The pathogenesis of chronic inflammation in experimental antigen-induced arthritis. I. The role of antigen on the local immune response

arthritis
and
rheumatism
Official Journal of The American Rheumatism Association
Section of The Arthritis Foundation
The Pathogenesis of Chronic Inflammation in Experimental
Antigen- Induced Arthritis.
1. The Role of Antigen on the Local Immune Response
T.Derek Cooke and Hugo E. Jasin
Using the model described by Dumonde and Glynn in which a chronic
synovitis is induced by the intraarticular injection of antigen in previously immunized rabbits, marked chronic local synthesis of immunoglobulin
comparable to that of lymph nodes and spleen was found as long as 6
weeks after the arthritis was induced. Between 30 and 40% of this newly
synthesized immunoglobulin was accounted for as specific antibody t o
. the locally injected antigen. Very low levels of synthesis were found in the
spleen, control synovia and regional lymph nodes not draining immunized
sites. The circulating antibody levels at the time of induction showed a
positive correlation ( I = 0.66) with the severity of the subsequent histologic findings in the synovium. The fate of the intraarticular antigen was
studied in arthritic and control animals. There was a selective local retention of antigen in animals previously immunized with the homologous
antigen. The retained intraarticular antigen was eliminated very slowly,
with a half-life of over 20 days. These data indicate that the chronic synovial inflammatory response is associated with a chronic local immune response in which prolonged active synthesis of immunoglobulin and specific antibody directed against the locally retained inducing antigen takes
place.
The antigen-induced arthritis described in
the rabbit by Dumonde and Glynn ( 1 ) has certain features in common with rheumatoid arthritis which suggest that it may serve as a sat-
isfactory model for this disease. The synovitis is
associated with a pronounced local response in
which lymphocyte and plasma-cell infiltration
is prominent (2); i t is chronic, often lasting
From the Department of Internal Medicine, Rheumatic
Diseases Unit, University of Texas Southwestern Medical
School, Dallas, Texas
This work was presented at the Interim Session of the
American Rheumatism Association, January 8, 197 1 .
This work was supported in part by National Institutes
of Health Research Grant AM 09989 and Training Grant
AM 05154.
T. DEREK COOKE, MD: Traveling Fellow, McLaughlin
Foundation, Ontario, Canada; HUGO E. JASIN, MD: Postdoctoral Fellow, The Arthritis Foundation; The University
of Texas Southwestern Medical School, 5323 Harry Hines
Blvd, Dallas, Texas 75235.
Reprint requests should be addressed to: Dr. Hugo E.
Jasin.
Submitted for publication Nov 23, 1971; arrepted Jan
31, 1972.
Arthritis and Rheumatism,Vol. 15, No. 4 (July-August 1972)
327
COOKE AND JASlN
several months after a single injection of antigen
into the joint (1-3). Glynn (2-4) has considered two mechanisms for the chronicity of
the arthritis: a) a maintained local response to a
persisting antigen and b) the induction of an
immune response to an autoantigen (2, 3, 5),
possibly a byproduct of the local inflammatory
reaction.
Active synthesis of immunoglobulin (Ig) has
been demonstrated in vitro ( 6 ) and in vivo (7) in
rheumatoid synovial membrane. It has been
suggested that this represents an immune response to an as yet unidentified local antigen.
The synovial response is believed to depend, in
addition, on the formation of complexes between IgG and rheumatoid factors of both the
IgG and IgM varieties (8-1 1).
In view of the similarities noted between the
antigen-induced arthritis and rheumatoid arthritis, it was of interest to characterize the antibody response in the inflamed rabbit synovial
membrane, and to investigate the fate of the intraarticular inducing antigen as well as its role
in the production of the local inflammatory response. The data reported here indicate that a
local immune response takes place in the involved synovial membrane in which lymphoid
infiltrates maintain intense, prolonged and
largely specific antibody synthesis directed at
the inducing antigen. They also show that the
inducing antigen is selectively retained within
the joint and eliminated at a very slow rate,
suggesting a causal relationship between the
persistence of antigen and the prolonged local
synthesis of antibody in the synovial membrane.
MATERIALS AND METHODS
Animals. New Zealand-White rabbits of either sex
weighing between 2.5 and 3 kg were used. They were fed
regular Purina chow and water ad libitum.
