close

Вход

Забыли?

вход по аккаунту

?

Interactions of synovial fluid immunoglobulins with chondrocytes.

код для вставкиСкачать
1502
INTERACTIONS OF SYNOVIAL FLUID
IMMUNOGLOBULINS WITH CHONDROCYTES
TOSHITAKA TAKAGI and HUGO E. JASIN
Objective. To study the interaction of synovial
fluid (SF) immunoglobulins with living chondrocytes,
and to evaluate the relative contribution of type I1
collagen (CII) antibodies.
Methods. SF of patients with rheumatoid arthritis (RA), osteoarthritis (OA), and gout were incubated
with isolated bovine articular chondrocytes. Ig binding
was measured by flow cytometry and by quantitation
with '251-labeled anti-IgG and anti-IgM. Complementdependent cytotoxicity was determined by
release.
Immunoglobulin binding and cytotoxicity were compared between chondrocytes obtained from the superficial and from the deep cartilage zones.
Results. Significantly greater IgG and IgM binding was found with RA SF compared with OA or gout
SF. Chondrocytes bound more Ig than did fibroblasts.
The relative contribution of anti-CII antibodies to Ig
binding was studied following absorption of the SF with
bovine CII, and by incubation with bacterial collagenase-treated chondrocytes. There was a small but significant reduction in IgG and IgM binding with SF
samples that were positive for anti-CII. RA SF exhibited
From the Rheumatic Diseases Division, Department of
Internal Medicine, University of Arkansas for Medical Sciences,
and the Veterans Administration Medical Center, Little Rock,
Arkansas.
Supported by NIH grant AR-16209.
Toshitaka Takagi, MD, PhD: Research Instructor, Department of Internal Medicine, University of Arkansas for Medical
Sciences; Hugo E. Jasin, MD: Professor of Internal Medicine, and
Director, Division of Rheumatology and Clinical Immunology,
Department of Internal Medicine, University of Arkansas for Medical Sciences.
Address reprint requests to Hugo E. Jasin, MD, University
of Arkansas for Medical Sciences, Department of Internal Medicine,
Mail Slot 509, 4301 West Markham, Little Rock, AR 72205.
Submitted for publication April 7, 1992; accepted in revised
form August 6, 1992.
Arthritis and Rheumatism, Vol. 35, No. 12 (December 1992)
modest, but significantly greater complement-dependent
cytotoxicity than OA SF. Gel chromatography fractionation indicated that IgM antibodies were responsible for
the cytotoxic activity. Additional studies showed that SF
IgM antibodies bound preferentially to, and killed
chondrocytes obtained from, the superficial layers of
cartilage.
Conclusion. Anti-CII antibodies contained in RA
SF represent one of many antibody specificities reacting
with chondrocyte membrane antigens. Chondrocytereactive SF antibodies may play an important pathogenic role in the processes leading to irreversible cartilage damage in RA. These deleterious effects appear to
be exerted particularly on chondrocytes located near the
articular surface of cartilage.
Previous studies have demonstrated the presence of autoantibodies reacting with plasma membrane components of articular chondrocytes in the
sera of patients with rheumatoid arthritis (RA) (1-3).
Although the specificity of these autoantibodies has
not been clearly defined, some antibody molecules
were shown to react with type I1 collagen (CII) (1,2).
Recent studies in our laboratory indicate that polyclonal and monoclonal anti-CII antibodies bind to
chondrocyte membranes and modulate the secretion
of proteolytic enzymes by these cells (4).
In light of the increased frequency and elevated
levels of anticollagen antibodies in the synovial fluid
(SF) of patients with RA (5-9), we undertook the
present studies to investigate the binding of SF IgG
and IgM to the plasma membrane of living bovine
chondrocytes and the relative contribution of anticollagen antibodies to this binding. Our findings demonstrated that RA SF mediate complement-dependent
SF Ig-CHONDROCYTE INTERACTIONS
1503
cytotoxic effects on a subpopulation of chondrocytes.
The data indicate that SF IgM antibodies mediate the
cytotoxic activity of, and appear to bind preferentially
to, the chondrocyte subpopulation localized close to
the surface of articular cartilage. These studies suggest
that the deleterious effects of SF antibodies on resident chondrocytes located near the articular cartilage
surface may be a major pathogenic mechanism contributing to irreversible cartilage damage in RA.
were detached by light trypsinization and treated as described above for the chondrocytes.
