close

Вход

Забыли?

вход по аккаунту

?

Enhanced chondrocyte destruction by lymphokine-activated killer cells. possible role in rheumatoid arthritis

код для вставкиСкачать
500
ENHANCED CHONDROCYTE DESTRUCTION BY
LYMPHOKINE-ACTIVATED KILLER CELLS
Possible Role in Rheumatoid Arthritis
KAREN M. YAMAGA, HARRIET BOLEN, LUCILLE KIMURA, and EUGENE M. LANCE
Objective. The lysis of chondrocytes, the parenchymal cells of cartilage, by lymphocytes may provide a
potent mechanism by which the immune system participates in sustaining joint damage in rheumatoid arthritis
(RA). We studied the capability of lymphocytes from
healthy individuals and patients with arthritis to lyse
chondrocytes.
Methods. Peripheral blood mononuclear cells
(PBMC) were tested for their ability to lyse chondrocytes in a "Cr-release assay. Enhancement of the chondrolytic activity was determined by preincubating the
cells with T cell growth factor (TCGF) or recombinant
interleukin-2 (rIL-2) before cytotoxic testing.
Results. PBMC from healthy individuals possessed a low ability to lyse chondrocytes, whereas cells
from the synovial fluid of patients with RA displayed
higher chondrolytic activity. In RA, modulating factors
must come into play because not all synovial fluid
sample cells showed high chondrolytic activity and cells
The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as
reflecting the views of the Department of the Army or the Department of Defense.
From the Shriners Hospital, the Departments of Tropical
Medicine and Surgery, John A. Burns School of Medicine, University of Hawaii, and the Department of Pathology, Tripler Army
Hospital, Honolulu, Hawaii.
Supported in part by Shriners Hospitals (project no. 15955).
Karen M. Yamaga, PhD: Department of Tropical Medicine,
John A. Burns School of Medicine, University of Hawaii, and
Shriners Hospital; Harriet Bolen, BS: Shriners Hospital; Lucille
Kimura, PhD: Department of Pathology, Tripler Army Hospital;
Eugene M. Lance, MD, PhD: Shriners Hospital, and Department of
Surgery, John A. Burns School of Medicine.
Address reprint requests to Karen M. Yamaga, PhD,
Shriners Hospitals for Crippled Children, 1310 Punahou Street,
Honolulu, HI 96826.
Submitted for publication May 12, 1992; accepted in revised
form November 17, 1992.
Arthritis and Rheumatism, Vol. 36, No. 4 (April 1993)
from synovial tissue had little or no lytic action on
chondrocytes. Chondrolytic activities of cells from all
sources, including PBMC from healthy subjects and
patients with arthritis and cells isolated from synovial
fluid or from the synovial tissue of RA patients, were
greatly increased by incubating the cells with TCGF or
rIL-2. In contrast, treatment of chondrocytes with
interferon-y, which enhances major histocompatibility
complex gene expression, decreased the susceptibility of
chondrocytes to lysis.
Conclusion. These observations suggest a mechanism for joint damage in which the destruction of
chondrocytes by lymphocytes is controlled by cytokines
released during the inflammatory process in arthritic
diseases.
Cartilage is one of the few immunologically
privileged tissues of the body. Its protection from
immune assault is due to the presence of an impervious matrix which isolates its parenchymal cell, the
chondrocyte (for review, see ref. 1). In rheumatoid
arthritis (RA), the integrity of cartilage is breached
and, in the presence of a chronic inflammatory process, is gradually destroyed (for review, see ref. 2).
One hypothesis, proposed by Gertzbein and Lance (3),
is that this process is immunologically mediated and
directed specifically against the tissue-differentiation
antigens of cartilage. Those authors reported that
coculturing lymphocytes with chondrocytes from inbred strains of rats produced remarkable results: syngeneic lymphocyte-chondrocyte mixtures resulted in a
lymphocyte proliferative response almost as great as
that found in major histocompatibility complex
(MHC)-mismatched mixed lymphocyte cultures.
While matrix degradation in inflammatory con-
LYSIS OF CHONDROCYTES
ditions may be due to the presence of enzymes which
break down collagen and proteoglycans (for review,
see ref. 4), the mechanisms whereby chondrocytes are
destroyed remain unelucidated. Malejczyk ( 5 ) first
reported that peripheral blood mononuclear cells
(PBMC) from healthy human subjects can lyse normal
epiphyseal chondrocytes. The lytic cells shared phenotypic and functional properties of natural killer
(NK) cells and were unrestricted by MHC differences.
Short-term incubation of PBMC with interferon-a
(IFNa), but not with interleukin-2 (IL-2), augmented
lytic activity.
We have confirmed and extended the findings
reported by Malejczyk (5). We found that lymphocytes
from the peripheral blood of healthy individuals and
patients with arthritis had low chondrolytic activity.
Cells from the synovial fluids of some patients with RA
showed greater chondrolytic activity than did cells
from any other source. This activity could easily be
augmented by incubating lymphocytes with cytokines.
Synovial tissue lymphocytes were generally refractory, but became potent killers when incubated with
cytokines.
PATIENTS AND METHODS
Isolation of PBMC. Samples of peripheral blood were
drawn into Vacutainer tubes containing preservative-free
heparin (Becton Dickinson, Rutherford, NJ), from surgical
patients undergoing joint replacement. For this study, blood
samples were obtained from 19 patients diagnosed as having
RA and from 25 patients with osteoarthritis (OA). Blood was
obtained also from 16 healthy volunteers. RA was diagnosed
using the American College of Rheumatology (formerly, the
American Rheumatism Association) criteria (6). PBMC were
isolated using Ficoll-Paque (Pharmacia LKB Biotechnology,
Piscataway, NJ) according to the manufacturer's directions,
washed with Hanks' balanced salt solution (HBSS) without
Mg++ and Ca++, and suspended in complete RPMI 1640
(media supplemented with 10 mM HEPES, 100 units/ml
penicillin, 100 pg/ml streptomycin, 0.3% NaHCO,, and 2
mM glutamine) with 10% fetal bovine serum (FBS). Reagents were purchased from Gibco (Gaithersburg, MD).
Synovial fluid samples. Synovial fluid for therapeutic
or diagnostic evaluation was obtained from 18 additional
patients with RA. Samples were collected in preservativefree Vacutainer tubes containing heparin. For highly viscous
samples, bovine testicular hyaluronidase I-S and DNase I
(Sigma, St. Louis, MO) were added, to final concentrations
of 1 mg/ml and 0.35 mg/ml of synovial fluid, respectively.
The mixture was incubated for 10 minutes at 37°C on a
rocker, then diluted with an equal volume of HBSS. Cells
from untreated or enzymatically treated samples were layered over Ficoll-Paque; the cells at the interface were
501
removed, washed twice, and resuspended in complete RPMI
1640 and 10% FBS.
Synovial tissue samples. Synovial tissues from 15 RA
patients who were undergoing joint replacement surgery
were also studied. The tissues were minced in complete
RPMI medium (-5 ml/gm). For some samples, collagenase
IV at 5 mg/ml, DNase I at 0.4 mg/ml, and hyaluronidase I-S
at 0.5 mg/ml were added to the suspension and the samples
incubated for 2 hours at 37°C. The mixture was then filtered
through a wire mesh, and the mononuclear cells were
isolated by Ficoll-Hypaque centrifugation, washed 3 times,
and placed in tissue culture at 1 x lo6 cells/ml of complete
RPMI with 10% FBS. At other times, cells from the minced
tissue were collected directly.