Materials. lC-amino acids (Schwartz-Mann BioRrsearch, Orangeburg, NY) consisted of a mixture of Larginine (125 mCi/mM), Lleucine (103 mCi/mM), L
lysine (180 inCi/mM) and L-valine (143 mCi/mM) at a
concentration of 250 pCi/ml for each of the four amino
acids. Carrier and reductant-free lZsI was obtained from
328
New England Nuclear Corp, Boston, Mass. Egg albumin
(EA), five times recrystallized, was obtained from Pentex
Corp, Kankakee, Ill, and bovine serum albumin (BSA)
from Armour Pharmaceutical Co, Chicago, 111. Dried Mycobacterium tuberculosis, H37 Ra, and incomplete
Freund's adjuvant were both obtained from Difco Labs,
Detroit, Mich. Rabbit anti-EA and anti-BSA antisera were
obtained by methods previously described (12). Goat antiserum to rabbit immunoglobulins were obtained from
Hyland Laboratories, Los Angeles, Calif.
Induction of the arthritis. Rabbits were immunized
with 2 ml of an emulsion of EA or BSA and complete
Freund's adjuvant (CFA) in a ratio of 1:1 v/v containing 5
mg of EA or BSA and 2 mg of dried tubercle bacilli. The
emulsion was injected into the footpads, glutei and multiple
intradermal sites in the neck. Seventeen days later, the animals were challenged with 2.5 mg of the respective antigen
in 0.5 ml of pyrogen-free saline solution (Ambot", Cutter
Labs, Berkeley, Calif) sterilized by micropore filtration, injected into each knee joint under aseptic conditions through
the patellar tendon. Control animals immunized in the
same way were challenged 17 days later with 5 mg of soluble antigen intraperitoneally and 0.5 ml of saline in each
knee.
Preparation of tissues and culture procedure.
Animals were sacrificed by exsanguination and air embolism at intervals of 2, 4, 6, or 8 weeks after intraarticular
challenge. The areas to be dissected were carefully shaved
and cleaned with alcohol. Using an aseptic technic, synovectomy was performed by vertically dividing the skin and
soft tissues directly over the knee joint. The synovium was
exposed by a transverse incision through the quadriceps
retinaculum and patellar tendon. The synovial membrane,
including the posterior compartment, was removed completely. The synovium was placed in freshly prepared,
sterile modified Eagle's Minimum Essential Medium
(MEM) and cut into pieces about 2 x 3 mm. The M E M
lacked L-arginine, L-leucine, L-lysine and L-valine, added
later as the radioactive amino acid mixture, and contained
5% inactivated normal rabbit serum, vitamins, glutamine
and penicillin. Samples of synovium, 300 mg in wet weight,
were cultured in 2 ml of M E M containing 10 pCi of IFlabeled amino acids. Cultures were incubated on shaker
platforms (Micromix, Cordis Laboratories, Miami, Fla) at
37OC for 21 hours in 5 ml Petri dishes (Falcon Plastics,
Oxnard, Calif) in an atmosphere of 95% air and 5%COz.
Spleen, popliteal and axillary lymph nodes, obtained by
aseptic dissection, were sliced and teased gently in MEM.
The debris was allowed to settle and the cell suspensions
collected and washed once in M E M by centrifugation at
600 g. Leukocyte counts were taken and the volumes adjusted to provide 5 x 10' nucleated cells. These were spun
again, resuspended in 2 ml of M E M containing 10 pCi of
lC-labeled amino acids and incubated in identical fashion
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
ANTIGEN-INDUCED ARTHRITIS. I
to the synovial cultures. After the incubation period, the
supernatants were removed a n d the remaining cells
separated by centrifugation at 600 g for 10 minutes. Supernatants were centrifuged again at 77,000 g for 1 hour to
remove cell debris, dialyzed against four changes of phosphate-buffered saline (PBS), p H 7.2, over a 36- to 48-hour
period and stored at -20" C until assayed.
Iodination of EA and BSA. N a Iz51 was used to iodinate EA by a modification of the ICI technic of McFarlane (13). IZ5I-BSA was prepared by the method of A.
Taurog (14), using human thyroid peroxidase kindly supplied by Dr. Taurog. After dialysis of the free iodine, the
radioactive antigens were at least 87% precipitable by their
respective antisera.
In vitro incorporation of "C-lobeled omino o c i d s
into protein ond immunoglobulin. T h e thawed supernatants of synovial, lymph node and spleen cell cultures
were centrifuged at 2000 g for 20 minutes and 50 rliter
samples were removed for total radioactivity and trichloracetic acid (TCA) precipitable radioactivity determinations. Incorporation of radioactive amino acids into the tissue culture supernatant immunoglobulin was assayed in 0.1
ml aliquots by immune precipitation using goat antirabbit
Ig in excess. T h e 5% rabbit serum in the M E M provided
sufficient carrier protein to yield a precipitate of about 0.5
mg of protein. In vitro synthesis of anti-EA was assayed on
a 0.1 ml aliquot of the radioactive supernatant by immune
precipitation, using carrier EA and rabbit anti-EA at
equivalence. T h e amounts of carrier antigen and antibody
used were such that immune precipitates of about 0.5 mg
were obtained. Nonspecific coprecipitable radioactivity was
determined using an unrelated antigen-antibody system.