SF immunoglobulin fractionation. RA SF with high
cytotoxic activity were incubated at 37°C for 30 minutes with
100 pglml testicular hyaluronidase, prior to gel filtration in a
70 x 5-cm BioGel A-5m column (Bio-Rad, Richmond, CA)
equilibrated with PBS-O.Ol% sodium azide. The protein
peaks in the effluent were monitored at 280 nm. Fivemilliliter fractions were collected, pooled into 4 separate
fractions, and stored at -20°C until used for cytotoxicity
studies.
Antibody binding studies. Chondrocyte monolayers.
Aliquots containing 2 x lo5 cells in 200 pl RPMI 1640-10%
FCS were dispensed into 96-well flat-bottom microtiter
plates (Costar, Cambridge, MA). After incubation for 24
hours at 37"C, the cell monolayers were washed 3 times with
culture medium. Triplicate wells were incubated with 100 pl
of 1:40 dilutions of SF in HBSS-10% FCS for 1 hour at 4°C.
After 3 washes, the chondrocyte monolayers were incubated
with 50 pl of 1.0 pdml affinity-purified '*'I-goat F(ab'),
anti-human IgG or IgM (1-2 x lo6 counts per minute/*;
Tago, Burlingame, CA) for 1 hour at 4°C. After 3 further
washes, cell-bound radioactivity was measured in a gamma
scintillation spectrometer.
Chondrocyte suspensions. Aliquots containing 2 x
lo5 cells in 200 pl RPMI 164&10% FCS were dispensed in
conical microcentrifuge tubes. Triplicate tubes were sequentially incubated with 50 pl of the SF dilutions and 50 pl
I2'I-anti-Igs as described above. The cell suspensions were
washed with 200 pl HBSS-10% FCS after centrifugation at
250g for 2 minutes.
Cytotoxicity studies. Chondrocyte monolayers. Cell
aliquots containing 2 x lo5 cells were dispensed into 96-well
flat-bottom microtiter plates, and were cultured for 2 days.
After 3 washes with culture medium, the chondrocytes were
incubated for 2 hours at 37°C with N+%rO,, 5.0 pCi/well.
Excess "Cr was removed, and the wells were further
incubated for 30 minutes prior to washing 3 times with
culture medium. The chondrocytes were then incubated with
100 pl of 1:40 dilutions of the SF samples in triplicate, for 1
hour at 37°C. After 3 washes, 100 pl of a 1:lO dilution of
fresh (or heat-inactivated as control) rabbit complement was
added, and the plates were incubated for 2 hours at 37°C.
"Cr release was measured using 50-pl aliquots of the cell
supernatants. Wells incubated with 0.1N NaOH were used
to determine 100% release. Results were expressed as the
percent 51Cr release after subtraction of radioactivity released by control wells containing the heat-inactivated rabbit
complement.
Chondrocyte suspensions. Chondrocytes suspended
in culture medium were incubated for 2 hours at 37°C with 25
pCi/106 cells Na,"CrO,.
After 3 washes, the cells were
incubated for a further 30 minutes. Aliquots containing 2 X
lo' cells in 200 pl RPMI 1640-10% FCS were dispensed into
conical microcentrifuge tubes. Subsequent steps were performed as described above for the plate assay, except that
the tubes were slowly rotated during the incubation steps.
Absorption studies. The antibody binding and cytotoxicity studies were repeated with selected SF samples
diluted 1:40 as described above, after addition of purified
bovine CII to a final concentration of 20 pdrnl.
,
MATERIALS AND METHODS
Synovial fluids. SF samples were obtained from patients with definite or classical RA (10) or with osteoarthritis
(OA) who were undergoing joint aspiration for diagnostic or
therapeutic purposes. All of the RA patients had clinically
active disease. Most were being treated with nonsteroidal
antiinflammatory drugs and gold salts or penicillamine; SF
from patients treated with cytotoxic agents were excluded
from these studies. Total and differential cell counts were
performed prior to the removal of cells and debris, by
centrifugation at 1,200g for 15 minutes at 4°C. The supernatants were aliquoted and stored at -70°C until used.
