated by scanning with a confocal laser microscope
focused downward from the cartilage surface. Separate XZ
scanning images for RSF, chondrocytes, and beads on the
cartilage surface were overlaid as 1 image using 3 pseudocolor
parameters. For comparison, histologic examination was performed. Cocultured RSF-cartilage samples were embedded in
LR white (Ted Pella, Redding, CA), then 10-pm sections were
cut and stained with toluidine blue (0.05%) for 15 minutes.
Assays for fibroblasts other than RSF. Second-passage
cultured human osteoarthritic synoviocytes (OAS), a longterm cultured normal synovial cell line, and second-passage
cultured human dermal fibroblasts were also studied. The
results were compared with those obtained using RSF.
Statistical analysis. All values were reported as the
mean t SEM obtained from at least triplicate experiments.
Statistical comparison was performed by Student’s or Welch’s
unpaired t-test.
Activation of RSF to penetrate cartilage slices.
IL-lP induced passaged RSF to penetrate into or
through cartilage slices, as demonstrated by both confo-
Figure 2. Three-dimensional scanning of invasion of cartilage by
rheumatoid synovial fibroblasts (RSF) using confocal laser microscopy. To directly observe the penetration of RSF into cartilage, in
addition to prestaining of RSF with Calcein-fluorescein isothiocyanate
(emitted green fluorescence), chondrocytes (C) were counterstained
with Hoechst dye (emitted blue fluorescence), and rhodamineconjugated 0.5-pm latex beads (b) (emitted red fluorescence) were
spread on the cartilage to mark the surface. The invasion of RSF into
a cartilage slice was demonstrated by scanning with a confocal laser
microscope, focused downward from the cartilage surface. A, Noninvasive RSF in a bovine serum albumin (BSA)-treated control (bar =
20 pm). B, Typical invasive RSF stimulated by 4 ng/ml interleukin-lp
(IL-1p) (bar = 20 pm). For comparison, C shows noninvasive RSF in
a BSA-treated control by histology (bar = 40 pm), and D shows the
IL-lp-stimulated invasion of RSF by histology (bar = 40 pm).
T **
Figure 3. Concentration-dependent effect of interleukin-lp (IL-1p)
on invasion of cartilage by rheumatoid synovial fibroblasts, with the
maximal effect at 4 ngiml.
o L
Upper pmtreaed
IL-Ip Gradient toward
Cartilage Surface
Figure 5. Requirement of an interleukin-lp (IL-1p) gradient toward
cartilage for the activation of invasion of cartilage by rheumatoid synovial
fibroblasts (RSF). A, Activation of RSF penetration of cartilage occurred
only when IL-lp (4 ngiml) was added to the lower transwell chamber
during the assay. No significant activation of RSF penetration of cartilage
occurred when IL-lp (4 ngiml) was added to the upper chamber o r when
RSF were preincubated with 4 ngiml IL-1p for 12 hours and then
washed free of IL-lp before addition to the upper chamber. B, To
establish an IL-lP gradient in the cartilage with the highest concentration at the lower cartilage surface, cartilage slices were preincubated
with IL-1p in the lower chamber and the media were then washed free
of IL-lp, producing a penetration of 3.9 ? 1.2 cellsimm‘ (mean z
SEM) in the cartilage surface. When the IL-Ip gradient was reversed,
with the highest concentration at the upper cartilage surface, by
preincubation of cartilage slice with IL-lp in the upper chamber, no
invasion occurred. *P < 0.05, ** = P < 0.01, compared with bovine
serum albumin (BSA)-treated control. Bars show the mean and SEM.
Figure 4. Effects of interleukins (IL) and growth factors on the
invasion of cartilage by rheumatoid synovial fibroblasts (RSF). The
penetration of RSF into cartilage could be reproducibly demonstrated
following IL-10 activation. When unstimulated second-passage RSF
were added to the upper chamber, 1.0 t 0.1 cellsimm’ (mean 2 SEM)
penetrated the cartilage slice. I L - l p at 4 ngiml increased the number
of RSF, which penetrated to 3.3 -+ 0.3 cellsimm’. ** = P < 0.001
compared with unstimulated control. Similar stimulation of RSF
penetration was also induced by IL-8 (4 ngiml), but was not induced by
IL-2 (4 ngiml), IL-6 (4 ngiml), granulocyte-macrophage colonystimulating factor (GM-CSF) (25 ngiml), transforming growth factor
pl (TGF-p1) (20 ngiml), or tumor necrosis factor 01 (TNF-a) (4
ngiml). All concentrations tested for individual cellular mediators were
the optimal concentrations predetermined in preliminary study. Bars
show the mean and SEM. BSA = bovine serum albumin.
