665 - F'LARE OF ANTIGEN-INDUCED ARTHRITIS IN MICE

AFTER INTRAVENOUS CHALLENGE

Kinetics of Antigen in the Circulation and Localization of Antigen in the Arthritic and Noninflamed Joint

JAN WILLEM LENS, WIM B. VAN DEN BERG, LEVINUS B. A. VAN DE PUTTE. and WIL A. ZWARTS

Intravenous injection of methylated bovine se- rum albumin (mBSA) in mice with unilateral, chronic, mBSA-induced arthritis has been shown to cause a flare of snnouldering arthritis without affecting the contralat- eral, noninflamed knee joint. We studied the kinetics of '251-labeled mBSA in the blood, and the accumulation of antigen in both arthritic and noninflamed joints. The bulk of the antigen was eliminated from the blood within 1&30 minutes, and accumulation of antigen in the joints occurred mainly within this period. The amounts of antigen found in the arthritic joints were of the same order of magnitude as the amounts found in nonin- flamed joints. Autoradiography of whole joint sections revealed that, in arthritic joints, antigen was located primlarily at the deep capillaries and large vessels, and in thie noninflamed joints, antigen was located at the small superficial capillaries. Antigen was handled by granulocytes in the arthritic joint and by synovial lining cells in the noninflamed joint. Our data indicate that tiny amounts of antigen reach the synovial stroma in both normal and arthritic joints but cause inflammation only in the arthritic joints, because of local hyper- reaclivity.

From the Department of Rheumatology , University Hospi- tal, St. Radboud, Nijmegen, The Netherlands.

Supported by a grant from the Nederlandse Vereniging tot Kheurnatiekbestrijding .

Jan Willem Lens, PhD (current address: Department of Periodontology, Faculty of Dentistry, University of Amsterdam, Amsterdam, The Netherlands); Wim B. van den Berg, PhD: Bio- chemist; Levinus B. A. van de Putte, MD: Professor of Rheu- matology; Wil A. Zwarts, Technician.

Address reprint requests to Dr. W. B. van den Berg, De- partment of Rheumatology, University Hospital, St. Radboud, Geert Grooteplein Zuid 8, 6525 GA Nijmegen, The Netherlands.

Submitted for publication May 23, 1985; accepted in revised form October 15, 1985.

A flare of chronic antigen-induced arthritis in mice has recently been demonstrated to occur after intravenous and oral administration of antigen (1,2). This reaction is demonstrable 6 hours after challenge and lasts for as long as 9 days. Histologic studies of such a joint have shown characteristics of an acute type of inflammation, i.e., an abundance of polymor- phonuclear cells in the infiltrate of the synovial tissue and in the exudate of the joint space, superimposed on the low-grade, chronic arthritis already present (1,3).

Studies on the mechanism involved in the flare reaction have shown that the chronically inflamed joint behaves as a hyperreactive area, probably due to the local presence of antigen-reactive cells in the synovial tissue infiltrate (4). Since the reaction is suppressed by pretreatment with anti-mouse lymphocyte serum, T lymphocytes are apparently operative in the induction of the flare (5) . In addition to local hyperreactivity of the chronically inflamed joint, an important principle in the induction of the flare is the leakage of antigen from the circulation into the chronically inflamed tissue.

In the present study, we investigated the arrival of the antigen at the joint and the way it is handled in the extravascular compartment, using radiolabeled antigen and autoradiography of whole joint sections. Data were obtained both on chronically inflamed and on noninflamed joints of mice with unilateral arthritis.

MATERIALS AND METHODS Animals. We used male C57.Bl0 mice aged 7-9

weeks and weighing 24-26 gm at the start of the immuniza- tion.

Iodination of antigen. lZ5Iodine labeling of methyl- ated bovine serum albumin (mBSA) was performed by the

Arthritis and Rheumatism, Vol. 29, No. 5 (May 1986)

666 LENS ET AL

Table 1. "'I-labeled methylated bovine serum albumin (mBSA) in immune and nonimmune mice*

Kinetics of antigen clearance in peripheral blood after intravenous injection of 300 pg of

Distribution of '''I % of (total loo%)$ % Amount of

Time administered 1'51-mBSA mBSA in after dose in total 5% % in the total blood

antigen blood protein- % cell- protein volume free bound fractions (CLP) challenge volumet bound

30 sec. 2 min.