Chondrocytes. Full-thickness articular cartilage
slices were obtained from the carpometacarpal joints of
1-year-old cows. The cartilage pieces were diced into 2 x
2-mm cubes, and were incubated with 10 ml of testicular
hyaluronidase (Sigma, St. Louis, MO), 1 mdml in Hanks'
basal salt solution (HBSS), for 45 minutes at room temperature. After washing with phosphate buffered saline solution
(PBS) containing 0.001M EDTA, the tissue pieces were
incubated with 0.25% trypsin for 20 minutes at 37°C. The
cartilage pieces were washed with HBSS-10% fetal calf
serum (FCS), and then incubated with 2 changes of clostridial collagenase (specific activity 150-200 unitdmg; Worthington, Freehold, NJ), 2 mglml in HBSS-FCS, for 18 hours at
37°C. The cell suspension was filtered through 2 layers of
gauze, washed 3 times, resuspended in RPMI 1640 tissue
culture medium (Gibco, Grand Island, NY) containing 10%
FCS, and incubated, with slow rotation (5 revolutions per
minute), for 24 hours at 37°C in 95% air-5% CO, prior to use.
Experiments with collagenase-treated chondrocytes were
performed immediately after the initial cell isolation and
washing. Cell viability, determined by trypan blue dye
exclusion, was >90%.
For binding studies, the cells were resuspended in
PBS containing 10% newborn calf serum at a concentration
of lo6 cells/ml. In some experiments, thin slices (60-160 pm)
were excised from the articular cartilage surface prior to the
removal of the remaining deeper tissue. In 1 experiment, the
thickness of 10 superficial-zone and deep-zone slices was
measured with a micrometer (Exakt Medical Instruments,
Oklahoma City, OK) (superficial layer mean & SEM 128
10 pm, deep layer 405 k 19 pm). The chondrocytes were
obtained as described above.
Bovine fibroblasts were grown from bovine joint
capsule explants on plastic Petri dishes. The monolayers
*
TAKAGI AND JASIN
1504
Radioimmunoassays. Radioimmunoassays were performed as previously described (1 1). Briefly, flexible polyvinyl chloride microtiter plates (Dynatech, Alexandria, VA)
were coated with affinity-purified anti-human IgM or IgG, 10
pg/ml. CII (6) was dissolved in 0.1N acetic acid and diluted
in PBS to a concentration of 20 &ml for incubation.
Appropriate dilutions of the SF in PBS-FCS were applied in
duplicate for overnight incubation, followed by washing and
incubation with affinity-purified ‘251-F(ab‘)2-goat anti-human
IgG or IgM.
Flow cytometry. Cell characteristics and antibody
binding were assessed with a FacStar Plus flow cytometer
(Becton Dickinson, Mountain View, CA). Data processing
was carried out with a Hewlett-Packard computer, using the
Lysys program. For the detection of bound antibody, the
chondrocytes were treated as described above, except that
the developing antibodies were labeled with fluorescein
isothiocyanate (Tago).
Statistical analysis. The data were analyzed using
Student’s 2-tailed t-test and linear correlation coefficients.
5.0
4.0
*
I
3.0
2.0
1.0
RESULTS
Immunoglobulin binding to bovine chondrocytes was measured in 38 RA SF, 17 OA SF, and 6
gout SF. Significantly greater IgG IgM binding was
found in the group of RA SF, compared with SF from
the other 2 patient groups (P < 0.001 for both) (Figure
1). There was no correlation between the RA SF IgG
or IgM concentrations and the magnitude of IgG or
IgM binding to chondrocytes. However, in the OA SF
as a group, a significant correlation was observed
between IgG and IgM concentrations and Ig binding (r
= 0.72, P < 0.001 for IgG; r = 0.48, P < 0.05 for IgM).
4.0
1
i
I r
I ,I
3.0
2.0
Figure 1. Binding to bovine chondrocytes, by synovial fluid IgG
and IgM from 38 patients with rheumatoid arthritis (open bars), 17
patients with osteoarthritis (hatched bars), and 6 patients with gout
(closed bars). Values are the mean and SEM. * = P < 0.001
compared with the other 2 groups.
I
0.0
RA
OA
Figure 2. Binding to chondrocytesby rheumatoid arthritis (RA; n =
9) and osteoarthritis (OA; n = 6) synovial fluid IgG after treatment
with the sulfhydryl reagent dithioerythritol.Values are the mean and
SEM. * = P < 0.05.
No correlation was found between the magnitude of Ig
binding and SF total leukocyte counts or polymorphonuclear or mononuclear cell counts.
Synovial fluid Ig bound to chondrocytes to a
greater extent than to fibroblasts, suggesting that a
significant proportion of the antibody response was
specific for chondrocyte antigens (mean ? SEM chondrocytes 4.1 0.6 ng anti-IgG/105 cells versus fibroblasts 2.1 k 0.4, P < 0.05, n = 5). In separate
experiments, heat-aggregated human IgG was shown
not to bind to bovine chondrocytes in detectable
amounts (results not shown).