In this procedure, sufficient BSA and anti-BSA to yield 0.5
mg of protein at equivalence were precipitated with a 0.1
ml aliquot of culture supernatant. T h e immune precipitates
were incubated at 37" C for 2 hours followed by overnight
incubation at 4" C. They were then centrifuged at 2000 g,
washed three times with 1 ml of cold PBS and dissolved in
0.5 ml of 0.5 N NaOH. T h e samples taken for measurement of total radioactivity, T C A precipitable radioactivity
and the radioactivity of the immune precipitates were
transferred to glass vials to which 15 ml of Bray's phosphor
containing a thixotropic gel (Cab-0-SiP, Packard Instruments Co, Downers Grove, Ill) were added. Radioactivity was measured with a Beckman scintillation spectrometer. T h e results were expressed as counts per minute
(counts/min) after quench correction using the external
standard technic. T h e small amounts of radioactivity
nonspecifically coprecipitated by the BSA anti-BSA system,
amounting from 2 to 5% of the total, were subtracted from
the radioactivity measured in the immune precipitates.
All the excised synovium, including the sample used for
tissue culture, was dessicated to constant weight under vacuum, and the dry weights obtained.
Histologic preparation of joint tissues. Specimens
of synovium, cartilage and ligament were fixed in 10% buffered formalin solution. They were stained with hematoxylin and eosin (H and E). T h e histologic findings were
graded blindly by one of us (HEJ) on a 0 to 4 point basis for
the degree of a) lining layer hyperplasia and b) mononuclear cell infiltration. T h e maximum score for the two knees
of each animal examined would therefore he 16 points.
Induction of arthritis in robbits using '251-labeled
ontigen. Groups of rabbits, immunized according to the
scheme described above with either EA or BSA in CFA,
were challenged 2 to 3 weeks later with bilateral IA injections of the respective antigen. Either i*51-labeled EA or
BSA was injected into the right knees at the same time. The
detailed schedule is outlined below. Rabbits were sacrificed
at 2, 4 and 6 weeks after induction of arthritis. They received Lugol's solution in the drinking water 24 hours before injection and during the period in which IZ'I-antigens
were being measured.
Measurements of Iz5I-EA or I2jI-BSA tissue radioactivity. T h e animals were exsanguinated under pentobarbital anesthesia and killed by air embolism. Both knee
joints were completely removed. Two milliliter volumes of
minced tissue were counted in a Gamma-guard 150 Spertromatic well-scintillation counter (Tracerlab, Oakland,
Calif). Counts per minute in the right knee tissue were corrected by subtracting the small number of counts per minute
in the corresponding tissue from the left knee.
External measurements of joint radioactivity and precordium, using a gamma scintillation external probe SPA-3
and Radiation Monitor R M 15 (Eberline Instrument
Corp, Santa Fe, NM), were made at 1- to 3-day intervals
after the injection of the '*jI-antigens.
Measurement of circuloting anti-EA titers. These
were measured by a modification of Farr's technic (15).
Sera obtained just prior to IA challenge and at sacrifice
were stored frozen at - 2 P until tested. Standard solutions
of IZ5I-EA containing 6.25 pg EA (1
N/ml) were allowed to react with serial dilutions of the antisera and the
gammaglobulin then precipitated by the addition of an
equal volume of saturated ammonium sulphate solution.
Serum antibody titers were expressed as the antigen binding
capacity (ABC-33), the amount of antigen nitrogen bound
per milliliter of undiluted serum, using as an end-point 33%
precipitation of the antigen. Least square regression analysis was applied to ABC-33 serum titers and histology
grades.
RESULTS
Histologic appearance of the arthritis
induced by EA. T h e arthritis induced by
intraarticular injection of 2.5 mg EA in saline
solution in previously immunized rabbits was
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
329
C W K E AND JASlN
Fig 1. Histologic appearance of synovium obtained from a rabbit immunized with
EA and injected IA with 2.5 m g EA in saline solution 4 weeks previously. Note the
lining layer hyperplasia and the infiltration with mononuclear cells resembling a
lymphoid follicle. Histology grade: 7 ( x 130).
Fig2. Synovium of immunized rabbit 8 weeks after challenge with 2.5 m g EA intraarticularly. Note the cell aggregates composed predominantly of lymphocytes
and plasma cells. Histology grade: 6 ( x 800).