cal laser scanning microscopy with vital fluorescent
staining (Figures 2A and B) and by conventional light
microscopy (Figures 2C and D). The IL-1p effect was
concentration dependent, with the maximal effect at 4
ng/ml (Figure 3). In the absence of IL-lp activation,
1.0 2 0.1 RSF/mm2 (mean 2 SEM) penetrated the
cartilage slice (Figure 4). When 4 ng/ml IL-1p was
placed in the lower chamber, the mean number of RSF
that penetrated the cartilage increased to 3.3 t 0.3
RSF/mm2 ( P < 0.001 compared with BSA-treated
Similar stimulation of RSF penetration was also
seen with IL-8 at 4 ng/ml (3.3 ? 0.4 RSF/mm2; P <
ILlB I L l h I L l r a IL8
IL-16 IL-lP + IL-1Ab
Figure 6. A, Effects of interleukin-1 receptor antagonist (IL-lra) on IL-1p- or IL-&induced invasion of cartilage by rheumatoid synovial fibroblasts
(RSF). IL-lra at 100 ngiml significantly inhibited both I L - l P and IL-%induced RSF invasion. The invading cells decreased from 3.3 i 0.3 to 0.7 ?
0.3 cellsimm’ (mean 2 SEM) when RSF were treated with the combination of IL-lp (4 ngiml) and IL-lra (100 ngiml). IL-lra had a similar inhibitory
effect on IL-8-induced RSF invasion of cartilage. ** = P < 0.05 compared with IL-10 or IL-8 alone. B, Effects of IL-1 antibody (Ab) on
IL-1/3-induced RSF invasion of cartilage. IL-1 antibody at a dilution of 1:200 significantly inhibited IL-lp-induced RSF invasion from 4.77 2 1.58
to 1.2 2 0.4 cells/mm’ ** = P < 0.05. Bars show the mean and SEM. BSA = bovine serum albumin.
0.001). Cartilage penetration was not significantly increased by the addition of 4 ng/ml IL-2 (1.6 +- 0.2
RSF/mm2; P > 0.05), 4 ng/ml IL-6 (1.6 t 0.4 RSF/mm2;
P > 0.05), 25 ng/ml GM-CSF (2.4 5 0.6 RSF/mm2; P >
0.05), 20 ng/ml TGFpl (1.3 5 0.4 RSF/mm2; P > 0.05),
or 4 ng/ml TNFa (2.4 +- 0.8 RSF/mm2; P > 0.05) (Figure
4). In timing studies, at least 18 hours of coincubation of
RSF and cartilage was required to identify RSF penetration of the cartilage slice (2.9 t 0.1 RSF/mm2) when
4 ng/ml IL-1p was added to the lower chamber.
Effect of IL-1P on invasion of the cartilage slices
by other fibroblasts. IL-lp activation of penetration of
cartilage was specific for RSF. IL-1p did not activate
penetration of cartilage by osteoarthritic synoviocytes,
normal synoviocytes, or dermal fibroblasts. Unstimulated second-passage cultured OAS showed no penetration of cartilage, and addition of IL-1p had no significant
effect on OAS penetration (0.2 2 0.1 cells/mm2). Fibroblasts that were derived from cultures of RSF maintained for >10 passages, from a long-term passage
normal synovial cell line, or from second-passage human
dermal fibroblasts did not invade cartilage and could not
be activated to invade with IL-1p.
Requirement of an IL-1P gradient toward cartilage for activation of RSF invasion of cartilage slices.
Activation of RSF penetration of cartilage occurred only
when IL-lP was added to the lower transwell chamber.
No activation of RSF penetration of cartilage occurred
when IL-1p was added to the upper chamber (IL-1p
0.9 +- 0.2 RSF/mm2 versus BSA-treated control 1.1 f 0.1
RSF/mm2; P > 0.05) or when RSF were preincubated
with 4 ng/ml IL-1p for 12 hours and then washed free of
IL-1p before addition to the upper chamber (IL-1p
1.3 +- 0.3 RSF/mm2 versus BSA-treated control 1.1 +0.04 RSF/mm2; P > 0.05) (Figure 5 ) . When cartilage
slices were preincubated with IL-1p in the lower chamber, so as to establish a gradient for IL-1p in the
cartilage, and the media were then washed free of IL-lp,
3.9 ? 1.2 RSF/mm2 penetrated the cartilage surface
(P < 0.05 compared with BSA-treated control). When
the IL-1p gradient was reversed to be highest at the
upper cartilage surface, no invasion occurred (1.7 +- 0.7
1 6
< A t t l h d l a to Ink@% Sobmmltr,
< I L l ~ p l u sAnUbodiataIntqrinsvbonlb~
Figure 7. Effect of antibodies to integrins on invasion of cartilage by
rheumatoid synovial fibroblasts (RSF). Activation of RSF penetration
by interleukin-1P (ILlp) (4 ngiml) was inhibited by antibodies (1:400)
to the integrins pl (inhibited by 70%), a4 (by 47.4%), a5 (by 40.2%),
and aV (by 63.1%), but not by antibodies to a l , a2, a3, or a6 integrins.