The possible artifactual increase in IgG binding
caused by the presence of IgM rheumatoid factor
contained in the majority of RA SF was investigated
by repeating the IgG binding experiments following
dissociation of IgM with 0.05M dithioerythritol. The
mean ? SEM decrease in IgG binding for the RA SF,
when compared with S F not treated with the sulfhydry1 reagent, was 42.9% k 5.5%. OA SF showed no
decrease in binding. In a separate experiment, 15 RA
and 4 OA SF were used to measure IgG binding in the
presence of dithioerythritol. As shown in Figure 2, IgG
binding by RA SF was still statistically higher compared with OA SF (mean ? SEM 4.4 & 0.3 ng/105 cells
versus 3.0 ? 0.4 ng/105 cells, P < 0.05).
*
SF Ig-CHONDROCYTE INTERACTIONS
1505
Table 1. Prevalence of anti-type I1 collagen antibodies in synovial
fluids of patients with rheumatoid arthritis (RA), osteoarthritis (OA),
and gout*
native CII. As shown in Figure 3, a small but significant reduction in antibody binding (IgG 9.6 5 3.4%
reduction, IgM 15.2 f 3.1% reduction) was observed
only with the SF that were positive for anti-CII antibodies. Moreover, there was a positive correlation
between IgM anti-native CII titers and reduction in
IgM binding after absorption (r = 0.47, P = 0.065 for
IgG; r = 0.71, P < 0.005 for IgM). However, no
correlation was found between anti-CII titers and the
magnitude of IgG and IgM binding. Similar results
were obtained when the antibody binding experiments
were carried out immediately following collagenase
digestion of the chondrocytes as when chondrocytes
were incubated for 24 hours after digestion. Increases
in antibody binding were observed with RA SF, only
for IgM, after recovery from the collagenase treatment
(Figure 4). These results suggest that anti-CII antibodies in SF represent one of several antibody specificities
with chondrocyte binding capacity.
The next set of experiments was performed to
investigate the complement-dependent cytotoxicity
characteristics of S F chondrocyte-binding antibodies.
Cytotoxicity was measured in 28 RA SF and 14 OA
SF. The results shown in Figure 5 indicate that RA S F
exhibited significantly greater cytotoxic activity than
did OA SF (P < 0.05). Of interest was the observation
that only the RA SF containing anti-native CII antibodies demonstrated a significant reduction in cytotoxicity after absorption with CII (anti-CII-positive
32.4 f 5.9% decrease [n = 71, anti-CII-negative 14.7
5.7% decrease [n = 51; P < 0.05).
The modest cytotoxic activity elicited by the
RA SF suggested that only a subpopulation of chon-
IgG
RA (n = 44)
OA (n = 16)
Gout (n = 6)
IgM
Native
Denatured
Native
Denatured
38.6
0
0
77.2
25.0
0
45.5
12.5
0
66.0
25.0
16.7
* Values are the percent positive.
The next series of experiments was designed to
investigate the relative contribution of anticollagen
antibody specificities to chondrocyte antibody binding. IgG and IgM anti-native CII antibodies were
detected in 38.6% and 45.5% of the RA SF used for
these experiments, respectively. Only 2 of 16 OA, and
0 of 6 gout S F were positive for these antibodies.
Anti-denatured collagen antibodies were found in
more than two-thirds of the RA SF and in a minority of
the OA and gout SF (Table 1).
The antibody binding studies were repeated
before and after absorption of the RA S F with excess
3
0
0.4
v)
sn
*
0.3
3M
c?
*-
3
0.2
0
I0
M
a
a
*-
0
E
Q
0
;
I
3.0
t
*
T
I
0.1
0.0
Anti-CII
Anti-CII
IgG
IgM
Figure 3. Effect of absorption of rheumatoid arthritis synovial fluid
(SF) with type I1 collagen (CII) on Ig binding to chondrocytes.
Chondrocyte monolayers were reacted with 8 anti-CII antibodypositive and 8 anti-CII antibody-negative SF. Values are the mean
and SEM. * = P < 0.05; ** = P < 0.01.
w
RA
OA
IgM
Figure 4. Binding of rheumatoid arthritis (RA; n = 10) and osteoarthritis (OA; n = 10) synovial fluid IgG and IgM to collagenasetreated chondrocytes at time 0 (open bars) and after incubation for
1 day (hatched bars). Values are the mean and SEM. * = P < 0.01.