10 min. 20 min. 30 min.

2 hrs. 6 hrs.

24 hrs.

36.0 26.0 84 15 0.7 100 65.0

6.8 66 32 2.3 100 14.0 4.3 59 39 2.3 84 6.5 2.7 (36) 24 74 0.3 70 1.4 2.7 (8.5) 15 84 0.1 25 0.5 3.1 (9.4) 5 95 0.1 10 0.1 0.2 (0.7) 36 64 0.1 10 0.1

* Values are the mean of measurements in 3 mice. t Total blood volume was calculated as one-fifteenth of the body weight (see ref. 8). Blood samples were collected by cardiac puncture. Numbers in parentheses are the values for nonimmune mice. $ Protein-bound and free "'1 were separated by Sephadex G25 fractionation. Cell-bound lZsI was measured in erythrocytes and leukocytes. See Materials and Methods for details. 5 Measured by immunoprecipitation with anti-mBSA antiserum. See Materials and Methods for details.

chloramine T method (6). '251-labeled mBSA was separated from free '"1 by Sephadex G25 fractionation.

Induction of arthritis and flare reaction. Arthritis was induced by injecting 60 pg of mBSA in 6 pl of saline into the right knee joint of mice that had been immunized 3 weeks earlier with mBSA emulsified in Freund's complete adjuvant (1,7). The contralateral knee joint, which served as a control, was injected with saline. In the chronic phase of the joint inflammation, 5 weeks after initial induction, the flare reac- tion was induced by intravenous injection of 300 pg of "'I-labeled mBSA (10 pCi = 0.37 MBq). Two days before challenge, potassium iodide (50 p g h l ) was added to the drinking water to prevent accumulation of free l2'I in the thyroid gland.

Measurement of mBSA in the blood. Blood samples were taken by cardiac puncture at various times after intravenous challenge with 300 pg of '251-labeled mBSA (10 pCi) and were collected in 3.8% sodium citrate solution to prevent clotting. At each time point, 3 mice were killed. The amount of radioactivity in the total blood volume was measured by gamma-counting 10-pl samples and was ex- pressed as the percentage of the administered dose at time 0, taking one-fifteenth of the body weight as the total blood volume (8). Blood samples were centrifuged, and the cell fractions were separated from the plasma.

The cells were washed twice with 20 volumes of saline to remove free '"I, and the radioactivity in the cell pellet was counted. Free '"1 in the plasma was separated frdm '"1 bound to high molecular weight material, by Sephadex G25 fractionation. We will refer to the high molecular weight fraction as the protein fraction. This pro- tein fraction probably consisted of '2SI-labeled mBSA and other '"I-labeled proteins (9).

The amount of '251-labeled mBSA in the protein fraction was determined by immunoprecipitation with anti- mBSA antiserum, which was produced in rabbits according to standard procedures (10). Anti-mBSA antiserum (100 1.1)

was added to 10 pl of the column protein fraction brought to an optimal mBSA concentration for precipitation (1 mg/ml). After incubation for 1 hour at 37°C and for 1 hour a t 4"C, the precipitate was spun down (30 minutes at 2 , 5 0 0 ~ and 4"C),

Figure 1. Autoradiograph of radiolabeled cells in the lumen of a synovial vessel (arrows), 30 minutes after intravenous injection of "'I-labeled methylated bovine serum albumin (hematoxylin and eosin stained, magnification x 250).

MU RINE ARTHRITIS FLARE 667

and the radioactivity in the pellet was counted. The percent mBSA corresponding with a given percentage of precipitate was read from a standard precipitation curve. The nonspe- cific precipitation was 10%.

Radioactivity measurements in the knee joint. At various times after injection of the radiolabeled antigen, mice were killed, and their knee joints were removed in toto by cutting the femur and the tibia at sites adjacent to the joint cavit,y. Bone marrow was removed from the extremities by extensive flushing with saline. Free "'1 released from the antigen after administration was removed from the joints by repealted washing with saline. This was performed after fixatiton in 4% phosphate buffered formalin to retain the labeled antigen.