330
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August1972)
ANTIGEN-INDUCED ARTHRITIS. I
WEEKS AFTER INTRAARTICULAR CHALLENGE
Fig 3. 14C-amino acid incorporated into Ig in 21
hour cultures of synovium. popliteal and axillary
lymph nodes and spleen expressed as percent of
TCA precipitable counts/min: 2,4 and 6 weeks after
IA challenge. 0 = Saline-injected joints; W =
Antigen-injected joints; illl = Spleen cell suspen= Lymph node cell suspensions.
sions and
Arthritis was induced by IA injection of 2.5 m g EA
in rabbits previously immunized with EA in the four
foot pads, glutei and neck. Control animals, also
immune, were challenged with 5 m g EA intraperitoneally and IA saline solution. Figures in parenthesis represent number of animals in each group
where all three tissues were cultured simultaneously.
radioactivity. It can be seen that Ig synthesis by
these inflamed synovial tissues comprised 15 to
30% of the newly synthesized TCA precipitable
proteins released into the supernatant. It is also
apparent that the synovial Ig synthesis compares well with that of the lymph nodes and
spleen at 2, 4 and 6 weeks after challenge. In
fact, at 6 weeks the percentages in these three
tissues were similar in magnitude. It is of interest that the saline-injected control synovium
synthesized some Ig, as shown by the small but
reproducible radioactivity in the immunoprecipitates. Synovial cultures from two normal
rabbits, not shown in the figure, did not synthesize any detectable Ig.
When these data were expressed as counts
per minute incorporated into Ig by all the syn-
r
"t
3.6
2.8
'f
2.4-
2
X
~
present in animals killed up to 8 weeks after the
injection (Figures 1 and 2). It was characterized
histologically by hyperplasia of the lining layer
and infiltration of the deeper layers with lymphocytes and plasma cells. T h e cell infiltration
was diffuse, but focal accumulations resembling
lymphoid follicles were often seen (Figure 1).
The joint fluids contained variable numbers of
cells, with polymorphonuclear neutrophiles as
the main cell type. T h e synovitis produced by
IA injection of BSA in BSA-immunized animals was indistinguishable from the synovitis
induced by EA in EA-immunized animals.
In vitro Ig synthesis by synovium and
lymphoid cells. Figure 3 shows the incorporation of '4C-amino acids into Ig by inflamed
and control synovial tissues, popliteal and axillary lymph node and spleen from the same animals, expressed as percent of TCA precipitable
2.0-
W
.-Ec
0
1.6-
\
c
c
s
V
...
A
0
1.2-
0.8-
.
0
..
.
-r
tl
0.4
1
WEEKS AFTER INTRAARTICULAR CHALLENGE
Fig 4. 14C-amino acid incorporation into Ig i n 21
hour cultures of synovium from immune rabbits
obtained 2, 4 and 6 weeks after IA challenge with
2.5 m g EA or 0.5 m l saline solution.
= Anttgeninjected and 0 = Saline-injected. The latter animals were challenged with 5 mg EA intraperitoneally. Results are expressed as c o u n t s h i n incorporated into Ig by all the synovium excised from each
joint.
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
331
COOKE AND JASIN
pressed as percent of radioactivity incorporated
into Ig is shown in Figure 5. At 2 , 4 and 6
weeks, between 30 to 40% of the Ig synthesized
in the synovium was anti-Ea. In comparison,
the control synovia and the spleen produced less
than 5%. T h e popliteal and axillary lymph
nodes, all of which were draining immunized
paws, synthesized between 15 and 20% of Ig as
specific antibody.
50r
(71
(6)
Influence of site of primary immunizat i o n o n p r o d u c t i o n of specific a n t i body. To examine this question, two rabbits
were immunized in the paws, gluteus muscle
and intradermally on one side only. T w o weeks
later they were challenged in both knees with
antigen. Tissues were obtained 4 weeks after
inducing the IA injection. A marked difference
in the production of specific antibody was found
in the lymph node cultures of each side. T h e
specific antibody synthesis in these tissues, expressed as percent of newly synthesized Ig, is
seen in Figure 6. Both popliteal and the axillary
lymph nodes draining the immunized paws
produced anti-EA at similar levels. O n the
unimmunized side, the lymph nodes were
smaller and the cultured popliteal nodes did not
synthesize detectable anti-EA. Specific antibody
synthesis by spleen was also undetectable.
However, the synovium from the knees of both
immunized and unimmunized sides demonstrated marked synthesis of anti-EA at similar
2
WEEKS AFTER INTRAARTICULAR CHALLENGE
Fig 5. 14C-amino acid incorporation into anti-EA
in 21 hour cultures of synovium. lymph node and
spleen cells expressed as percent of counts/min
incorporated into Ig.
= Antigen-injected joints;
0 = Saline-injected joints; Nl = Spleen cell suspension and El = Lymph node cell suspension.