TAKAGI AND JASIN
1506
drocytes appeared to be susceptible to cytotoxic damage by a fraction of the S F tested. Separation of IgG
and IgM from one RA S F by gel chromatography
showed that only the IgM fraction mediated complement-dependent cytotoxicity (Figure 6). Moreover,
flow cytometric studies indicated that no more than
15% of the chondrocytes were stained by the S F Ig
(results not shown).
In light of recent studies suggesting that the
superficial chondrocyte subpopulation exhibits phenotypic and functional characteristics different from
those of cells in the deeper layers (12-14), comparative
studies on the binding characteristics of S F Ig between
these 2 subpopulations were carried out. When the
cells were analyzed by size and complexity (forward
versus side scatter), the dot-plot of the chondrocytes
obtained from thin superficial-layer cartilage slices
separated into 2 distinct subpopulations: a subpopulation of smaller size and greater complexity comprising
35% of the cells, and a second subpopulation of larger
size and lesser complexity comprising 65% of the
chondrocyte (results not shown). The chondrocytes
from the superficial-zone slices in this particular experiment comprised 33% of the total population, so it
can be calculated that the smaller, superficial-layer cell
population (12) accounted for no more than 11% of the
total. In this experiment, immunofluorescence staining
revealed that 22% of the smaller, superficial-layer
chondrocytes showed bright IgM staining with RA S F
only; OA S F failed to show significant (i.e., 1%) IgM
immunofluorescence. IgG staining was also found in
the smaller, superficial-layer chondrocytes, but no
-
I
k
8
ir
4-*
81
0
0
8
-L.
a*
0.0
RA
OA
Figure 5. Complement-dependent cytotoxicity of rheumatoid arthritis (RA; n = 28) and osteoarthritis (OA; n = 14) synovial fluid.
Horizontal lines represent the means. * = P < 0.05.
i
30
h
Y
.
1
0
'R0
20
Y
0
Y
h
0
Y
8Ll
&
0
10
Fraction 1 Fraction 2
Fraction 3
Fraction 4
v1
I
a
0
In
2X
2E
1
0
P
M
c?
.
1
Y
9
M
E
Figure 6. Cytotoxicity profile (top) and IgG and IgM binding profiles (bottom) of a rheumatoid arthritis synovial fluid sample fractionated by column chromatography. The eluted proteins were
pooled into 4 fractions. These were used to measure chondrocyte
binding of IgG and IgM and complement-dependent cytotoxicity, as
described in Materials and Methods.
significant differences were observed between OA S F
(8%) and RA S F (16%).
The results of a separate experiment are shown
in Figure 7. The chondrocyte population derived from
the superficial tissue slices failed to show a distinct
separation between the smaller and larger subpopulations in this instance, probably because the slices were
somewhat thicker. When the whole cell populations
from superficial and deep slices were analyzed ungated
(Figure 7A), incubation with RA S F revealed that 40%
of the superficial-layer chondrocytes stained for IgM,
whereas the deep-layer chondrocytes failed to show
positive immunofluorescence with either RA or OA
SF. Further analysis of the superficial-layer cells gated
SF Ig-CHONDROCYTE INTERACTIONS
twice as much IgM bound to superficial chondrocytes
as compared with the deep-layer subpopulation (mean
+- SEM superficial chondrocytes 4.15
0.46 ng antiIgM bound/105 cells versus deep chondrocytes 2.3 ?
0.22, P < 0.001). Concomitant cytotoxicity studies
with 4 RA SF comparing superficial and deep chondrocyte killing showed that 3 of 4 RA SF exhibited
greater cytotoxic effects on the superficial subpopulation (Table 2). However, there was no correlation
between IgM binding and cytotoxicity in the individual
SF tested.
1
DEEP
SUPERFICIAL
10.
10'
10'
10'
10'
10.
10'
10'
1507
*
10'
10.
FLUORESCENCE
DISCUSSION
3
SUPERFICIAL- RA
0
50
1W
150
100
10'
10'
10'
10.
100
10'
10'
10'
10.