The amount of '2sI-labeled mBSA in the entire joint (i.e., in the blood, the blood vessel wall, and the joint tissue) was measured by gamma-counting and was expressed as a percentage of the administered dose of radiolabeled mBSA. To calculate the amount of '"I-labeled mBSA associated with the joint tissue, i.e., in the blood vessel walls and the extravascular joint tissue, the radiolabel measurements of the total joint were corrected for the amount of iodide bound to protein or to cells in the total blood volume of the joint (see Table I ) .

The blood volumes of arthritic and noninflamed joints were measured in a separate group of mice with unilateral chronic arthritis, using "Cr-labeled erythrocytes. Briefly, 5'Cr-labeled erythrocytes were injected intrave- nously, and blood samples were then taken. Arthritic and noninflamed joints were immediately removed, flushed, and washed as described above. The joint blood volume was calculated from measurements of the amount of "Cr in blood samples and joints. The mean * SD joint blood volume in this group of mice with arthritic (40 days) and noninflamed joints appeared to be 3.0 * 0.3 pl and 2.7 2 0.2 pl, respectively (n = 5 ) .

Localization of antigen in the knee joint detected by autoradiography. Knee joints were removed in toto and fixed in 4% phosphate buRered formalin. After decalcification

with 5% formic acid, the joints were processed and embed- ded in paraffin. Sections (7 pm) of the whole knee were prepared and mounted on gelatin-coated slides. These were dipped in K5 emulsion (Ilford, Basildon, Essex, England) and exposed for 9 weeks. After this period, the slides were developed and stained with hematoxylin and eosin. Labeling intensity was graded on a 4-point scale (0, t, + , and + + 1. Two of the authors (JWL and WBvdB). who had n o knowl- edge of which experimental groups the mice came from, scored the sections. In order to identify the labeled cells in the autoradiographs, the metallic silver granules were re- moved by washing the sections with a 5% potassium fer- ricyanide solution, and the photographs taken before and after this treatment were compared.

RESULTS '*'I-labeled mBSA in the circulation. Blood sam-

ples were collected at various times after intravenous injection of "'I-labeled mBSA. Table I shows the rapid disappearance of the antigen from the circulation of immunized mice, as compared with nonimmune mice. In the first 20 minutes, most of the "'I-label was still bound to high molecular weight material, but thereafter, the percentage of free lZ5I increased mark- edly. Immune precipitation with anti-mBSA showed that in the first 30 minutes, the bulk of the radiolabeled high molecular weight material represented mBSA.

In addition to the free (or complexed) mBSA in the circulation, some of the antigen was found associ- ated with circulating cells (Table 1). Figure 1 shows an autoradiograph of a labeled cell in the lumen of a synovial blood vessel.

Amounts of '251-labeled mBSA in arthritic and noninflamed joints. Knee joints were removed in toto at various times after antigen challenge. To obtain a

Table 2. intravenous injection of "CI-labeled methylated bovine serum albumin (mBSA)*

Measurements of radioactivity in the arthritic (right) and the noninflamed (left) knee joints of mice at various times after

Radioactivity in the joints Time: after Arthritic knee Noninflamed knee

antigen Total Corrected Total Corrected challenge n amount value+ P$ amount valuet P+

- 30 sec. 2 8.6 2.3 7.3 I .6 2 min. 4 12.6 2 2.2 8.7 ? 1.5 0.01 9.6 t 2.6 6.1 t 1.7 0.0s

10 min. 4 12.6 f 3.9 11.8 i 3.7 N S 10.0 t 1.7 9.6 t 1.6 NS

30 min. 6 11.4 i_ 2.3 11.3 ? 2.3 NS 9.4 ? 1.8 9.3 t 1.8 NS

-

- 20 min. 5 13.4 t 1.7 12.9 ? 1.6 10.8 ? 3.0 10.1 ? 2.9 -

2 hrs. 7 9.6 t 2.8 9.5 2 2.7 0.05 6.8 t 1.7 6.7 2 1.7 0.05 6 hrs. 9 7.9 t 3.0 7.9 r 3.0 0.01 4.0 t_ 1.4 4.0 2 1.4 0.01

- 24 hrs. 2 0.9 0.9 0.6 0.6 -

* Values are % x lo-' of administered dose, mean t SD. t These values represent radiolabeled mBSA associated with the synovial tissue and the blood vessel walls (see Results for details). t: Significance versus 20-minute value (the highest level detected). by Student's 2-tailed t-test. NS = not significant.