Immune rabbits were injected IA with 2.5 rng EA
or saline solution 2. 4 and 6 weeks previously.
Saline-injected animals were challenged with an
equivalent dose of antigen intraperitoneally.
ovium excised from the joints (Figure 4), fourfold to eightfold greater mean incorporation
was observed in the antigen-injected joints
compared with the saline-injected joints. It is
again apparent that inflamed synovial tissues
continued to synthesize Ig very actively 6 weeks
after induction of the arthritis.
Specific antibody synthesis by synovium
and lymphoid cells. T h e radioactive amino
acid incorporation into anti-EA antibody, ex50 r
Immunized Side
I
;
Nonimmunized I
Side
I
!
a
W
30
kg
a
.-c
20
E \
\ -
:
; lo
s*
0
1
Synovium
1
I
Pop1I t eal
332
t
Axlllary Synovium
1
t
Fig 6. In vitro 14C-amino acid incorporation into anti-EA by synoviurn,
spleen and lymph node cells expressed
as percent of counts/min incorporated
into Ig. W = Antigen-injected joints;
F
3 = Lymph node cell suspensions
and
= Spleen cell suspensions.
Two rabbits were immunized unilaterally with EA in CFA and subsequently
challenged in both knees with 2.5 rng
EA in saline solution. The tissues were
obtained 4 weeks after the I A injection.
Spleen
Pop1 Iteal
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
ANTIGEN-INDUCED ARTHRITIS. I
Table 1. Sum of Histologic Grades of Synovitis in
Right and Left Knee after the lntraarticular
Injection of EA*
Duration of
arthritis
(wk)
2
4
6
No. of
animals
8
10
4
Average
grade
7.8
7.3
4.5
Range
6-11
2-10
3-6
"2.5 m g injected into each joint
levels. These findings indicate that, although
injection of antigen with CFA is a prerequisite
for the prolonged synthesis of anti-EA by the
local lymph nodes, the occurrence of synovitis
and its persistence is dependent on systemic
immunity, as well as the intraarticular deposition of inducing antigen. It appears to be independent of the immune response in the adjacent
lymph nodes.
Correlation of serum antibody level
with severity of arthritis. In an attempt to
correlate the initial serum antibody response
with the severity of the arthritis, sera obtained
just before intraarticular challenge with antigen
were assayed for anti-EA titers by the ammonium sulphate precipitation technic, using radioactive EA. These were plotted against the sum
of the histology grades assigned to the right and
left knees of each rabbit. Since the histologic
examination showed a decrease in average
severity of arthritis 6 weeks after induction
(Table 1 ) and the anti-EA titers were all determined immediately prior to intraarticular
challenge, a linear relationship between these
parameters would not be expected. Therefore,
only the 18 rabbits with arthritis of 2 and 4
weeks duration were included. When the values
at these time intervals were correlated with
serum antibody levels at the time of indyction of
arthritis, a positive correlation ( r = 0.66) and a
statistically significant straight-line relationship
( P < 0.005) were obtained (Figure 7).
The fate of '251-labeledantigen. T w o
groups of rabbits immunized with either BSA
(experimental) or EA (control) were challenged
in both knees 3 weeks later by the IA injection
of the respective inducing antigen. T h e experimental group received 2.5 mg I2jI-BSA
containing 15 x lo6 counts/min in the right
knee and the control group received 2.5 mg EA
plus 15 x 106 counts/min of '251-BSA. In addition, 2.5 mg 1251-BSA(28 x lo6 counts/min)
was injected into both knees of two normal,
nonimmune rabbits. Using the external probe,
the percent of initially injected radioactivity retained in the right knees was measured and
plotted against time. In Figure 8 are seen the
elimination patterns of Iz5I-BSA in these groups
of rabbits. T h e early patterns are similar in the
normal, nonimmune and control group with an
EA induced synovitis. There is an initial rapid
decrease in radioactivity, lasting for about 5
days, which probably represents equilibration
with the rest of the body. T h e experimental
group, preimmunized and injected locally with
BSA, showed a more rapid initial decrease,
reaching approximately 0.25% of the initial
radioactivity by 5 days. At this time, the normal
I0
??
2l
L
I
100
I
200
I
300
I
400
I
500
ANTIGEN BINDING CAPACITY (pgN1
Fig7. Positive correlation (r = 0.66; P < 0.005)
of antigen-binding capacity of sera (ABC-33) of
individual rabbits prior to IA challenge with antigen
and subsequent severity of arthritis at 2 and 4
weeks assessed by the added histology grades from
both knees. Lining layer hyperplasia and mononuclear cell infiltration each graded 0 to4.