* = P < 0.05 and ** = P < 0.01, compared with ILlp alone. None of
the antibodies to integrin subunits showed a significant effect on
unstimulated RSF (all P > 0.05 when compared with the bovine serum
albumin [BSAI-treated control). Bars show the mean and SEM.
RSF/mm2; P > 0.05 compared with BSA-treated control) (Figure 5).
Inhibitors of IL-lp-activated RSF invasion of
cartilage. The activation of RSF invasion stimulated by
IL-1p was inhibited by IL-1 receptor antagonist (IL1Ra). IL-1Ra at 100 ng/ml completely inhibited the
IL-16 A-Pain
- 1.5
__ 1.0
- 0.5
IL-16 >
Figure 9. Complete inhibition of interleukin-lp (IL-1p)-stimulated
rheumatoid synovial fibroblast (RSF) penetration of cartilage by both
1,lO-o-phenanthroline (PNT) and phosphoramidon (Ramid) at 10 f l ,
with no significant effect seen with anti-pain (A-Pain), a serine
protease, or leupeptin (Leup), a cysteinyl protease. * = P < 0.05; ** =
P < 0.01, compared with IL-lP alone. Bars show the mean and SEM.
stimulation of RSF invasion induced by 4 ng/ml IL-1p
(IL-1p + IL-1Ra 0.7 t 0.3 RSF/mm2; P < 0.001
compared with IL-1p alone [3.3 ? 0.3 RSF/mm2] and
P = 0.3875 compared with BSA control [1.0 A 0.1
RSF/mm2]) (Figure 6A). IL-1p-activated RSF invasion
was inhibited by a 1:200 dilution of anti-IL-lp antibody
(1.2 % 0.4 RSF/mm2) (Figure 6B).
Antibodies to integrin receptors. The antibodies
to p l , a4, a5, and a V integrins inhibited RSF invasion of
cartilage. The lowest dilution at which inhibition was
detected was 1:400. Nonspecific IgG of the same species
did not effect RSF invasion. IL-lp-activated RSF penetration of cartilage slices was reduced by 47.4% by
anti-a4 antibody, 40.2% by a n t i d antibody, 63.1% by
anti-aV antibody, and 70.0% by anti-pl antibody (P <
0.05 for all comparisons with controls). At the same
dilution, IL-lp-activated RSF penetration was not inhibited by the antibodies to a l , a2, a3, or a6 integrins
(Figure 7). None of the antibodies to integrin subunits
significantly inhibited unstimulated RSF. The RGD
derivative, 1 mM GRGNSP, significantly inhibited RSF
invasion of cartilage (1.1 t 0.4 RSF/mm2).
Protease inhibitors. The collagenase inhibitors PNT
and phosphoramidon markedly inhibited IL-1@stimulated
RSF penetration of cartilage in a concentration-dependent
manner (Figure 8). The minimal effective concentration
of these 2 inhibitors was 10 pM, at which IL-1pstimulated RSF penetration was reduced from 4.4 ? 0.9
RSF/mm2 to 1.5 f 0.6 RSF/mm2 by PNT and to 1.1 k
0.3 RSF/mm2 by phosphoramidon (P < 0.05 and P <
0.01, respectively, compared with IL-1p alone) (Figure
9). PNT and phosphoramidon inhibited IL-8-induced
RSF invasion of cartilage at concentrations similar to
those that inhibited induction by IL-1p. Neither the
serine protease inhibitors PMSF, aprotinin, and antipain dichloride, nor the cysteinyl protease inhibitors
E-64 and leupeptin blocked RSF penetration of cartilage
at concentrations lower than 100 @.