668 LENS ET AL

X(xlO-')of administered dose of'251-mBSA 16 r- 14-

12 - 10 -

8-

I T

I/ I/ I

24 hrs 30 sec time after i.v injection

Figure 2. Measurements of radioactivity in the arthritic (right) knee (0) and the noninflamed (left) knee (0) at the indicated times after intravenous (i.v.) injection of 1251-labeled methylated bovine serum albumin (1251-mBSA; 300 pg). Values represent radiolabeled mBSA associated with the synovial tissue and the blood vessel walls (mean 2 SD) (see Table 2). * P < 0.05; **P < 0.01.

measure of the amount of '251-labeled mBSA associ- ated with the joint tissue, the bone marrow was removed from the open ends of the femur and tibia by extensive flushing. In addition, values were corrected for the amount of radiolabeled protein and cells present in the joint blood volume (see Materials and Methods). In view of the rapid disappearance of mBSA from the circulation (Table l ) , the accumula- tion of antigen into the joints probably occurs in the

early phase after intravenous challenge with antigen, and in this period, most of the radiolabel represents mBSA (Table 1). Table 2 shows the amounts of '251-labeled mBSA in the knee joints, with and without correction for the amount of radiolabel in the joint blood volume. The corrected values represent antigen extravasated into the joint tissue and antigen associ- ated with the walls of blood vessels.

At 30 seconds after antigen challenge, antigen was found in the right (arthritic) and the left (normal) joint (Table 2). The highest values (between k300-400 ng of mBSA) were found at 20 minutes for both arthritic and noninflamed joints, and the amounts were of the same order of magnitude (Figure 2). After this period, the amount of '251-labeled mBSA in the joints decreased; the decline was most pronounced in the left, noninflamed joint. At 2 hours and 6 hours postinjection, significantly lower levels were detected in noninflamed joints, when compared with arthritic joints (Figure 2).

Localization of '251-labeled mBSA in arthritic and noninflamed knee joints. Radiographs of whole joint sections were obtained at various times after antigen challenge. The most important findings are given in Table 3. At 2 minutes after intravenous injection, radiolabel was present in the lumen of blood vessels in the superficial and deep layers of the ar- thritic and noninflamed synovial tissue. In non- inflamed joints at 30 minutes postinjection, '251-labeled mBSA localized in and around the small capillaries, just beneath the synovial lining cells (Figure 3). The

Table 3. after intravenous injection of 1251-labeled methylated bovine serum albumin*

Localization of radiolabel in noninflamed (left) and arthritic (right) knee joints of mice

Time after antigen challenge Compartment, localization 2 min. 30 min. 2 hrs. 6 hrs. 24 hrs.

Blood Superficial capillaries Deep capillaries and large vessels

Synovial membrane Stroma Cells in infiltratet Lining cells

Joint space Scattered Cells in exudate

Periarticular tissue Cells in infiltrate

+ + I+ + + + I+ +

010 010 010

010 010

010

+ +I+ +I+ +

o/o 010 o/o 010 010

010

+Lk o/+ + 012 01 + + I0

f I0 010

01 +

010 010

010 o/+ + + I0

010 01+ + o/+ +

010 010

010 01 + o/o 010 01 + 01 +

~ ~ ~ ~ ~ ~~

* Results are left knee jointlright knee joint, scored on hematoxylin and eosin-stained sections, using a scale of 0, ?, + , and + + , where 0 = no staining and + + = intense staining. t Labeled cells were predominantly polymorphonuclear cells (see Figures 6A and B and 7A and B).

MURINE ARTHRITIS FLARE 669

hours after intravenous challenge with 1251-labeled mBSA, radiolabel was still present above the walls of capillaries and large vessels of the deep synovial tissue (Figure 6A) and above a few cells in the synovial tissue (Figure 6B). At 6 hours after challenge, 1251-labeled mBSA was predominantly found above cells. The labeling intensity varied from a few black dots to intensive blackening which covered the cell com- pletely. Heavily labeled cells were found scattered between unlabeled cells in the synovial tissue (Figure 7A), the joint cavity (Figure 7B), and the periarticular tissue.