Human articular chondrocytes. Articular chondrocytes were obtained from 69 patients undergoing amputation
or joint replacement surgery (60 cartilage samples from OA
patients, 5 from RA patients, and 4 from patients with
avascular necrosis [AVN]). The chondrocytes were processed from cartilage as described by Lance (7), using a
mixture of trypsin 11, collagenase IV, hyaluronidase type
I-S, and calcium chloride dihydrate (Sigma) in the absence of
FBS. The suspension was rocked overnight at 37"C, then
mixed with FBS to a final concentration of 10% to stop
digestion. The suspension was filtered through sterile nylon
mesh, and the chondrocytes were washed twice, placed in
tissue culture flasks at 5 x lo5 celldm1 in complete medium
with 10% FBS, and then incubated for at least 24 hours to
allow the cells to recover from the digestion process.
The chondrocytes were never cultured for more than
3 days so that they would maintain their differentiated state,
as determined by monoclonal antibodies to proteoglycan
(Yamaga KM, Kimura L, Glant T, Lance E: unpublished
data). The chondrocytes were harvested by treating the cells
with 0.25% trypsin in HBSS, washed twice with RPMI 1640,
and either used immediately or cryopreserved in liquid
nitrogen. In some experiments, chondrocytes were treated
with IFNy for 5-7 days to induce the expression of MHC
class I1 molecules and enhance the amount of MHC class I
gene products (8).
Cell lines. Target cell lines K562 and Raji were
obtained through the American Type Culture Collection
(Rockville, MD). These cell lines were maintained in complete RPMI medium containing 10% FBS. The RA128 B cell
line was generated from PBMC obtained from an RA patient
and was transformed with Epstein-Barr virus using spent
media from a marmoset cell line infected with the virus.
Cytotoxicity assay. Cytotoxicity assays using 51Crrelease were conducted as described elsewhere (9). Briefly,
5 x lo6 target cells were pelleted, resuspended in 0.2 rnl of
FBS, and 100 pCi of 51Cr(-450 mCi/mg) was added. The
cells were rocked gently for 90 minutes at 37"C, washed 3
times with 5% FBS in HBSS, counted, and suspended to 1 X
lo' cells/ml. In some experiments, 2-fold serial dilutions of
attackers were made, starting with 5 x lo6 cellslml. Attackers (0.1 ml) and labeled targets (1 x lo4 cells/O.l ml) were
added in triplicate to 96-well round-bottom plates, and the
cells were centrifuged at 50g for 5 minutes. Control wells
contained target cells with either 0.1 ml of 5% Triton (for
maximum counts per minute release) or 0.1 ml of RPMI
medium with 10% FBS (for spontaneous release). After
YAMAGA ET AL
502
incubating the plate for either 4 hours or 18 hours, the
supernatants were harvested (100 pl/well) and counted in a
gamma counter. The percent lysis was determined using the
following equation:
% lysis =
experimental cpm - spontaneous cpm
maximum cpm - spontaneous cpm
x 100
In other experiments, complete titration curves were
not done because of limited cell numbers. For comparative
purposes, the results for the highest attacker-to-target cell
ratio tested are presented. The spontaneous release was
520% of the maximum cpm released. The average specific
activities of 50 separate labeling experiments of each of the
target cells used were as follows: 0.05 pCi (5,500 cpm) per
lo4 chondrocytes, 0.02 pCi (1,750 cpm) per lo4 K562 cells,
and 0.008 pCi (840 cpm) per lo4 Raji cells. In some tests,
lysis was <O% (e.g., see Table 2) due to counting errors or to
low-level uptake of 51Crby lymphocytes, which resulted in
counts lower than those for spontaneous release. For calculation of averages and for construction of the illustrations
herein, these results were considered to be 0%. When one of
the triplicate values differed by more than twice the other
values, it was discarded. All averages and standard deviations were calculated from triplicate values or, rarely, from
duplicate values. Values were compared using Student’s
r-test.
Fractionation of PBMC and synovial fluid cells. In
some experiments, monocytes and macrophages were removed from PBMC by plastic adherence (lo), and the
remaining cells were further fractionated into sheep erythrocyte rosetting (E+) and erythrocyte nonrosetting (E-)
populations. To 1 ml of packed sheep red blood cells
(SRBC), 10 ml of 2-aminoethylisothiouranium bromide hydrobromine solution (AET; Sigma) (20 mg/ml in HBSS, pH
8.0) was added, and samples were incubated at 37°C for 15
minutes with intermittent shaking. The cells were washed 5
times and brought to 10% by volume with complete RPMI
medium with 10% FBS. To 3 X lo6 whole or nonadherent
cells (per ml) was added 100 pl of AET-treated SRBC in
complete RPMI with 10% FBS. The cells were centrifuged at
35g for 5 minutes, and the pellet was incubated for 1 hour at
room temperature. The cells were resuspended and separated on Ficoll-Hypaque.
The E- cells at the interface were washed thrice
with complete RPMI in 10% FBS. To isolate the E + cells,
the Ficoll-Hypaque solution was removed and 9 ml of sterile
distilled H,O was added to the washed pellet, followed
immediately by 1 ml of lox HBSS and 40 ml of complete
RPMI medium with 10% FBS. E-rosette-positive and Erosette-negative PBMC were further separated by indirect
panning using monoclonal antibody to CD3 (Coulter, Hialeah, FL) and affinity-purified, human Ig-absorbed goat
anti-mouse Ig (Tago, Burlingame, CA) as described (10).
Activation of cells. Unfractionated PBMC, cells isolated from synovial fluid or synovial tissue, and subpopulations of cells were placed in tissue culture for 1-7 days in the
presence of various amounts of T cell growth factor (TCGF;
Collaborative Research, Bedford, MA), an enriched mixture
of lymphokines produced as conditioned medium from
phytohemagglutinin P-stimulated human blood lympho-
cytes, or recombinant IL-2 (rIL-2; Boehringer Mannheim,
Indianapolis, IN), produced as a recombinant DNA product
in Escherichia coli. The activities of the factors were expressed as Biological Response Modifiers Program (BRMP)
units, which has been established as a standard by the
National Institutes of Health (Bethesda, MD).
Immunofluorescence and flow cytometry Fluorescein
isothiocyanate (F1TC)-labeled or phycoerythrin (PE)labeled monoclonal antibodies against CD3, 4, 8, 56, and y
chain of the T cell receptor and control mouse immunoglobulin (MsIg) were obtained from Coulter Immunology. Twocolor immunofluorescence was performed in 96-well plates
using approximately 2 x lo5 ceIIs/weII in a volume of 0.2 mI
containing the antibody diluted in 5% FBS and 0.01M
sodium azide in Dulbecco’s phosphate buffered saline (PBS;
Gibco), pH 7.2. After a 30-minute incubation at room
temperature, the cells were washed 4 times, suspended in
1% paraformaldehyde in PBS, and analyzed on a Coulter
Epics C flow cytometer equipped with an Innova 90 argon
laser (Coherent, Palo Alto, CA) tuned at 488 nm and set at
200 mW. A total of 5,000 cells were analyzed for green and
red fluorescence. The percentage of fluorescent cells was
calculated with the Coulter Quadrant Statistics Program
using FITC-conjugated and PE-conjugated MsIg-treated
negative control cells for comparison.
.