IL-lp, when added to the lower transwell chamber, reproducibly induced the penetration of RSF into
cartilage. IL-1 activates both chondrocytes (12) and
synoviocytes to synthesize and release proteases (13). It
induces decreased synthesis and increased degradation
of cartilage matrix-associated components, including
collagens and proteoglycans (14,15). The RSF in pannus
synthesize and release IL-1 and metalloproteinases in
response to cytokine activation (16). IL-1 stimulates
RSF adherence to cartilage explants, which is further
enhanced by pretreating cartilage with collagenase (17).
IL-l-activated RSF are able to penetrate and degrade
cartilage-associated matrix. This effect is augmented by
RA synovial fluid, IL-lp, or TGFPl (18). The invasiveness of the reconstituted matrices is positively correlated
with expression of interstitial collagenase in pannus
tissue and with clinical progression of RA in patients
from whom the RSF are obtained (19). We observed
that IL-lp-activated RSF penetrated cartilage segments
at IL-1p concentrations similar to those required to
induce degradation of cartilage proteoglycans. Thus, the
IL-1 effect we assayed in vitro could contribute to RSF
invasion of cartilage in vivo.
IL-8 also stimulated RSF penetration, but IL-2,
IL-6, GM-CSF, TGF, and TNFa did not. IL-8 is released
from articular cartilage following induction by IL-1p
(20). IL-8 is present in most RA synovial lining cells and
synovial fluid (21). More synovial fluids from RA patients contain IL-8 than do those from OA patients (22),
while only a few normal synovial fluids contain IL-8 (23).
In the pannus-junction tissues, the concentration of IL-8
is 3 times greater in RA than in OA tissues (23). The
IL-1 effect could result in part from IL-l-stimulated
release of IL-8. However, we did not test this possibility.
Tuan et a1 and Frye et a1 reported that, in the presence
of IL-1, a greater number of RSF penetrated cartilage-
associated matrix (lSJ9). In contrast, TNFa decreased
RSF invasion of matrix (18,19). In one study, TNFa was
found to increase RSF binding to cartilage fragments
(24). It induces proteinase production by connective
tissue cells, and thereby can contribute to cartilage
matrix destruction in RA (25). However, TNFa is much
less potent than IL-1, and it may only augment the
production or the effect of IL-1 (26). RSF produce very
little TNFa (27).
The results using GM-CSF and TNFa, which,
although not statistically significant, did produce an
increase in RSF invasion, are partly due to the large
variances. In addition, most data obtained from these
studies have been statistically analyzed by Student’s
t-test, which assumes that the populations have equal
standard deviations. However, since the standard deviations using GM-CSF or TNFa were significantly different from those obtained with the controls, we used
Welch’s test, which produced results of lower significance than did Student’s t-test. The studies to detect
induction of RSF invasion by GM-CSF and TNF in this
system do not exclude the possibility that these cytokines, in combination with each other or other cytokines
such as IL-1, might contribute to the process in vivo. The
synergistic effects of a combination of cytokines on RSF
invasion will be explored in future studies using this
A great proportion of RA patients develop erosive joint disease. Invasion of cartilage by synoviocytes,
including RSF, and inflammatory mononuclear cells is
characteristic of RA. Normal synoviocytes do not adhere
to the cartilage surface, whereas in RA, at the point of
contact between pannus and cartilage, RSF adhere
directly to the cartilage surface (17). Destruction of
underlying cartilage matrix occurs at the sites of
cartilage-pannus contact (28). In the assays using this
model, we did not induce invasion of cartilage by other
fibroblasts, such as osteoarthritic synoviocytes, normal
synoviocytes, or dermal fibroblasts treated with IL-1p.
We speculate that RSF may have developed an invasive
phenotype in vivo through exposure to various mononuclear cell mediators including IL-1. IL-1 is detected in
a greater percentage of RA synovium than in OA or
normal synovium (29), and RSF synthesize and release a
greater concentration of IL-1 into synovial fluid than do
OAS (30).
Activation of RSF penetration of cartilage occurred only when IL-1p was present in the lower transwell chamber during the assay. No significant activation
of RSF penetration of cartilage occurred when IL-lp
was added to the upper chamber or when RSF were
preincubated with IL-1p and then washed free of it prior
to addition to the upper chamber. A gradient of IL-1p
inside cartilage, with the lowest concentration at the
upper cartilage surface, also induced RSF invasion.
When the gradient was reversed, with the highest concentration at the upper cartilage surface, no invasion
occurred. Therefore, the IL-1p induction of RSF invasion of cartilage requires a gradient of IL-1p from the
interior of the cartilage to the surface. This suggests that
a chemotactic effect may contribute to the invasive
process. Such a gradient could occur in vivo if mediators
released by mononuclear cells or RSF induce chondrocytes to release IL-1p in RA. We have previously
reported that both IL-1p and IL-8 are chemotactic for
RSF, as shown both in a double-compartment transwell
and with a live-cell imaging assay, which could distinguish directed migration from random migration (4).