Identification of labeled cells in the synovial tissue demonstrated that approximately 80% of them were polymorphonuclear cells (Figures 8A and B). At 24 hours postinjection, the radiolabel was sparsely detected above cells in the tissues.

Figure 3. Intercondylar synovial tissue (s) of a noninflamed knee joint, 30 minutes after intravenous injection of '251-labeled methyl- atad bovine serum albumin. Radiolabel is present in and around small capillaries (arrows), just beneath the synovial lining cells (SIC). js = joint space; f = femur (hematoxylin and eosin stained, magnification X 250).

lapel was found sparsely in the capillaries or larger vejssels of the deep layers of the synovial tissue.

In arthritic joints, '251-labeled mBSA was found to, be primarily localized in and around the capillaries an,d larger vessels of the deep layers of the synovial tissue (Figure 4). The capillaries of the subsynovial plt+xus were sparsely labeled.

Extravasated antigen was distributed differ- enfly in normal synovial tissue, as compared with chf-oriically inflamed synovial tissue (see Table 3). In nopinflamed joints 2 hours postinjection, 1251 accumu- latFd above the synovial lining cells and was diffusely present in the joint cavity ( ~ i ~ ~ ~ ~ 5 ) . These findings

antigen challenge, antigen could only be detected as a few black dots above some cells. In arthritic joints 2

,

Figure 4. Synovium (s) from an arthntic jolnt, 30 minutes after intravenous injection of '2SI-labeled methylated bovine serum albu-

synovial tissue. j s = joint space; = synovial lining cells (hematoxylin and eosin stained, magnification x 250).

were at hours postinjection* At 24 hours after min. Radiolabel Is present in the vessels in the deep layer of the

670 LENS ET AL

Figure 5. Section of a noninflamed joint, 2 hours after intravenous injection of 1251-labeled methylated bovine serum albumin. Radiola- be1 is found above synovial lining cells (SIC) and is diffusely present in thejoint space (js). c = capsule; f = femur (hematoxylin and eosin stained, magnification x 250).

Arthritis flare. Leakage of antigen in arthritic joints resulted in a flare of the chronic inflammation already present, whereas no sign of acute inflamma- tion was observed after leakage of antigen into noniriflamed joints. At 2 hours after intravenous chal- lenge, a few polymorphonuclear cells were present in the infiltrate of the arthritic joint synovial tissue (Fig- ure 6). At 6 hours, the numbers increased both in the synovial tissue (Figure 7A) and in the joint cavity (Figure 7B). This reaction was even more pronounced at 24 hours postinjection.

DISCUSSION Injection of a sufficient amount of mBSA into

the circulation of mice with chronic joint inflammation induces a flare of the smouldering joint inflammation,

without affecting the contralateral, noninflamed knee joint (1). Results of this study show that circulating mBSA reaches the synovial tissue of both the arthritic and the noninflamed joints. The leakage into the joints takes place mainly in the first 10-20 minutes after intravenous challenge with mBSA, and over this pe- riod, the amount of accumulated antigen seems to be of the same order of magnitude in the arthritic and the noninflamed joints. The location of mBSA at the synovial tissue and the reaction to extravascular mBSA were different between these 2 joints. In ar- thritic joints, antigen was found primarily in and around capillaries and larger vessels of the deep layers of the synovial membrane (Figures 4 and 6A) and was handled by polymorphonuclear cells in the infiltrate of the synovial tissue (Figures 6B and 7A) and in the exudate of the joint space (Figure 7B). In noninflamed joints, antigen was found above small capillaries of the subsynovial plexus and was handled by synovial lining cells (Figures 3 and 5).

The blood supply to the synovial membrane is rather complex. In general, the large synovial arteries run in the deeper layer of the membrane, close to the fibrous capsular tissue. They branch via precapillary arterioles, into capillary networks located immediately beneath the synovial lining. The capillaries join to postcapillary venules and end in veins.