RESULTS
Natural antichondrocyte cytotoxicity activity in
healthy subjects and patients with arthritis. Malejczyk
reported that healthy individuals possessed cells with
the ability to lyse allogeneic chondrocytes (5). To
confirm this result, PBMC were obtained from 16
different healthy individuals and from 25 OA and 19
R A surgical patients, and the chondrolytic abilities
were tested (Figure 1). E a c h value is the percent
chondrolysis that was obtained using a different
attacker-chondrocyte combination (see below and Table 7 for discussion of variation). Among the PBMC
samples, no significant differences in the mean (+.SD)
percent lysis were found in healthy individuals (6 +
5%), O A patients (5 ? 5%), or R A patients (4 2 6%)
under the conditions used.
Analysis of cells from the synovial tissue or
synovial fluid of R A patients, however, yielded different results (Figure 1). Cells isolated from 15 different
R A synovial tissues displayed little (<lo% lysis) or no
activity for chondrocytes (mean k SD 2 k 2%). T h e
mean level of lytic activity of R A synovial tissue cells
was significantly lower than that of PBMC from
healthy individuals and from OA patients (P = 0.05).
In contrast, mononuclear cells from 6 of 18 synovial
fluid samples had enhanced lytic activity toward chondrocytes (>20% lysis), and the mean + SD for the 18
samples was 14 + 12% lysis (P = 0.05 versus any of
LYSIS OF CHONDROCYTES
503
60
50
0
40
E
i/)
>
J
30
w
20
10
0
NORMAL
PBMC
(n=46)
OA
PBMC
RA
PBMC
RA
(n=33)
(n=21)
(n=18)
SF
RA
ST
(n=15)
Figure 1. Lysis of allogeneic chondrocytes by killer celis from peripheral blood mononuclear cells (PBMC) from healthy subjects (normal), osteoarthritis (OA) patients, or
rheumatoid arthritis (RA) patients, or from RA synovial fluid (SF) or synovial tissue
(ST). Values are % lysis of triplicate cultures. Numbers in parentheses are the number
of attacker and target combinations tested. Bars show the group mean. The attacker:
target ratio for normal and OA PBMC was 50: 1; that for RA patients was 50: 1 , except for
2 PBMC samples, which were tested at 25:l and gave lytic values of 3% and 1%; 8 SF
samples, which were tested at 25:l and gave values of 35%. 27%, 18%, 17%, 5%, 3%,
1%, and 0%; 7 ST samples, which were tested at 25:l and gave values of 7%, 6% (2
samples), 2%, and 0% (3 samples), and 1 ST sample, which was tested at 12:l and gave
a value of 1%.
the PBMC samples). These data suggest that some
synovial fluid samples were enriched for cells with
chondrolytic activity, whereas cells from synovial
tissue had little or no such activity.
Fractionation of cells by sheep erythrocyte rosetting techniques. Since Malejczyk ( 5 ) reported that the
chondrolytic activity could be identified in the Epopulation, which contained predominantly N K cells,
PBMC were separated into rosetting and nonrosetting
subpopulations. The unfractionated, E+ and E- subpopulations from PBMC were analyzed for lytic activity against chondrocytes, K562, and Raji cells. The
K562 cell line is NK sensitive, whereas, the Raji cell
line is NK resistant but sensitive to lymphokineactivated killer (LAK) cells. As shown in Table 1, no
significant differences in PBMC from healthy individuals, OA patients, or RA patients were noted.
Thirteen synovial fluid samples contained sufficient numbers of cells to perform E-rosetting fractionation. Six of these 13 samples yielded cells that were
lytic for chondrocytes at 10% or more (Table 2). Most
of the chondrolytic activity was localized in the Esubpopulations. Lytic activity of the other 7 samples
never exceeded 10% in any of the subpopulations
tested. Immunofluorescence analysis was performed
on synovial fluid cells obtained from patients RA774,
YAMAGA ET AL
504
Table 1. Lysis of chondrocytes, natural killer (NKtsensitive cells
(K562), and NK-resistant cells (Raji) by peripheral blood mononuclear cells (PBMC) from healthy subjects and patients with
osteoarthritis (OA) and with rheumatoid arthritis (RA)*
PBMC source,
cell population
Normal
Unfractionated
E+
EOA patients
Unfractionated
E+
E ~RA patients
Unfractionated
E+
E-
% lysis of targets (no. of samples)
Chondrocytes
K562 cells
Raji cells
3 ? 3 (15)
I 2 2 (15)
2 t 3 (15)
43 t 21 (12)
46 t 28 (12)
54 t 27 (12)
2 2 3 (11)
I f 2 (11)
5
2
3
67 + 19 (18)
54 t 23 (18)
64 t 24 (18)
6 r+_ 14 (17)
6 t I4 (17)
5 f 14 (17)
84 2 25 (9)
77 f 2 (9)
79 4 2 (9)
6 t 9 (9)
5 2 10 (9)
4 2 10 (9)
?
?
?
4 (21)
3 (21)
3 (21)
4 2 7 (9)
2 2 6 (9)
3 2 4 (9)
1 k 1 (11)
* PBMC were separated into E+ and E- fractions by sheep
erythrocyte-resetting techniques. lo lysis (mean & SD) calculated as
shown in Patients and Methods. Target chondrocytes were from OA
patients or patients with avascular necrosis. All cells were tested at
an attacker:target ratio of 50:l.
RA799, and RA809, whose samples contained sufficient cells (Table 3). In the E- subpopulation, in
which most of the chondrolytic activity was found,
samples RA774 and RA809 had low percentages of
C D 3 t T cells and slightly enriched NK cells (relative
to T cells) (9% and 11%, respectively). The remaining
lymphocytes in the E- preparations were negative for
all markers tested. Synovial fluid cells from patient
RA799 contained an unusually high percentage of
CD3+, CD4+ T cells in the E- fraction and elicited
the highest chondrolytic activity of the 3 synovial fluid
samples tested.
Enhancement of chondrocyte lytic ability by incubation of PBMC with lymphokines. To determine the
possible role of cytokines in chondrolytic activity,
various concentrations of TCGF (either 0, 1.3, 6.3,
12.7, or 31.5 BRMP unitdml) or rIL-2 (either 0, 2, 5,
10, 20, or 50 BRMP unitslml) were added to PBMC
from healthy individuals, at 5 x lo5 cellslml in a total
volume of 10 ml. A preliminary experiment established
that concentrations of TCGF >50 BRMP unitdm1
were toxic to the cells, whereas concentrations of
rIL-2 as high as 250 unitdm1 supported cells in vitro.
Activated cells were tested on days 1 and 5 for their
ability to lyse allogeneic chondrocytes (Figure 2). High
chondrolytic activity was obtained with approximately
10-15 unitdm1 of TCGF after just I day of incubation.
The high activity remained for at least 5 days. Lysis of
K562 and Raji cells was increased by TCGF activation
of PBMC.
Activation with rIL-2 showed some differences
compared with TCGF (Figure 2). Maximal lytic ability
did not occur until day 5, when the degree of lysis
surpassed that obtained with TCGF. Only modest
enhancement of lytic ability was exhibited after I day
of incubation, and only with the highest concentration
of rIL-2 tested. Thus, TCGF appeared to activate
PBMC more quickly and with lower doses, but rIL-2
treatment of PBMC ultimately resulted in higher chondrolytic ability. Recombinant IL-2 activation of PBMC
enhanced their ability to lyse K562 and Raji cells.