The effect of IL-1 is regulated by IL-lRa, a
specific receptor antagonist of IL-1, which binds to the
IL-1 receptor without initiating an intracellular response
(31). The ratio of IL-1Ra to IL-1 is decreased in RA
(32). Cultured RSF contain intracellular IL-lRa, but
secrete little IL-1Ra into the culture supernatant (33). In
this study, IL-1Ra significantly inhibited IL-1pstimulated RSF penetration of cartilage. This suggests
that the IL-1p activation of invasion involves the activation of IL-1 receptors.
Integrins are the transmembrane receptors for
cell adhesion (34,35). Integrins are heterodimers and are
usually composed of an a and a p subunit. The p l
subunit can noncovalently link with different a units to
form receptors for collagen, laminin, fibronectin, and
vitronectin. The pl integrin subunit is abundantly expressed in synovial lining cells (36). a2p1, a4p1, a5p1,
and a6pl integrins have been detected in the normal
synovial membrane, although there are discrepancies
regarding these findings among different reports (36,37).
In severely inflammatory synovium, including RA synovium, both macrophage-like and fibroblast-like synoviocytes express a3, a4, a5, and a6 integrin subunits (6,7).
IL-lp, TNFa, or interferon-? (IFNy) can induce expression of a1 integrin, which is not expressed in unstimulated synoviocytes. IL-1p and IFNy can also up-regulate
the expression of a3 and a5 integrins (6). TNFa can
further enhance the effects induced by IL-10 (6). RSF at
the cartilage-pannus junction express integrin subunits,
including pl, a4, and a5 (7). It is possible that the IL-1
and IL-8 effects are mediated through changes in RSF
integrins and in integrin-cell interior signaling. We
report herein that IL-lp-activated RSF penetration was
inhibited by antibodies to pl, a4, a5, and aV integrins,
but not by antibodies to al, a2, a3, or a6 integrin. These
results suggest that RSF invasion of cartilage requires p l
integrins, and several a subunits, including 134, a5, and
aV, may also participate in the RSF invasion process.
All these integrins are associated with the cell surface
receptor for fibronectin.
RGD-containing peptides also inhibited invasion
of cartilage by RSF. This suggests that both RGDdependent and non-RGD-dependent binding sites are
required for RSF binding to cartilage. The RGD sequence within the major cell binding domain of fibronectin is specifically recognized by the a5pl adhesive receptor (38), while the a 4 p l integrin recognizes the
alternatively spliced IIICS domain (39), which is an
RGD-independent but LDV-dependent cell binding region (40). This alternative spliced region has been shown
to be present in rheumatoid synovial fluid and to be
synthesized by rheumatoid synoviocytes (41,42). The
results obtained suggest that more than one class of
integrin receptors may be required for the RSF invasion
of cartilage. The major integrin subunits we found to be
involved were a4, a5, and aV, all of which bind the
matrix component of fibronectin. Vitronectin could also
be involved, since aV is also a receptor for vitronectin.
In unpublished studies using isolated matrix components
on microtest plates, we found that type I1 collagen
supports RSF adherence. However, in these experiments, antibodies to collagen-binding integrin receptors,
including al, 1x2,and a3, did not inhibit RSF invasion.
The receptors involved in RSF adherence and invasion
may be different, despite the close relationship between
these two phenomena. Future studies are needed to test
the role of otherp integrin subunits in the RSF invasion
of cartilage.
The collagenase inhibitors PNT and phosphoramidon completely inhibited the IL-1p-stimulated RSF
penetration of cartilage, suggesting that collagenase
plays an important role in the invasion of cartilage.
Serine protease inhibitors or cysteinyl protease inhibitors were weak inhibitors of the RSF penetration of
cartilage. We used cartilage slices for the in vitro studies.
In these slices, potential binding sites on cartilage matrix
proteins may be exposed. In intact cartilage in vivo, these
sites are probably sheltered by cartilage surface proteins
and proteoglycans. Degradation of these molecules by
enzymes released from pannus may enhance RSF invasion. IL-1 activates both chondrocytes and synoviocytes
to synthesize and release proteases, including collagenase (43), stromelysin (44,45), and plasminogen activator (46). IL-1 activation results in decreased synthesis
and increased degradation of type I1 and IX collagens
(12) and proteoglycans (13) in the cartilage matrix
remote from the cartilage-pannus junction.