200 250
FORWARD SCAlTER
FLUORESCENCE
Figure 7. Flow cytometric analysis of superficial-layer and deeplayer chondrocytes stained with rheumatoid arthritis (RA) or osteoarthritis (OA) synovial fluid (SF) and developed with fluoresceinlabeled anti-human IgM. A, Ungated chondrocytes obtained from
superficial-layer and deep-layer cartilage slices. Approximately 40%
of the superficial-layer chondrocytes stained positively with the RA
SF. B, Dot-plot of forward scatter (horizontal axis) versus side
scatter (vertical axis) of the superficial chondrocyte subpopulation
stained with RA S F (left panel). IgM fluorescence staining of the
superficial-layerchondrocytes was analyzed separately for high (top
right panel) and low (bottom right panel) side scatter.
according to complexity (Figure 7B) again revealed
that 56% of the high side scatter, more complex cells
showed positive IgM staining, suggesting that most of
the IgM-stained cells belonged to this subpopulation.
Gating simultaneously for low forward scatter and
high side scatter cells yielded similar results.
Binding studies using '251-anti-IgM antibodies
were performed with 12 RA SF. In these experiments,
chondrocytes obtained from superficial-layer cartilage
slices were compared with cells obtained from the
deeper cartilage layers. The former cell subpopulation
contained no more than 35% of the small-sized chondrocytes when analyzed by flow cytometry. Almost
Humoral and cellular reactivities against chondrocyte membrane antigens have been detected in the
sera and circulating lymphocytes of patients with RA
(1,3,15). Although the specificity of most of the reactive antibodies has not been completely characterized,
CII appears to be one of the antigens involved (1).
Recent studies in our laboratory indicate that polyclonal and monoclonal anti-CII antibodies bind to
chondrocyte membranes and modulate the secretion
of proteolytic enzymes by these cells (4). In light of the
increased frequency and elevated levels of anticollagen antibodies in the SF of patients with RA (5-9), we
report herein our findings with regard to the binding
properties of SF IgG and IgM to the plasma membrane
of living bovine chondrocytes and the relative contribution of anticollagen antibodies to this binding.
RA SF exhibited significantly greater IgG and
IgM binding compared with OA or gout SF. There was
no correlation between the binding by and the concentrations of IgM or IgG in RA SF, whereas binding of
IgM and of IgG was positively correlated with their
respective concentrations in OA SF. The experiments
comparing immunoglobulin binding to bovine fibroTable 2. Differential binding and complement-dependent cytotoxicity of rheumatoid arthritis (RA) synovial fluid immunoglobulins on
superficial-layer and deep-layer chondrocytes
Cytotoxicity*
IgM binding?
Synovial
fluid
sample
Superficiallayer
chondrocytes
Deeplayer
chondrocytes
Superficiallayer
chondrocytes
Deeplayer
chondrocytes
RA 1
RA 2
RA 3
RA 4
14.0
8.8
33.1
3.4
36.5
1.4
0
0
2.2
3.8
3.2
5.1
1.8
2.0
2.6
2.3
* Values are the percent 5'Cr release.
t Values are ng anti-IgM/105cells.
1508
blasts and to chondrocytes confirmed and extended
previous observations indicating that serum antibodies
from RA patients bound preferentially to chondrocytes (1,3). It is unlikely that the observed Ig binding
was due to nonspecific interactions between immune
complexes contained in the SF and chondrocyte membrane receptors (16). We have been unable to demonstrate significant binding of heat-aggregated human
12'I-IgG to fresh bovine chondrocytes. Chondrocytes
showed IgG aggregate binding capacity only after
several days of in vitro culture (17), whereas the
experiments reported here involved chondrocytes incubated in vitro for only 24 hours.
Taken as a whole, these observations indicate
that SF Ig binding to chondrocytes was due to specific
antigen-antibody interactions. The positive correlation with IgM and IgG levels in OA S F suggests that in
this particular group of fluids, the low level of Ig
binding may have been due to nonspecific absorption
to the cell membranes, inasmuch as anti-CII antibodies were generally absent in OA SF.
In view of recent observations in our laboratory
indicating that monoclonal and polyclonal anti-type I1
collagen antibodies bind to chondrocytes and modulate protease secretion (4), it was of interest to study
the relative contribution of this S F antibody specificity
to chondrocyte binding. Absorption of the S F with
excess CII resulted in a modest but significant decrease in binding only in the S F that were positive for
this specificity, supporting the view that several other
antibodies contribute to immunoglobulin binding. Similar results were obtained when the antibody binding
experiments were carried out with collagenase-treated
chondrocytes, and with cells allowed to recover for 24
hours. IgM binding increases were observed with RA
SF only; no increase was seen with OA SF, and no
differences in IgG binding were observed with either
SF group. Thus, these studies suggest that S F anti-CII
antibodies, particularly IgM, constitute only a relatively small fraction of chondrocyte-binding immunoglobulins.