To analyze the long-term kinetics of activation,
PBMC from a healthy individual were tested for
cytotoxic activities and then incubated with 10 units of
rIL-2. Cytotoxic activity was assessed at days 4, 9,
and 13. At day 13, additional rIL-2 was added to a
duplicate culture and tested 3 days later, giving a total
culture time of 16 days. As shown in Table 4, chondrolysis peaked at days 4 and 9, and declined to almost
prestimulation levels by day 16. High chondrolytic
activity was found 3 days after restimulation with
rIL-2. Lysis of K562 cells occurred at all time points
tested, but the highest activities were noted at day 4
Table 2. Lytic ability of synovial fluid mononuclear cells from RA
uatients*
RA patient no.,
cell population
RA634
Unfractionated
E+
ERA672
Unfractionated
E+
ERA774
Unfractionated
E+
ERA783
Unfrac tionated
E+
ERA799
Unfractionated
E+
ERA809
Unfractionated
E+
E-
*
% lysis of targets
Chondrocytes
K562 cells
Raji cells
21 t 4
I t 2
33 t 3
83 t 2
75 ? 6
83 4 4
27 t 5
35 2 0
23 4 2
h t l
-2 2 2
10 & 6
21 t o
17 ? 1
24 t 1
7+3
23 t 4
3 5 1
24 t 4
64 4 3
35 t 1
66 ? 0
39
18 f 1
14 t 1
20 t 1
22 t 3
6 2 2
44 2 1
23 t 1
13 t 2
28 2 2
32 t 3
53
5 5 5
I t 6
2 3
18 ? 4
11 + 4
5
102 1
5 t h
11 4 I
53 t 2
6024
7 5 1
16 + 0
I
78 2 3
75 + 0
78 t 4
35
2
7'1
26 t 1
?
ND
ND
ND
Cells were tested at an attacker:target ratio of 50:1, except for
RA783 and RA799 which were 25: 1. ND = not done. See Table 1 for
details and other definitions.
LYSIS OF CHONDROCYTES
505
Table 3. Cell phenotype and lytic ability of unfractionated cells and subpopulations of mononuclear
cells from RA synovial fluid*
% fluorescence positive
Attacker cell
population
RA774
Unfractionated
E+
ERA799
Unfractionated
Ef
ERA809
Unfractionated
E+
E-
CD3+,
CD4+
CD3+,
CD8+
CD3+,
Ty+
46
49
6
31
27
2
3
2
1
48
42
30
22
37
40
8
% lysis
Chondrocytes
K562
cells
Raji
cells
6
1
8
23
3
24
64
35
66
39
18
13
1
18
32
5
53
53
11
10
7
9
2
ND
2
60
16
21
29
2
4
4
1
15
35
7
26
78
75
78
ND
ND
ND
35
CD3-,
CD56+
7
4
11
* The cell phenotype was determined by 2-color immunofluorescence using phycoerythrin-conjugated
monoclonal antibody to CD3 and fluorescein-conjugated monoclonal antibody to CD4, CD8, CD56, or
T r , and then gating in a flow cytometer for green and red fluorescence. See Tables 1 and 2 for other
details and definitions.
""
I
-1
2o
o
I
1
0
l
I
10
10
20
30
TCCF UNITS/ML
?
80
80
-
0
30
20
TCGF UNITS/MI
TCGF UNITS/MI
100
RAJI
K562
CHONDROCYTES
-
60
>
Lo
82
40
20
0
0
20
10
rlL-2 UNITS/ML
6C
0
20
40
r l L - 2 UNITS/ML
0
20
40
60
rlL-2 UNITS/ML
Figure 2. Generation of killer cells with enhanced ability to lyse chondrocytes. Aliquots of peripheral blood mononuclear
cells from a healthy subject (1 X 10' cellsiml) were incubated with either 0, 1.3, 6.3, 12.7, or 31.5 unitsiml of T cell growth
factor (TCGF) or with 2, 5 , 10, 20, or 50 unitshl of recombinant interleukin-2 (rIL-2) in a total volume of 10 ml. At days
I and 5 , the cells were counted and tested for their ability lyse chondrocytes, K562 cells, or Raji cells. Day-I samples were
tested at an attacker:target ratio of 50: I , and day-5 samples at 25: 1 .
YAMAGA ET AL
506
Table 4.
Kinetics of PBMC activation by recombinant interleukin-2 (rIL-2)*
% lysis by PBMC cultured with rIL-2
~~
~~~
Targets
Day 0
Day 4
Day 9
Day 13
Day 16
Day 16R
OA745 chondrocytes
K562 cells
Raji cells
1 + 1
4924
7+4
67k8
8624
6226
6427
6824
5529
13 2 2
48 5 3
10 2 5
8 2 2
49 2 3
2 f 4
42
83
I1
f
k
2
16
4
3
* PBMC were isolated from a healthy individual and either tested for cytotoxicity immediately (day 0)
or cultured at 1 x lo6 cells/ml in the presence of 10 units/ml of rIL-2 for varying times, counted, and
then tested. All assays were performed at an attacker:target ratio of 25:l. Complete media (without
rIL-2) was added to the cultures at day 9. For day 16R cultures, an additional 10 units of rIL-2 was
added at day 13, and tested 3 days later. See Tables 1 and 2 for details and other definitions.
after initiation of culture and 3 days after restimulation. Lysis of Raji cells paralleled that of chondrocytes. Fluorescence analysis of both 16-day cultures
revealed an increase in CD4+ cells in the restimulated
(56%) versus the once-stimulated (38%) cultures,
whereas the other CD8+, CD56+, and Ty cells were
not selectively stimulated. Paired immunofluorescence
values for each of the markers of the once-stimulated
cultures and restimulated cultures, respectively, were:
28% and 14%, respectively, for CD3+, CD8+ cells;
7% and 7% for CD3+, Ty+ cells; and 10% and 4%,
respectively, for CD3-, CD56+ cells. Apparently,
restimulation resulted in the selective activation of
CD3+, CD4+ cells.
Activation of PBMC, synovial tissue cells, and
synovial fluid cells by TCGF or rIL-2. PBMC from 7
healthy individuals, 9 OA patients, and 9 RA patients
and synovial fluid cells from 8 RA patients were
fractionated into E+ and E- subpopulations. The E +
and E- fractions were cultured in the presence of 8-10
unitdm1 of TCGF. After 4-7 days, the cells were
counted and tested for their ability to lyse allogeneic
chondrocytes, K562, and Raji cells. The E+ subpopulation of all PBMC and synovial fluid samples yielded
higher chondrolytic activity after culturing with TCGF
(increase in chondrolytic activity 18-94%), with no
major differences among the groups. The E- subpopulation gave more variable chondrolytic activity, but
chondrolytic activity increased by >lo% in most of
the samples (23 of 33) after TCGF stimulation.
There were fewer mononuclear cells available
from synovial tissues, and separation into E+ and Epopulations was not done. Unfractionated synovial
tissue cells from patients RA568, R A W , and RA735
were incubated with TCGF and were activated to lyse
chondrocytes (Table 5). The synovial tissue cells that
had been cultured in the absence of TCGF (from
patients RA571 and RA735) failed to lyse chondro-
cytes. No significant differences in lytic susceptibility
were noted between autologous chondrocytes and
chondrocytes from a different RA patient.
Characterization of lymphokine-activated cells.
To better define the LAK cells capable of lysing
chondrocytes, PBMC from a normal individual were
subjected to additional fractionation procedures. They
were passed through a nylon-wool column to remove
B cells, allowed to adhere to plastic to remove monocytes, subjected to E rosetting, and panned twice
using a monoclonal antibody to CD3. Each of the
populations (E+, CD3+; E + , CD3-; E-, CD3+; and
Table 5. Chondrocyte lytic ability of synovial tissue cells from RA
patients before and after culturing in the presence or absence of T
cell growth factor (TCGF)*
Source,
no. days in
culture
RA568
0
4
RA.571
0
4
RA588
10
RA735
3
3
3
3
Target
chondrocytes
(patient no.)