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
333
COOKE AND JASlN
100.0 r
Fig 8. Elimination curves of IA
lZ51-BSA in immune arthritis and
control rabbits obtained with an
external probe. Experimental rabbits (BSA-immune) -A-: arthritis
was induced with 2.5 m g BSA bilaterally. The r i ht knees were injected w i t h 1'51-BSA (15 x lo6
counts/min). Control arthritic rabarthritis
bits (EA-immune) ---Or--:
was induced with 2.5m g EA bilaterally. I n addition. the right knees
were i n j e c t e d w i t h (15 x lo6
counts/min) lZ51-BSA. Normal rabbits (nonimmune) - 0 - injected
only with 2.5 ~TI~"~I-BSA (28 x
1o6counts/min).
I
0.001;
2
3
4-
WEEKS
and control rabbits knees retained four times
this amount. T h e difference presumably represents immune elimination in the rabbits preimmunized to BSA. After the initial 5-day period,
the unimmunized and control animals with
EA-induced arthritis then eliminated I2jI-BSA
at a steady and comparable rate. Between 10 to
14 days however, a n abrupt change in slope was
seen which coincides with the development of
the expected primary immune response to BSA.
This response decreased the remaining articu-
lar antigen to lower levels than those remaining
in the experimental group, and was followed by
a very slow, constant rate of elimination. A
similar slow component in the experimental
group began immediately after the initial rapid
phase. Sufficient radioactivity remained in the
experimental group to allow computation of the
half-life of the remaining intraarticular antigen. T h i s was estimated to be 20 days or more.
A reciprocal, but otherwise identical, experiment was carried out in EA preimmunized
Table 2. Retention of lz5I-BSA in Joints of Normal or Preimmunized Rabbits
lz5I-BSA
in iected
Immunizing
antigen
BSA
EA
BSA
EA
-
Weeks after
lntraarticular intraarticular Counts/min
antigen
injection
x 106
BSA
EA
BSA
EA
BSA
2
2
4
4
4
15
15
15
15
28
lZ51-BSAretained
Cou nts/min
32.930"
12.120"
24.600
4,310
9.200-t
Countshnin (BSA)
010
0.220
0.094
0.165
0.029
0.033
Countshnin (EA)
2.3
-
5.7
-
*Average counts/min for the right knees of two rabbits
?Average counts/min for both knees of two rabbits
334
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
ANTIGEN-INDUCED ARTHRITIS. I
Table 3. Retention of 1251-EA in Arthritic Joints of Immunized Rabbits Induced Previously
by EA or BSA. Both Groups Received lz5I-EA* in the Right Knees
I mrnunizing
antigen
Inducing
antigen
EA
BSA
EA
BSA
EA
BSA
EA
BSA
Duration of
arthritis
(wks)
2
2
4
4
1251-EAretained
Counts/min (EA)
Cou nts/mi n t
Counts/min (BSA)
24,350
4,970
21,250
2.900
4.9
7.1
”30 x 106counts/min injected
tAverage counts/min of the right knees of two rabbits
animals by injecting 30 x lo6 counts/min of
I2jI-EA into knees of rabbits with EA (experimental) and BSA (control) arthritis. T h e
elimination patterns of 1251-EA by external
monitoring were similar to the first experiment.
Selective retention of radioactive antigen. Selective retention of radioactive antigen
in rabbit knees with a synovitis induced by the
same antigen was confirmed when the rabbits
were sacrificed, the knee joints removed and
their radioactivity determined. After injection
of radioactive BSA (Table 2), the joints of rabbits immunized to BSA retained 2.3 and 5.7
times as much antigen as the control rabbits
with arthritis induced by EA and injected with
12’I-BSA. T h e normal, nonimmune rabbits
sacrificed 4 weeks after the IA injection retained
0.03% of the initial dose, a level similar to that
measured in control arthritic joints. T h e results
using l2jI-EA are seen in Table 3. Again,
marked retention of radioactivity is evident in
the joints of rabbits immune to EA; 4.9 and 7.1
times as much 1251-EAwere retained at 2 and 4
weeks than in the BSA inducedjoints, which
also had been injected with 12’I-EA.
DISCUSSION
In this investigation, a number of aspects of
an immunologically-induced, chronic inflammatory reaction have been studied in the antigen-induced a r t h r i t i s of D u m o n d e a n d
Glynn (1,4). T h e arthritis was induced by in-
jection of EA or BSA into the knees of rabbits
previously immunized with either antigen in
CFA. T h e experiments were designed to quantitate immunoglobulin and specific antibody
synthesis concurrently in synovium, lymph
nodes and spleen. These were measured by immune precipitation of 14C-amino acid-labeled
immunoglobulin and anti-EA synthesized in in
vitro cultures of these tissues. T h e method
utilized has been previously described from this
laboratory (6, 12).