In the light of the results described above, it was
deemed important to establish whether SF immunoglobulin binding to chondrocytes resulted in significant
biologic effects on the target cells. Complementdependent cytotoxicity was an obvious effect to investigate, particularly since IgM antibodies were prominently represented among the chondrocyte-binding
antibodies: This immunoglobulin is most efficient in its
ability to activate complement and to mediate complement-dependent cytotoxicity (18). In accord with the
results obtained with the Ig binding studies, RA SF
TAKAGI AND JASIN
exhibited significantly greater complement-dependent
cytotoxic activity than did OA SF. Anticollagen antibodies in the SF that were positive for this specificity
played a prominent role in this, because absorption
with CII resulted in >30% reduction of their cytotoxic
activity.
The modest cytotoxicity elicited by the RA SF
suggested that it might be only a subpopulation of
chondrocytes that was susceptible to cytotoxic damage. Two observations provided an explanation for the
findings. First, immunoglobulin separation of the RA
SF by molecular sieve chromatography indicated that
the cytotoxic activity resided within the IgM antibody
fraction. Second, flow cytometric analysis showed
that only a small fraction of the chondrocytes stained
positively for IgM with the RA SF.
In light of recent studies suggesting that superficial-layer and deep-layer chondrocytes differ not only
morphologically, but also phenotypically and functionally (12-14), it was important to examine whether
there were differences in S F Ig binding and cytotoxicity between these 2 populations. Flow cytometric
analysis of the chondrocyte subpopulation obtained
from superficial-layer cartilage slices showed that the
smaller, superficial-layer cells (12) comprised 25-35%
of the total subpopulation, depending on the thickness
of the tissue slices. Since the cell yield from the
superficial slices amounts to -3340% of the total, it
may be calculated that the small cell population constitutes no more than 14% of the resident chondrocytes.
When the chondrocytes obtained from the superficial-layer cartilage slices were stained with RA S F
and the small and larger cells analyzed separately by
flow cytometry, it was apparent that only the smaller
(lower forward scatter), more complex (higher side
scatter) chondrocytes were stained with IgM. Thus,
the proportion of cells staining with IgM (-7% of the
total cell population) correlated well with the modest
cytotoxicity observed. These findings were confirmed
by binding studies using '251-labeledanti-IgM. In these
experiments, the superficial chondrocyte population
containing 25-35% of the small cells bound almost
twice as much SF IgM as the deep chondrocytes.
Concomitant cytotoxicity studies carried out with 4
RA S F showed increased killing of the superficial
population in 3 of the 4 fluids. However, there was no
correlation between the level of cytotoxicity and the
magnitude of IgM binding.
Taken together, the results obtained suggest
that S F antibodies against chondrocyte membrane
antigens react preferentially with the chondrocytes
residing near the articular surface. The preferential
SF Ig-CHONDROCYTE INTERACTIONS
1509
binding of SF IgM to superficial-layer chondrocytes is
not readily explainable. It is possible that this cell
subpopulation may exhibit membrane antigens different from those expressed in deep-layer chondrocytes.
Recent observations by Hauselmann et a1 (19) suggesting that the superficial chondrocytes may express the
interleukin- 1 receptor differently from the deep chondrocyte subpopulation would support this possibility.
We have previously shown that an anti-CII
monoclonal antibody was able to modulate secretion
of proteases by cytokine-activated chondrocytes (4).
Similar experiments with RA SF have yielded negative
results (Takagi T, Jasin HE: unpublished observations). It is possible that the antibody effect may have
been masked by the vast excess of nonresponsive
deep-layer chondrocytes in the cell populations used
for these studies. Further experiments are in progress
using highly purified superficial-layer chondrocytes to
clarify this issue.