A:T
ratio
%
lysis
OA562
OA562
50: I
50: 1
0
24
-
OA562
OA562
50: I
50: 1
3
0
+
OA582
50: 1
48
-
OA734
RA73.5
OA734
RA735
25: 1
25: 1
25: 1
25: I
6
TCGF
-
+
f
+
1
30
29
* Synovial tissue cells from patient RA568 were either tested
immediately (day 0) or placed in tissue culture with 10 Biological
Response Modifiers Program (BRMP) unitdm1 of TCGF. Cells from
patient RA571 were layered over Ficoll-Hypaque and collected at
the interface, washed 3 times, and tested immediately or after 4 days
in tissue culture without TCGF. Cells from patient RA588 were
placed immediately in culture with 10 BRMP units/ml of TCGF.
Cells from patient RA735 were placed immediately in culture in the
presence or absence of 10 BRMP units/ml of TCGF. A:T =
attackertarget. See Table 1 for details and other definitions.
LYSIS OF CHONDROCYTES
507
Table 6. Cell phenotype and lytic ability of unfractionated cells and subpopulations of mononuclear
cells from peripheral blood of a healthy subject*
% fluorescence positive
% lysis
Attacker cell
population
CD3+,
CD4+
CD3+,
CD8+
CD3-,
CD56+
CD3-,
CD19+
Chondrocytes
Unfractionated
E + , CD3+
E + , CD3E-, CD3E-, CD3+
35 (45)
50 (54)
18 (19)
2 (2)
QNS
27 (30)
42 (44)
30 (29)
2 (3)
QNS
14(11)
2 (1)
42 (30)
55 (76)
QNS
4
0
0
5
QNS
56
4
44
78
4
K562
cells
Raji
cells
80
39
66
64
62
9
34
26
QNS
13
____
~
* PBMC were isolated by Ficoll-Hypaque centrifugation, passed through a nylon-wool column, and
the nonadherent fraction collected and allowed to adhere to plastic to remove residual monocytes.
Cells were subjected to rosetting, and subpopulations were panned using anti-CD3 monoclonal
antibody, yielding the 4 subpopulations. Each (1 x lo6 cells/ml) was treated with rIL-2 (10 units/ml)
for 4 days. Cell phenotype was determined by 2-color immunofluorescence using phycoerythrinconjugated monoclonal antibody to CD3 and fluorescein-conjugated monoclonal antibody to CD4,
CD8, CD56, Ty,or CD19, a B cell marker (see Table 3). Values are % positive after fractionation
(before rIL-2 culture); values in parentheses are those obtained after 4 days’ culture with rIL-2.
Samples were tested for % lysis at an attacker:target ratio of 25:l (ISD was 58). QNS = quantity not
sufficient. See Tables 1-3 for other details and definitions.
E- , CD3-) was incubated with rIL-2, tested individually for lytic ability, and phenotyped (Table 6). The
chondrolytic activity after incubation with rIL-2 was
found in the E+ , CD3- and E- , CD3 - subpopulations. The E-, CD3- subpopulation contained predominantly NK cells; the E+ , CD3- cells contained
cells with both T and NK phenotypes.
Notably, the E+, CD3+ subpopulation containing high-density T cell surface markers in ~ 9 0 %
of
the cells failed to lyse chondrocytes. To test whether
this failure was due to a requirement of “feeder” cells
for activation of this highly purified population, this
experiment was modified such that the PBMC were
first activated with rIL-2 and 4 days later separated
into E+ and E- subpopulations. Each of these fractions was panned twice using anti-CD3 monoclonal
antibodies. In 2 experiments, the E + , CD3+ cells
showed high chondrolytic activity; in addition, lytic
activity was observed in the E-, CD3+ and E-,
CD3- populations. Thus, if activation with rIL-2 was
performed before fractionation, the chondrolytic activity could be generated in the T cell population.
Variation in lytic activity among cells from different healthy individuals and sources of chondrocyte
targets. To assess the degree of variation in chondrolytic activity of cells from different sources, PBMC
from 3 healthy individuals were tested for their ability
to lyse chondrocytes from 4 different sources (Table
Table 7. Lysis of allogeneic chondrocytes or control cells by peripheral blood mononuclear cells
(PBMC) from healthy individuals before or after culturing in T cell growth factor*
~~
~
~~
~
% lysis after culture
% lysis before culture
Subject
1
Subject
2
Subject
3
Subject
Targets
K562 cells
RA128 B cells
Fischer chondrocytes
OA580 chondrocytes
AVN603 chondrocytes
OA614 chondrocytes
OA621 chondrocytes
57 f 15
-22J
8210
-1 2 1
824
724
324
62 f 7
-427
ND
8*1
0 2 6
20*5
5?1
1222
Ilk2
J423
2023
-5k5
15*4
725
622
6*1
1
6 2 4
50216
3427
2523
1923
Subject
2
Subject
3
7223
17t3
323
6029
40*14
1924
2523
J5?5
1120
121
28?4
25211
17?4
3124
* PBMC were tested either immediately or after culture with 8 unitdm1 of T cell growth factor for 3
days. All assays were performed at an attacker:target ratio of 50: 1. Values are the mean 2 SD. Patient
RA128 B cells is an Epstein-Barr virus-transformed cell line from PBMC; Fischer chondrocytes were
from neonatal Fischer rats; human chondrocytes were obtained from surgical patients with osteoarthritis (OA) or avascular necrosis (AVN). ND = not done.
YAMAGA ET AL
508
Table 8. Lysis of syngeneic and alIogeneic chondrocytes by PBMC before and after culture with T
cell growth factor*
70lysis after culture
% lysis before culture
Attacker cell
population
OA647
Unfractionated
E+
EOA648
Unfractionated
E+
E-
OA647
chondrocytes
OA648
chondrocytes
5 2 2
0'1
8*1
1 2 1
0'1
0-0
OA647
chondrocytes
OA648
chondrocytes
4 2 1
10'1
0 + 1
8+1
78 8
o c 2
10 c 1
81 c 5
-3 '0
o r 0
-2 k 2
-1 r 1
30 2 6
22 5
ND
47 c 3
58 % 10
-2 2 2
'
'
* Unfractionated, E+, or E- cells were tested immediately before or after 4 days of culture in the
presence of 8 u n i t s h l of T cell growth factor. Attacker cells were counted and tested at an
attacker:target ratio of 50:l. See Tables 1-3 for details and other definitions.
7). Lysis of K562 cells, B cells (an EBV-transformed
cell line) generated from the PBMC from patient
RA128, and Fischer rat chondrocytes were tested as
controls. The percentage lysis of chondrocytes by
PBMC varied from 0% to 20% at an attacker:target
ratio of 50: 1, and significant differences were observed
among the various chondrocyte preparations. For example, PBMC from subject 1 did not lyse chondrocytes from patient OA580 but lysed 8% of chondrocytes from patient AVN603, which was significantly
different at P = 0.05, by paired ?-test. Furthermore,
cells from different individuals lysed the same chondrocyte preparation to various degrees. Thus, cells
from subject 1 did not lyse OA580 patient chondrocytes, whereas those from subject 3 lysed 20% of those
cells ( P = 0.05). Lysis of K562 cells was >50% for the
2 samples tested, whereas lysis of RA128 B cells and
rat chondrocyte was insignificant. Thus, utilization of
different chondrocyte sources resulted in a variation in
lysis of 0% to 20%.