T h e synovitis induced was associated with an
intense and prolonged synthesis of immunoglobulin for as long as 8 weeks after induction,
the longest interval tested. This finding was
consistent with the prominent infiltration of the
synovial m e m b r a n e w i t h lymphocytes a n d
plasma cells. In the EA-injected joints, anti-EA
comprised 30 to 40% of the total Ig synthesized
throughout the 6-week observation period.
Anti-EA synthesis in these joints made up a
much larger fraction of the total Ig synthesized
than observed in control joints and spleen. Prolonged synthesis of large amounts of specific
antibody, directed at the locally injected antigen, provided indirect evidence for the persistence of the antigen.
Control cultures of synovium from immunized animals which were not challenged intraarticularly, synthesized small but reproducible amounts of Ig and, on occasion, small
amounts of specific antibody. T h i s is consistent
with the observation of Jasin and Ziff (12) who
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
335
COOKE AND JASlN
demonstrated synthesis of antibody to intravenously-injected BSA in t h e synovial
membrane of an animal injected intraarticularly with the nonspecific mitogens streptolysin-S
and phytohemagglutinin. Moreover, the synovium from saline-injected joints also demonstrated mild inflammatory changes as long as 6
weeks after IA injection. It is possible that the
mild acute inflammatory response produced by
the saline solution may have attracted antigen
and immunocompetent cells to the joint, leading
to the accumulation of cells synthesizing specific
antibody to the intraperitoneally injected EA.
In unilaterally primed rabbits, anti-EA synthesis in lymphoid tissue occurred only in
lymph nodes from the immunized side, confirming previous work by Askonas and
White (16). In contrast, synovial synthesis of
anti-EA was similar on both immunized and
unimmunized sides, indicating that the accumulation of antibody producing cells in the
synovium was not influenced either by the site
of immunization or the immune status of the
adjacent lymph nodes and depended only on the
systemic immunity of the animals.
T h e level of the circulating antibody present
at the time of intraarticular challenge appeared
to influence the severity of the subsequent arthritis, since a positive correlation was found
between pre-intraarticular challenge serum
antibody levels and the degree of subsequent
inflammation. It would appear from this that
the severity of the synovitis was related to the
magnitude of the humoral antibody response.
Glynn (2) has emphasized a requirement for
the presence of delayed hypersensitivity to the
immunizing antigen for the development of the
arthritis on the basis of a) the necessity for immunization with CFA and b) the correlation of
the arthritis with a positive delayed skin reaction to the antigen. T h e present findings provide evidence that the humoral antibody response also plays a direct role.
T h e finding of prolonged specific antibody
synthesis in the inflamed synovial cultures
supports the concept that a local immune re336
sponse is being maintained by persisting antigen. Webb, Ford and Glynn (17) have shown
that systemic immunization of rabbits up to 4
weeks after intraarticular injection of antigen
results in the appearance of a chronic synovitis
in the antigen-injected joints. Studies in a subsequent paper (18), describing the retention of
antigens in this model of arthritis and the site of
their localization, indicate that persisting antigen may indeed be directly responsible for the
chronicity of the arthritis.
T h e fate of '251-labeled antigen in normal
and arthritic rabbits was markedly influenced
by the immune status of the animal at the time
of joint challenge, since animals immune to the
labeled antigen retained it in greater amounts
than control rabbits with arthritis induced by
another antigen or normal rabbits. Similar results were recently reported by Consden et
a1 (4); intraarticular retention of EA was increased in preimmune rabbits. This suggests
that intraarticular antigen interacts with antibody to form complexes which are retained in
the joint. A similar mechanism has been found
to operate in germinal centers of lymph nodes
of immunized animals where antigen and antibody, presumably in the form of complexes, are
r e t a i n e d f o r c o n s i d e r a b l e p e r i o d s of
time (19-21). T h e role of antibody would also
be compatible with the observed positive correlation between severity of arthritis and preexisting serum antibody levels.
A similar pathogenesis may underlie the
chronic arthritis produced in piglets by injection of live Mycoplasma hyorhinis (22). In this
arthritis, inflammatory activity continues long
after the disappearance of viable organisms,
suggesting that antigen may be sequestered locally. In this arthritis, synovial fluid antibody
titers were frequently higher than in serum,
indicating that a local immune response to the
Mycoplasma antigens was also present.
Following IA injection of antigen, there was
a rapid phase of immune elimination, thereafter, the remaining intraarticular antigen was
eliminated at a very slow rate. In the immune
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
ANTIGEN-INDUCED ARTHRITIS. I
rabbits, the half-life of the intraarticular antigen was estimated to be greater than 20 days.
These findings provide support for the concept
that the prolonged local immune response may
be a consequence of the persistence of antigen.