In summary, we have described an in vitro system
for the study of RSF invasion of cartilage, which may be
useful for studying the processes and mechanisms by
which RSF invade cartilage. In this system, IL-1p and
IL-8 activated RSF penetration of cartilage. IL-1activated cartilage invasion was specific to RSF and
required an IL-1 gradient into the cartilage. IL-1pactivated RSF invasion of cartilage was blocked by
antibodies to pl, a4, a5, and aV integrin subunits, by
RGD peptides, and by collagenase inhibitors.
We thank Dr. P. Z . Cohen of the Department of
Orthopedics, Allegheny General Hospital, for providing synovial tissues from the patients with RA and OA. We also thank
Dr. T. Whiteside of the Department of Pathology, University
of Pittsburgh School of Medicine, for providing human dermal
fibroblasts. We are grateful to Dr. D. Farkas of the Science and
Technology Center, Carnegie Mellon University, for his assistance in imaging microscopy. We thank Dr. J. M. Chen, a
postdoctoral fellow in the Medical Research Laboratory,
Western Pennsylvania Hospital, for his assistance in the technology and research of synoviocyte invasion of cartilage.
1. Shiozawa S, Tokushisa T: Contribution of synovial mesenchymal
cells to the pathogenesis of rheumatoid arthritis. Semin Arthritis
Rheum 21:267-273, 1992
2. Shiozawa S, Yoshihara R, Kuroki Y, Fujita T, Shiozawa K, lmura
S: Pathogenic importance of fibronectin in the superficial region of
articular cartilage as a local factor for the induction of pannus
extension on rheumatoid articular cartilage. Ann Rheum Dis
512369473, 1992
3. McCarthy JB, Hagen ST, Furcht L T Human fibronectin contains
distinct adhesion- and motility-promoting domains for metastatic
melanoma cells. J Cell Biol 102:179-188, 1986
4. Wang AZ, Chen JM, Fisher GW, Wang JC, Diamond HS:
Improved in vitro models for assay of rheumatoid synoviocyte
chemotaxis. Clin Exp Rheumatol 12:293-299, 1994
5. Frye C, Yocum D, Benjamin J, Hendrix M: Transformed phenotype of the rheumatoid synoviocyte shows invasion of basement
membrane matrix and enhanced migration (abstract). Arthritis
Rheum 35 (suppl 9):S139, 1992
6. Rinaldi N, Barth T, H e m e C, Mechtersheimer G, Moller P:
Synoviocytes in chronic synovitis in situ and cytokine stimulated
synovial cells in vitro neo-express alpha 1, alpha 3 and alpha 5
chains of beta 1 integrins. Virchows Arch 425:171-180, 1994
7. Ishikawa H, Hirata S, Nishibayashi Y, Imura S, Kubo H, Ohno 0:
The role of adhesion molecules in synovial pannus formation in
rheumatoid arthritis. Clin Orthop 300:297-303, 1994
8. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,
Cooper NS, Healey LA, Kaplan SR, Liang MH, Luthra HS,
Medsger TA Jr, Mitchell DM, Neustadt DH, Pinals RS, Schaller
JG, Sharp JT, Wilder RL, Hunder GG: The American Rheumatism Association 1987 revised criteria for the classification of
rheumatoid arthritis. Arthritis Rheum 31:315-324, 1988
9. Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K,
Christy W, Cooke TD, Greenwald R, Hochberg M, Howell D,
Kaplan D, Koopman W, Longley S 111, Mankin H, McShane DJ,
Medsger T Jr, Meenan R, Mikkelsen W, Moskowitz R, Murphy W,
Rothschild B, Segal M, Sokoloff L, Wolfe F: Development of
criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee. Arthritis Rheum 29:10391049, 1986
10. Comper WD: Cartilage-Molecular Aspects. Boca Raton, FL,
CRC Press, 1991
11. Hauselmann HJ, Flechtenmacher J, Michal L, Thonar EJ-MA,
Shinmei M, Kuettner KE, Aydelotte MB: The superficial layer of
human articular cartilage is more susceptible to interleukin-linduced damage than the deeper layers. Arthritis Rheum 39:478488, 1996
12. Arend WP, J o s h FG, Massoni RJ: Characteristics of chondrocyte
responses to a human interleukin 1-like factor. Clin Immunol
lmmnopathol 36:358-370, 1985