To determine the degree of variation in chondrolytic activity by TCGF-activated cells, PBMC from
the 3 healthy subjects were incubated with 8 unitdm1
of TCGF for 3 days and then tested for their ability to
lyse the 4 chondrocyte preparations. Activated PBMC
from the 3 individuals showed significantly enhanced
lytic ability toward all chondrocytes (Table 7), compared with freshly isolated PBMC. Chondrocytes that
were resistant to lysis by unactivated PBMC became
susceptible to lysis by activated PBMC.
Susceptibility of autologous versus allogeneic
chondrocytes to lysis. Most of the target chondrocytes
used in the analyses were from allogeneic donors. To
establish that autologous chondrocytes were susceptible to lysis, PBMC and chondrocytes from 2 patients
with osteoarthritis (OA647 and OA648) were isolated,
fractionated, and cultured with or without TCGF
(Table 8). Although the degree of lysis for the different
preparations varied, both autologous and allogeneic
chondrocytes were susceptible to lysis.
Lower susceptibility to lysis of IFNy-treated
chondrocytes. Since IFN y enhances the expression of
MHC class I and I1 antigens (8,l l), the susceptibility
of chondrocytes treated with IFNy was determined.
Allogeneic chondrocytes incubated for 5 days in the
presence of IFNy showed a reduction in lysis by
unfractionated, E + , and E- attacker cells (Figure 3)
by freshly isolated cells from a healthy individual or
from the synovial fluid of an RA patient. When the
PBMC subpopulations from the healthy individual
were activated with TCGF, each lysed the I F N y
treated chondrocytes to a significantly lower degree
than the untreated chondrocytes ( P = 0.01). The
TCGF-activated synovial fluid cells lysed more of the
untreated than the treated chondrocytes, but only the
unfractionated cells yielded a significant difference
( P = 0.01).
This experiment was repeated twice using
PBMC from a healthy individual and an OA patient,
and using allogeneic OA chondrocytes as targets. Both
experiments yielded similar results. Chondrocytes
treated with IFNy were significantly less susceptible
to lysis by PBMC before and after culturing with
TCGF.
DISCUSSION
Despite the immunologic privilege cartilage enjoys as an autotransplant (12), it appears to be the
target of the immune response in inflammatory dis-
LYSIS OF CHONDROCYTES
40
509
-
100
6. AFTER CULTURE
A. BEFORE CULTURE
0UNTREATED CHOND
30
-
cn
cn
3
80
IFN TREATED CHOND
cn
3
60
T
Ul
20-
x
40
20
0
I
W
E+
NORMU.
E-
w
E+
E-
E+
E-
w
E+
E-
SF562
Figure 3. Lowered susceptibility to lysis of interferon-?treated chondrocytes (chond). Peripheral blood mononuclear cells from
a healthy subject (normal) and synovial fluid (SF) cells from a rheumatoid arthritis patient (SF562) were fractionated into
E-rosetting (E+) and non-E-rosetting (E-) populations. The unfractionated, or whole (W), population and E+ and Epopulations were tested immediately (A) or after 5 days in culture with T cell growth factor at 8 units/ml (B), for ability to lyse
allogeneic untreated (open bars) or interferon (1FN)-treated (hatched bars) chondrocytes. The attacker:target ratio was 50:1 for
those tested before culture, and 50:l for the normal sample and 12:l for the synovial fluid samples tested after culture.
eases such as RA (2). Understanding this paradox has
been a major goal at our laboratory. At the cellular
level, rat chondrocytes elicit a vigorous proliferative
response when removed from their immunologically
protective matrix (3). Mixing spleen cells with allogeneic or syngeneic chondrocytes results in stimulation
similar to that found in a mixed lymphocyte reaction.
In humans, however, an equally potent response has
not been observed. Several possible explanations for
the differences between humans and animals include
the presence of MHC class I1 antigens on normal
chondrocytes of rodents and rabbits, but not humans;
rodent chondrocytes are isolated from neonates,
whereas human chondrocyte sources are usually
adults; and human and rat chondrocytes may produce
substances that either inhibit (in the case of humans)
or promote (in the case of rats) proliferation of lymphocytes (7). In humans, a proliferative response to chondrocytes can be obtained by using chondrocyte membranes and T cells from either blood or synovial tissue
from some RA patients (13). These responses are
monocyte dependent, and the responses of PBMC
from RA and OA patients are higher than those of
PBMC from normal controls.
There is relatively little information regarding
the cytotoxic responses by cells of the immune system, even though it is clear that lysis of chondrocytes
offers a potent effector mechanism for promoting
cartilage damage. The existence of MHC-restricted
cytotoxic T cell responses against chondrocytes has
been difficult to demonstrate (7,14,15). A major problem is that chondrocytes appear to inhibit proliferative
T cell responses, and this prevents the generation of
cytotoxic T cells in vitro (7).
Malejczyk first reported that freshly isolated
human chondrocytes are susceptible to NK cells (5).
Our report confirms and extends those observations to
include 3 additional points. First, cells from the synovial fluid of some RA patients demonstrate higher
chondrolytic activity than cells from synovial tissue or
peripheral blood. Second, activated mononuclear cells
can vigorously attack chondrocytes. Finally, since
chondrocytes induced to express higher levels of
MHC antigens were more resistant to lysis than were
untreated chondrocytes, chondrolytic activity may be
down-regulated by IFN y.
With regard to the first point, a comment should
be made on the differences between the percent lysis
of chondrocytes by PBMC from healthy individuals as
reported by Malejczyk ( 5 ) and as found in our study
(Table 7), which was generally lower (120%) and more
varied (620%). Several factors may contribute to this
variation. One is the source of donor chondrocytes.
Malejczyk used epiphyseal chondrocytes from spontaneously aborted fetuses, whereas most of our chondrocyte preparations were obtained from adult surgi-
510
cal patients with varying disease severity. The amount
of matrix production, the progression of disease and
its effect on the joint, and the treatment regimen may
influence chondrocyte susceptibility. Human chondrocytes from adult surgical patients do not readily divide
under normal, low-density tissue culture conditions,
but rather, dedifferentiate after a month in tissue
culture into cells that assume an appearance of fibroblasts (16), thus limiting the number of chondrocytes
available from a single patient. Another source of
variation is technical. Chondrocytes were released
from cartilage specimens by collagenase, hyaluronidase, and trypsin. These enzymes may have varying
effects on the target cells.
The cytotoxicity assays were performed after a
16-hour incubation period, using the conditions reported by Malejczyk (5). Slightly lower lytic activity
was achieved when the incubation period was reduced
to 4 hours (data not shown). Using the 16-hour incubation period, chondrolytic activities of the PBMC
from neither the healthy individuals nor the patients
with arthritis exceeded 20%. Yet, under the same
conditions, synovial fluid cells from 6 of 18 RA patients were found to have chondrolytic activity >20%
and cells from synovial tissue demonstrated a lower
cytotoxic activity (510%). Statistical analyses revealed that synovial fluid cells displayed higher chondrolytic activity than the other samples tested.
Of the 18 synovial fluid samples, 11 were fractionated further into E+ and E- populations. Six of
the samples exhibited lysis >lo% in at least 1 of the
fractions. Most of the chondrolytic activity was concentrated in the E- population in the 6 samples.
Two-color fluorescent analysis of the E- subpopulation from 3 of the synovial fluid samples failed to
identify a predominant cell type. In 2 of the samples,
85% of the lymphocytes present in the E- population
did not react with any of the monoclonal markers
tested; in the other sample, 30% of the lymphocytes
were CD3+, CD4+ and 18% were CD3-, CD56+.