In another paper (18), evidence is provided that
the antigen is retained in the form of immune
complexes bound with complement components. This indicates a dual role for the intraarticular antigen: a) as a stimulus for the
prolonged local antibody synthesis and b) as a
source of antigen-antibody complexes mediating continuous inflammatory activity.
REFERENCES
1. Dumonde DC, Glynn L E : The production of
arthritis in rabbits by an immunological reaction
to fibrin. Br J Exp Pathol43:373-383, 1962
2. Glynn LE: T h e chronicity of inflammation and
its significance in rheumatoid arthritis. Ann
Rheum Dis27:105-121,1968
3. Glynn L E : Aetiology of rheumatoid arthritis
with regard to chronicity. Ann Kheum Dis
(Suppl) 28:3-4. 1969
4. Consden R, Doble A, Glynn LE, et al: Production of a chronic arthritis with Ovalbumin. Its
retention in rabbit knee joints. Ann Rheum Dis
30:307-3 IS, 197 1
5. Phillips J M , Kaklamanis P, Glynn L E : Experimental arthritis associated with autoimmunization to inflammatory exudates. Ann Rheum
Dis 2s: 165-174, 1966
6. Smile) JD, Sachs C, Ziff M : In vitro synthesis of
immunoglobulin by rheumatoid synovial membrane. ~JClin Invest 47:624-632, 1968
7. Sliwinski AS, Zvaifler N J : In vivo synthesis of
IgG by the rheumatoid synovium. J Lab Clin
bled 76:304-310, 1970
8. Vaughan J H , Barnett EV, Sobel MV, et al: Intracytoplasmic inclusions of immunoglobulins in
rheumatoid arthritis and other diseases. Arthritis Kheum 1 1 :125-134,1968
9, Hannestad K: Presence of aggregated globulin
in certain rheumatoid synovial effusions. Clin
Exp Immunol2:511-529, 1968
10. Winchester RJ, Agnello V, Kunkel H G : T h e
joint fluid and gamma globulin complexes and
their relationship to intraarticular complement
diminution. Ann NY Acad Sci 168:195-203.
1969
11. Kinsella T D , Baum J, Ziff M : Studies on isolated synovial lining cells of rheumatoid and
non-rheumatoid synovial membranes. Arthritis
Rheum 13:734-753,1970
12. Jasin H E , Ziff M : Immunoglobulin and specific
antibody synthesis in a chronic inflammatory
focus: antigen-induced synovitis. J Immunol
102:355-369, 1969
13. MacFarlane AS: In vivo behavior of
fibrinogen. J Clin Invest 42:346-361, 1963
14. Taurog A: Thyroid peroxidase and thyroxine
biosynthesis. Recent Progr H o r m Res
26:189-247, 1970
15. Minden P,Farr RS: T h e ammonium sulphate
method to measure antigen-binding capacity,
Handbook of Experimental Immunology. Edited
by I)M Weir. Great Britain, Blackwell Scientific Publications, 1967, pp 463-492
16. Askonas B, White R G : Sites of antibody production in the guinea-pig. T h e relation between
in vitro synthesis of anti-ovalbumin and yglobulin and distribution of antibody-containing
plasma-cells. Br J Exp Pathol 37:61-74, 1956
17. Webb FWS, Ford P M , Glynn L E : Persistence
of antigen in rabbit synovial membrane. Br J
Exp Pathol52:31-35, I971
18. Cooke T D , Hurd ER, Ziff M, et al: T h e pathogenesis of chronic inflammation in experimental antigen-induced arthritis. 11. Preferential
localization of antigen-antibody complexes to
collagenous tissues. J Exp Med 135:323, 1972
19. Ada GL, Nossal GJV, Austin CM: Antigens in
immunity. IV. Cellular localization of 12jl and
l 3 I I labeled flagella in lymph nodes. Aust J Exp
Biol Med Sci 42:311-330, 1964
20. White KG, French V, Stark J M : Germinal center formation and antigen localization in Malpighian bodies of the chicken spleen. Germinal
Centers in Immune Responses. Edited by H.
Cottier a n d W . O d a r t c h e n k o . New York,
Springer Verlag Inc, 1967, pp 131-144
21. Sordat B, Sordat M, Hess ,MW, et al: Specific
antibody within lymphoid germinal center cells
of mice after primary immunization with horseradish peroxidase: a light and electron microscopicstudy. J Exp Med 131:77-91, I970
22. Barden JA, Decker J L : Mycoplasma hyorhinis
swine arthritis. I. Clinical and microbiological
features. Arthritis Rheum 14: 193-20 1. 197 1
Arthritis and Rheumatism, Vol. 15, No. 4 (July-August 1972)
337