Morphologic studies on articular cartilage from
patients with RA reveal evidence of widespread damage or death of the superficial articular chondrocytes
(20). The observations reported herein may provide an
explanation for this finding. Moreover, they underscore the important role that chondrocyte-reactive SF
antibodies may play in the pathogenic mechanisms
leading to irreversible cartilage damage in RA.
rheumatoid arthritis and osteoarthritis. Arthritis Rheum
28:241-248, 1985
7. Andriopoulos NA, Mestecky J, Miller EJ, Bennett JC:
Antibodies to human native and denatured collagens in
synovial fluids of patients with rheumatoid arthritis. Clin
Immunol Immunopathol6:209-212, 1976
8. Menzel J, Steffen C, Kolarz G, Eberl R, Frank 0,
Thumb N: Demonstration of antibodies to collagen and
of collagen-anticollagen immune complexes in rheumatoid arthritis synovial fluids. Ann Rheum Dis 35:44&
450, 1976
9. Clague RB, Moore LJ: IgG and IgM antibody to native
type I1 collagen in rheumatoid arthritis serum and synovial
fluid: evidence for the presence of collagen-anticollagen
immune complexes in synovial fluid. Arthritis Rheum
27:1370-1377, 1984
10. Ropes MW, Bennett GA, Cobb S, Jacox R, Jessar RA:
1958 revision of diagnostic criteria for rheumatoid arthritis. Bull Rheum Dis 9: 175-176, 1958
11. Olsen NJ, Stastny P, Jasin HE: High levels of in vitro
IgM rheumatoid factor synthesis correlate with HLADR4 in normal individuals. Arthritis Rheum 30:841-848,
1987
12. Aydelotte MB, Kuettner KE: Differences between subpopulations of cultured bovine articular chondrocytes. I.
Morphology and cartilage matrix production. Connect
Tissue Res 18:205-222, 1988
13. Aydelotte MB, Greenhill RR, Kuettner KE: Differences
between sub-populations of cultured bovine articular
chondrocytes. 11. Proteoglycan metabolism. Connect
Tissue Res 18:223-234, 1988
14. Aydelotte MB, Schmid TM, Greenhill RR, Luchene L,
Schumacher BL, Kuettner KE: Synthesis of collagen by
cultured bovine chondrocytes derived from different
depths of articular cartilage (abstract). Trans Orthop Res
SOC16:36, 1991
15. Alsalameh S, Mollenhauer J, Hain N, Stock K-P, Kalden JR, Burmester GR: Cellular immune response toward human articular chondrocytes: T cell reactivities
against chondrocyte and fibroblast membranes in destructive joint diseases. Arthritis Rheum 33: 1477-1486,
1990
16. Uno K, Cooke TDV, Scudamore RA: Interaction of
cultured chondrocytes with heat-aggregated immunoglobulin (abstract). Trans Orthop Res SOC14581, 1989
17. Takagi T, Jasin HE: Reply (letter). Arthritis Rheum
35:1251-1252, 1992
18. Colomb MG, Arlaud GJ, Villiers CL: Activation of C1.
Trans R SOCLond 306:283-292, 1984
19. Hauselmann HJ, Mok-Loh SS, Schumacher BL, Aydelotte MB, Kuettner KE: Different IL-1 and IRAP responses of subpopulations of human articular chondrocytes (abstract). Trans Orthop Res SOC17:61, 1992
20. Mitchell N, Shepard N: The ultrastructure of articular
cartilage in rheumatoid arthritis: a preliminary report. J
Bone Joint Surg [Am] 52:1405-1423, 1970
ACKNOWLEDGMENT
We thank Dr. Joshua Epstein for help with the flow
cytometry studies.
REFERENCES
1. Mollenhauer J, von der Mark K, Burmester G, Glucker
K, Lutjen-DrecollE, Brune K: Serum antibodies against
chondrocyte cell surface proteins in osteoarthritis and
rheumatoid arthritis. J Rheumatol 15:1811-1817, 1988
2. Greenbury CL, Skingle J: Anti-cartilage antibody. J Clin
Pathol 32:82&831, 1979
3. Mettal D, Brune K, Mollenhauer J: Cytotoxic effects of
rheumatoid arthritis sera on chondrocytes. Biochim
Biophys Acta 1138:85-92, 1992
4. Takagi T, Jasin HE: Interactions between anticollagen
antibodies and chondrocytes. Arthritis Rheum 35:224230, 1992
5 . Steffen C: Collagen as an autoantigen in rheumatoid
arthritis, Studies in Joint Disease. Edited by A Maroudas, EJ Holborow. Tunbridge Wells, Pitman Medical
Publishing, 1981
6. Jasin HE: Autoantibody specificities in immune complexes sequestered in articular cartilage of patients with
Документ
Категория
Без категории
Просмотров
2
Размер файла
794 Кб
Теги
interactions, immunoglobulin, synovial, chondrocyte, fluid
1/--страниц
Пожаловаться на содержимое документа