CD56 has been reported to be up-regulated to varying
degrees depending on the cytokine used to treat the
PBMC (17). Thus, the cell type responsible for the
lytic activity in the E- population of synovial fluid is
unknown, and their identification will require additional testing with other surface markers. Synovial
fluid cells may be enriched for activated cells capable
of enhanced lysis of chondrocytes, and therefore need
only be present in low amounts to display lytic activity. Comparisons of matched synovial fluid cells and
PBMC from the same patient need to be performed.
YAMAGA ET AL
Malejczyk concluded that in healthy subjects,
the PBMC responsible for lysing chondrocytes are NK
cells, based primarily on the results of E- fractionation and the fact that a 3-hour incubation of PBMC
with IFNa slightly augmented, while IL-2 did not
increase, the ability to lyse chondrocytes ( 5 ) . In our
studies, high K562 lysis was found both in the E+ and
the E- subpopulations. It is not clear that clean
separation can be achieved simply by E rosetting.
Indeed, NK cells have been reported to fractionate in
both E+ and E- subpopulations (18,19).
With regard to lymphokine-activated killer
cells, it appears that a heterogeneous mixture of
attackers participate in chondrocyte killing, based on 3
observations. First, restimulation of 13-day-old cultures resulted in high chondrolytic activity and the
preferential stimulation of CD3+, CD4+ cells, indicating that T cells may be partially responsible for the
enhanced chondrolytic activity. The second observation attesting to the heterogeneous nature of the attacker cells was that the E+ population cultured in the
presence of TCGF yielded higher and more consistent
chondrolytic activity than did the activated E- population. Third, with highly purified E+, CD3+ T cells,
monocytes or feeder cells were needed before IL-2
was able to stimulate, whereas the E-, CD3- cells
were easily activated. It is likely that appropriately
activated cells include NK cells, CD3+ and CD3- T
cells (which exhibit “promiscuous” killing) (20), and
cells expressing the $6 T cell receptor (21). Direct
evidence for the presence of these different cell types
in synovial fluid will require better separation of
attacker cells by cell sorting, cell cloning, or specific
deletion methods with monoclonal antibodies or selective reagents (10,22).
Only a few reports have described the ability of
NK or LAK cells to lyse freshly isolated autologous or
allogeneic human cells. NK cells appear to be cytotoxic for microvascular endothelial cells (23). Human
keratinocytes are lysed by LAK cells from normal
PBMC (24). Monocyte-derived macrophages also act
as targets of LAK cells (25,26). Furthermore, it has
been found that alloreactive CD3- NK clones can kill
phytohemagglutinin blast cells from some, but not all,
individuals (27,28). Resistance to lysis is dominant and
appears to segregate with HLA haplotype.
The presence of phenotypic and functional evidence for in vivo activation of lymphocytes in the
synovial fluids of RA patients has been a subject of
controversy. There is general agreement that cells in
some samples are activated (29-37). One reason for
LYSIS OF CHONDROCYTES
the lack of uniformity may be the diverse effects of
cytokines in RA (38). In our study, IL-2 enhanced
chondrolytic ability, but chondrocytes treated with
IFNy increased their resistance to lysis. Treatment of
chondrocytes with IFNy results in the increased expression of class I molecules and the appearance of
HLA-DP and DR gene products (8). It is not clear
whether the increased class I expression or the class I1
molecules played a principal role in the decreased
susceptibility to lysis. An inverse relationship between
MHC class I expression and NK sensitivity has been
reported in a variety of systems, including the attack
of &-microglobulin-deficient cells from “knock-out
mice” by NK cells from normal syngeneic mice (39).
These results support the hypothesis proposed by
Ljungren and Karre (40): the function of NK cells
might be to eliminate cells that fail to express MHC
molecules.
Clarifying the reason for the increased resistance of IFN ytreated chondrocytes may help identify
the target antigens of these cells. This task promises to
be a difficult one because of the complex nature of the
receptor-ligand interactions involved in NK and LAK
activity (41,42). We have preliminary evidence that
chondrolytic activity requires cell contact between
lymphocytes and chondrocytes because activated lymphocytes separated from ”Cr-labeled chondrocytes by
cellagen discs and incubated overnight failed to cause
the lysis of chondrocytes. Lytic activity is not due to
the production of soluble toxic substances (unpublished observations).
The diverse effects of cytokines on chondrolytic activities must be elucidated. Understanding
these control mechanisms is important to explaining
the failure of some synovial fluid cells to demonstrate
high chondrolytic activity, the absence of chondrolytic
activity in synovial tissue unless cultured in vitro with
cytokines, and the decreased susceptibility of IFN y
treated chondrocytes. The dynamic interplay between
activation caused by inflammatory processes and suppression due to regulatory mechanisms or drug treatment must be understood.
Mechanisms of cartilage damage thus far explored have centered on the presence of proteolytic
enzymes that cleave matrix components. In this report, we give evidence of another mechanism that may
protract cartilage damage and inhibit cartilage repair
processes. The inflammatory reaction generates cytokines which activate lymphocytes to kill chondrocytes, preventing them from functioning in normal
cartilage repair. The fact that cells from RA and OA
51 1
patients and healthy subjects have the potential to lyse
chondrocytes may not be the critical issue. The key
factor is the presence of relevant cytokines generated
as a result of the inflammatory process. In diseases in
which there is chronic inflammation, such as RA,
ample opportunity to produce an array of cytokines
may be one of the differences between the pathology
associated with chronic inflammation and that associated with diseases in which chronic inflammation is
not a hallmark, as in OA. Establishing lymphokineactivated killer cells directed to chondrocytes as one
of the mechanisms of pathology in RA and devising
strategies for controlling the activity offer an exciting
challenge for further research.
ACKNOWLEDGMENTS
We thank the many physicians and surgeons at
Queen’s Medical Center, Honolulu, Hawaii, for providing
clinical specimens, especially Drs. Jeffrey Fong and Larry
Levin for obtaining some of the synovial fluid samples. The
helpful discussions with Drs. Tibor Glant and Peter Lipsky
are gratefully acknowledged. We thank Lucienne Abe, Remedios Gose, and Brian Tanabe for technical assistance and
Cathy Iwai for administrative support.
REFERENCES
1. Glant TT, Kuettner KE, Thonar EJ, Mikecz K: Immune
responses to cartilage proteoglycans in inflammatory
animal models and human diseases, Cartilage Degradation: Basic and Clinical Aspects. Edited by J F Woessner, DS Howell. New York, Marcel Dekker, 1992
2. Cush JJ, Lipsky PE: Cellular basis for rheumatoid
inflammation. Clin Orthop 2659-22, 1991
3. Gertzbein SD, Lance EM: The stimulation of lymphocytes by chondrocytes in mixed cultures. Clin Exp
Immunol24:102-109, 1976
4. DiPasquale G: Interleukins and metalloproteinases in
arthritis, Inflammatory Disease and Therapy. Edited by
TF Kresina. New York, Marcel Dekker, 1991
5 . Malejczyk J: Natural anti-chondrocyte cytotoxicity of
normal human peripheral blood mononuclear cells. Clin
Immunol Immunopathol50:42-52, 1989
6. 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
7. Lance EM: Immunological reactivity towards chondrocytes in rat and man: relevance to autoimmune arthritis.
Immunol Lett 21:63-73, 1989
512
8. Lance EM, Kimura L, Manibog CN: The expression of
major histocompatibility antigens on human articular
chondrocytes. Clin Orthop (in press)
9. Simpson E, Chandler P: Analysis of cytotoxic T cell
responses, Handbook of Experimental Immunology.