13. Hembry RM, Bagga M, Reynolds JJ, Hamblen D L Immunolocalisation studies on six matrix metalloproteinase and their inhibitors, TIMP-1 and TIMP-2, in synovia from patients with osteoand rheumatoid arthritis. Ann Rheum Dis 54:25-32, 1995
14. Goldring MB, Birkhead J, Sandell LJ, Kimurd T, b a n e SM:
Interleukin 1 suppresses expression of cartilage-specific types I1
and IX collagens and increases types I and 111 collagens in human
chondrocytes. J Clin Invest 82:2026-2037, 1988
15. Smith RJ, Chin JE, Sam LM, Justen JM: Biologic effects of an
interleukin-1 receptor antagonist protein on interleukin-lstimulated cartilage erosion and chondrocyte responsiveness. Arthritis Rheum 34:78-83, 1991
16. Tetlow LC, Lees M, Ogata Y, Nagase H, Woolley DE. Differential
expression of gelatinase B (MMP-9) and stromelysin-1 (MMP-3)
by rheumatoid synovial cells in vitro and in vivo. Rheumatol Int
1353-59, 1993
17. Ishikawa H, Ohno 0, Saura R, Matsubara T, Kuroda T, Hirohata
K: Cytokjne enhancement of monocyteisynovial cell attachment to
the surface of cartilage: a possible trigger of pannus formation in
arthritis. Rheumatol Int 11:31-36, 1991
18. Tuan R, Frye CA, Yocum DE, Seftor EA, Moore TL, Hendrix
MJC: An in vitro model for studying mechanisms underlying
synovial fibroblast mediated cartilage destruction in rheumatoid
arthritis (abstract). Arthritis Rheum 37 (suppl 9):S311, 1994
19. Frye CA, Tuan R, Yocum DE, Hendrix MJC: An in vitro model
for studying mechanisms underlying cartilage destruction associated with rheumatoid arthritis (abstract). Arthritis Rheum 36
(suppl 9):S174, 1993
20. Recklies AD, Golds EE: Induction of synthesis and release of
interleukin-8 from human articular chondrocytes and cartilage
explants. Arthritis Rheum 35:1510-1519, 1992
21. Miyasaka N: Cytokines and chemical mediators in rheumatoid
arthritis. Nippon Rinsho 50:463-469, 1992
22. Endo H, Akahoshi T, Takagishi K, Kashiwazaki S, Matsushima K
Elevation of interleukin-8 (IL-8) levels in joint fluids of patients
with rheumatoid arthritis and the induction by IL-8 of leukocyte
infiltration and synovitis in rabbit joints. Lymphokine Cytokine
Res 10:245-252, 1991
23. Koch AE, Kunkel SL, Burrows JC, Evanoff HL, Haines GK, Pope
RM, Strieter RM: Synovial tissue macrophage as a source of the
chemotactic cytokine IL-8. J Immunol 147:2187-2195, 1991
24. Ishikawa H, Hirata S, Nishibayashi Y, Kubo H, Osamu 0, Imura
S: Cytokine enhancement of adhesion molecule expression on
cultured synovial cells at the surface of cartilage: a possible trigger
of pannus formation in arthritis (abstract). Arthritis Rheum 36
(suppl 9):S265, 1993
25. Taylor DJ: Interleukin-4 (IL-4) induces down-modulation and
shedding of the p55 tnmour necrosis factor receptor and inhibits
TNF alpha’s effect on rheumatoid synovial fibroblasts. Rheumatol
Int 14:21-25, 1994
26. Henderson B, Pettipher ER: Arthritogenic actions of recombinant
IL-1 and tumor necrosis factor a in the rabbit: evidence for
synergistic interactions between cytokines in vivo. Clin Exp Immuno1 7.5306-310, 1989
27. Chu CQ, Field M, Feldmann M, Maini RN: Localization of tumor
necrosis factor a in synovial tissues and at the cartilage-pannus
junction in patients with rheumatoid arthritis. Arthritis Rheum
28. Mohamed-Ali H: Influence of synovial cells on cartilage in vitro:
induction of breakdown and inhibition of synthesis. Virchows Arch
B Cell Pathol 62:227-236, 1992
29. Farahat MN, Yanni G, Poston R, Panayi GS: Cytokine expression
in synovial membranes of patients with rheumatoid arthritis and
osteoarthritis. Ann Rheum Dis 52:870-875, 1993
30. Kahle P, Saal JG, Schaudt K, Zacher J, Fritz P, Pawelec G:
Determination of cytokines in synovial fluids: correlation with
diagnosis and histomorphological characteristics of synovial tissue.