Edited by DM Weir. Oxford, Blackwell Scientific Publications, 1987
10. Coligan JE, Kruisbeek AM, Margulies DH, Shevach
EM, Strober W: Current Protocols in Immunology. New
York, John Wiley & Sons, 1991
11. Jahn B, Burmester GR, Schmid H , Weseloh G, Rohwer
P, Kalden JR: Changes in cell surface antigen expression
on human articular chondrocytes induced by gammainterferon: induction of Ia antigen. Arthritis Rheum
30:64-74, 1987
12. Moskalewski S: Transplantation of isolated chondrocytes. Clin Orthop 272:16-20, 1991
13. 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
14. Strober S, Holoshitz J: Mechanisms of immune injury in
rheumatoid arthritis: evidence for the involvement of T
cells and heat-shock protein. Immunol Rev 118:233-255,
1990
15. Miltenburn AM, van Laar JM, deKuiper P, Daha MR,
Breedveld FC: Cytolytic activity in T cell clones derived
from human synovial rheumatoid membrane: inhibition
by synovial fluid. Clin Exp Immunol 82:49%503, 1990
16. Adolphe A: Articular chondrocytes in culture: applications in pharmacology. Adv Cell Culture 6:1942, 1987
17. Naume B, Gately M, Espevik T: A comparative study of
IL-12 (cytotoxic lymphocyte maturation factor)-, IL-2-,
and IL-7-induced effects on immunomagnetically purified CD56+ NK cells. J Immunol 148:2429-2436, 1992
18. Froelich C, Bankhurst AD: Natural killing and antibody
dependent cellular cytotoxicity: characterization of effector cells by E-rosetting and monoclonal antibodies.
Cell Immunol 78:33-42, 1983
19. Jantscheff P, Milleck J, Thranhardt H: Different Erosetting properties of human peripheral blood NK- and
K-cell. Archiv Fuer Geschwulstfurshung 56:335-339,
1986
20. Thiele DL, Lipsky PE: The role of cell surface recognition structures in the initiation of MHC-unrestricted
‘promiscuous’ killing by T cells. Immunol Today 10:
375-381, 1989
21. Fisch P, Malkovsky M, Braakman E, Sturm E , Bolhuis
RLH, Prieve A, Sosman JA, Lam VA, Sonde1 PM: y/6T
cell clones and natural killer cell clones mediate distinct
patterns of non-major histocompatibility complexrestricted cytolysis. J Exp Med 171: 1567-1579, 1990
22. Thiele DL, Lipsky PE: Tumor cell lysis by T cells
YAMAGA ET AL
distinct from NK cells and alloantigen-specific cytotoxic
T cells. Clin Immunol Immunopathol 49:405-423, 1988
23. Bender JR, Pardi R, Engleman E: T cell receptornegative natural killer cells display antigen-specific cytotoxicity for microvascular endothelial cells. Proc Natl
Acad Sci U S A 87:6949-6953, 1990
24. Kalish RS: Non-specifically activated human peripheral
blood mononuclear cells are cytotoxic for human keratinocytes in vitro. J Immunol 142:74-80, 1989
25. Djeu JY, Blanchard DK: Lysis of human monocytes by
lymphokine activated killer cells. Cell Immunol 111:5565, 1988
26. Streck RJ, Helinski EH, Ovak GM, Pauly JL: Lysis of
autologous human macrophages by lymphokineactivated killer cells: interaction of effector cell and
target cell conjugates analyzed by scanning electron
microscopy. J Leukoc Biol 48:237-246, 1990
27. Gallagher G, Findlay J, Al-Azzawi FAL: Human lymphokine activated killer cells develop syngeneic killing
ability. Immunology 66:471474, 1989
28. Ciccone E, Pende D, Viale 0, Tambussi G, Ferrini S,
Biassoni R, Longo A, Guardiola J , Moretta A, Moretta
L: Specific recognition of human CD3 -CD 16+ natural
killer cells requires the expression of an autosomic
recessive gene on target cells. J Exp Med 172:47-52,
1990
29. Hemler ME, Glass D, Coblyn JS, Jacobson JG: Very
late activation antigens on rheumatoid synovial fluid T
lymphocytes: association with stages of T cell activation. J Clin Invest 78:696-702, 1986
30. Ofusu-Appiah WA, Warrington RJ, Wilkins JA: Characterization of IL-2 responsive synovial T lymphocytes
in rheumatoid arthritis. 11. Functional properties. Rheumatol Int 7:147-151, 1987
31. Dauphinee MJ, Dang H, Flescher E, Wilson-Burris K,
Galarza D, Hempel K, Tala1 N: Characterization of the
IL2 receptor on rheumatoid arthritis synovial fluid T
cells. J Autoimmun 2:813-824, 1989
32. Hovdenes J, Gaudernack G, Kvien TK, Hovdenes AB,
Egeland T: Mitogen-induced interleukin 2 and gamma
interferon production by CD4+ and CD8+ cells of
patients with inflammatory arthritides: a comparison
between cells from synovial fluid and peripheral blood.
Scand J Immunol 30597-603, 1989
33. Hain N, Alsalameh S, Bertling WM, Kalden JR, Burmester GR: Stimulation of rheumatoid synovial and
blood T cells and lines by synovial fluid and interleukin
2: characterization of clones and recognition of a costimulatory effect. Rheumatol Int 10:203-210, 1990
34. Fox DA, Millard JA, Kan L , Zeldes WS, Davis W ,
Higgs J, Emmrich F , Kinne RW: Activation pathways of
synovial T lymphocytes: expression and function of the
UM4D4/CDW 60 antigen. J Clin Invest 86: 1124-1 135,
1990
35. Konttinen YT, Pettersson T , Kemppinen P, Friman C:
LYSIS OF CHONDROCYTES
36.
31.
38.
39.
Spontaneous and in vitro activation of synovial fluid and
peripheral blood lymphocytes in rheumatoid arthritis.
Clin Rheumatol 9:325-332, 1990
DeMaria AF, Malnati MS, Poggi A, Pende D, Cottafava
F, Moretta L: Clonal analysis of joint fluid T lymphocytes in patients with juvenile rheumatoid arthritis. J
Rheumatol 17:1073-1078, 1990
Aaron S, Paetkau B: Synovial cell secretion of IL-2 in
vitro: a limiting dilution analysis. Clin Exp Rheumatol
9: 113-1 18, 1991
Westacott C, Swan A, Dieppe P, Whicher J, Thompson
D: Cytokines in rheumatoid arthritis. Ann Rheum Dis
50:405-406, 1991
Liao NS, Bix M, Zijlstra M, Jaenisch R,Raulet D: MHC
513
class I deficiency: susceptibility to natural killer (NK)
cells and impaired NK activity. Science 253: 199-202,
1991
40. Ljungren HG, Karre K: In search of the ‘missing self‘:
MHC molecules and NK cell recognition. Immunol
Today 11:237-244, 1990
41. Kaare K, Hansson M, Kiessling R:Multiple interactions
at the natural killer workshop. Immunol Today 12:343345, 1991
42. Karlhofer FM, Yokoyama WM: Stimulation of murine
natural killer (NK) cells by a monoclonal antibody
specific for the NK1.1 antigen: IL-2-activated NK cells
possess additional specific stimulation pathways. J Immunol 146:3662-3673, 1991
Документ
Категория
Без категории
Просмотров
0
Размер файла
1 235 Кб
Теги
lymphokine, destruction, killer, arthritis, role, possible, enhance, rheumatoid, chondrocyte, cells, activated
1/--страниц
Пожаловаться на содержимое документа