Ann Rheum Dis 51:731-734, 1992
31. Arend WP, Welgus HG, Thompson RC, Eisenberg SP: Biological
properties of recombinant human monocyte-derived interleukin 1
receptor antagonist. J Clin Invest 85:1694-1697, 1990
32. Firestein GS, Boyle DL, Yu C, Paine MM, Whisenand TD,
Zvaifler NJ, Arend WP: Synovial interleukin-1 receptor antagonist
and interleukin-1 balance in rheumatoid arthritis. Arthritis Rheum
37544452, 1994
33. Deleurdn BW, Chu CQ, Field M, Brennan FM, Katsikis P,
Feldmann M, Maini RN: Localization of interleukin-la, type 1
interleukin-1 receptor and interleukin-1 receptor antagonist in the
synovial membrane and cartilageipannus junction in rheumatoid
arthritis. Br J Rheumatol 31:801-809, 1992
34. Haas TA, Plow EF: Integrin-ligand interactions: a year in review.
Curr Opin Cell Biol 6656-662, 1994
35. Yamada KM, Aota S, Akiyama SK, La Flamme SE: Mechanisms
of fibronectin and integrin function during cell adhesion and
migration. Cold Spring Harb Symp Quant Biol 152:203-212, 1992
36. Nikkari L, Aho H, Yli-Jama T, Larjava H, Jalkanen M, Heino J:
Expression of integrin family of cell adhesion receptors in rheumatoid synovium: alpha 6 integrin subunit in normal and hyperplastic synovial lining cell layer. Am J Pathol 142:1019-1027, 1993
37. Schneider M, Voss B, Rauterberg J, Menke M, Pauly TH, Miehlke
RK, Friemann J, Gerlach U: Basement membrane proteins in
synovial membrane: distribution in rheumatoid arthritis and synthesis by fibroblast-like cells. Clin Rheumatol 13:90-97, 1994
38. Demaziere A, Athanasou NA: Adhesion receptors of intimal and
subintimal cells of the normal synovial membrane. J Pathol
1681209-215, 1992
39. Hocking DC, Smith RK, McKeown-Longo PJ: A novel role for the
integrin-binding 111-10 module in fibronectin matrix assembly. J
Cell Biol 133:431-444, 1996
40. Makarem R, Newham P, Askari JA, Green LJ, Clements J,
Edwards M, Humphries MJ, Mould AP: Competitive binding of
vascular cell adhesion molecule-1 and the HepIIiIIICS domain of
fibronectin to the integrin alpha 4 beta 1. J Biol Chem 269:40054011, 1994
41. Makarem R, Humphries MJ: LDV: a novel cell adhesion motif
recognized by the integrin alpha 4 beta 1 . Biochem Soc Trans
19:380S, 1991
42. Huhtala P, Humphries MJ, McCarthy JB, Tremble PM, Werb Z,
Damsky CH: Cooperative signaling by alpha 5 beta 1 and alpha 4
beta 1 integrins regulates metalloproteinase gene expression in
fibroblasts adhering to fibronectin. J Cell Biol 129267479, 1995
43. Trabandt A, Aicher W, Gay R, Sukhatme VP, Nilson-Hamilton M,
Hamilton RT, Mcghee JR, Fassbender HG, Gay S: Spontaneous
expression of immediately-early response genes c-fos and egr-1 in
collagenase-producing rheumatoid synovial fibroblasts. Matrix 10:
349-361, 1990
44. Hutchinson NI, Lark MW, MacNaul KL, Harper C, Hoerrner LA,
McDonnell J, Donatelli S, Moore V, Bayne E K In vivo expression
of stromelysin in synovium and cartilage of rabbits injected
intraarticularly with interleukin-lp. Arthritis Rheum 35:12271233, 1992
45. Case JP, Lafyatis R, Remmers EF, Kumkumian GK, Wilder RL:
Transinistromelysin expression in rheumatoid synovium: a
transformation-associated metalloproteinase secreted by phenotypically invasive synoviocytes. Am J Pathol 135:1055-1064, 1989
46. Leizer T, Clarris BJ, Ash PE, van Damme J, Saklatvala J,
Hamilton J A Interleukin-1p and interleukin-la stimulate the
plasminogen activator activity and prostaglandin E, levels of
human synovial cells. Arthritis Rheum 30:562-566, 1987
Без категории
Размер файла
1 094 Кб
invasion, antibodies, inhibited, collagenase, inhibitors, specific, integrins, receptors, interleukin, synovial, cartilage, stimulate, articular, rheumatoid, fibroblasts
Пожаловаться на содержимое документа