Leakage of material from the blood into the synovial tissues has been described as coming from the lumen of venules or capillaries, and seems not to occur across the walls of precapillary arterioles (1 1 ) . Data on passage across vessel walls in the synovium seem conflicting. Schumacher, in his studies using rabbits (12), reported that large particles, such as carbon, leaked from deep large vessels (sizes were increased by histamine) and suggested that only smaller particles emerge across superficial capillaries. This opinion is in agreement with the results reported by Mitnick et a1 (13) in their experiments with rabbits. Those authors showed that horseradish peroxidase (a small protein with a molecular weight of t45,OOO) passed from the capillary lumen to the joint space, primarily across superficial fenestrated vessels. In contrast, Hamanishi et a1 (14), who studied rats, did not observe any leakage of the proteins catalase (MW 240,000, subunits 60,000) or horseradish peroxidase across superficial fenestrated vessels.

Our data may point to a different location in arthritic and noninflamed synovial tissue. Methylated BSA (MW 65,000) was found to be associated prima- rily with deep capillaries and larger vessels in the

MURINE ARTHRITIS FLARE 67 1

A B Fiigure 6. Synovium (s) from an arthritic joint, 2 hours after intravenous injection of '251-labeled methylated bovine serum albumin. Radiolabel is present in vessel walls of the deep synovial tissue (A), and above cells (arrows) (B). Note the presence of polymorphonuclear cells in the synovial tissue. c = capsule; f = femur; js = joint space. (Hematoxylin and eosin stained, magnification x 250.)

arthritic joint, and primarily with superficial capillaries in noiiinflamed joints. This suggests that the leakage through deep vessels or superficial vessels may be dependent on both the size of the particles and the presence of inflammation or tissue damage in the synovial membrane.

The antigen used in this study is a cationic protein with a high isoelectric point (PI -8.5). In general, charge-modified antigens are rapidly cleared from the circulation by the liver and the spleen (15-17). In addition, cationic proteins may bind to negatively charged tissues or cells (18,19). This prin- ciple may also explain the extremely rapid disappear- ance of the bulk of the mBSA (64% of administered label) within 30 seconds after injection. This rapid clearance may also be caused by some aggregation of mBSA occurring upon contact with serum proteins

(unpublished observations). Accelerated clearance of circulating mBSA in immune mice, due to the well- known immune elimination phenomenon, was still demonstrable for this antigen (Table 1).

The main deposition of antigen from the circu- lation probably takes place shortly after intravenous injection. Within 30 minutes, the amount of mBSA in the circulation of immune mice had already decreased to < 1 % of the initial dose, and our quantitative data (Table 2) did not show a net increase of antigen in the joint after this period of time. Similar data were obtained by Vogt et a1 (20) in their study of the binding of cationized ferritin to rat glomerular basement mem- brane. Results of their ligation experiments made it clear that cationic ferritin capable of binding to the glomerular basement membrane was removed from the circulation within 15 minutes. It seems likely that a

672 LENS ET AL

B Figure 7. Arthritic joint, 6 hours after antigen challenge. Radiolab- eled cells are present in the synovial tissue (A), and the joint space (B). The cells in the infiltrate of the synovial tissue and in the exudate of the joint space are predominantly polymorphonuclear cells. s = synovium; SIC = synovial lining cells; j s = joint space; f =

femur. (Hematoxylin and eosin stained, magnification x 250 in A and x 400 in B.)

radiographs showed that the vessel wall-associated antigen remained clearly present for 2 hours (Figure 6A), indicating that this part of the antigen is cleared only slowly.

Our immune mice possess moderate antibody titers (19). Shortly after intravenous injection, part of the antigen will probably be complexed with antibod- ies, and antigen entrance into the synovial tissue may, to some extent, reflect deposition of preformed com- plexes. Since no large differences were found in the amounts of antigen associated with the arthritic versus the noninflamed, control joints (Table 21, increased deposition of immune complexes in the arthritic joint, due to augmented vascular permeability or tissue damage, seems to be a minimal contribution.

B small part of the m~~~ sticks to the vessel wall

charge interactions (19); whereas, the bulk of the

Figure 8. Autoradiograph of synovial tissue, 6 hours after intrave- nous injection of IZ5I-labeled methylated bovine serum albumin (A),

(B). The radiolabeled cell is identified as a polymorphonuclear cell after intravenous Of due to and after removal ofthe "blackening" (see Materials and Methods)

. . .

antigen is rapidly cleared from the circulation. Auto- (arrow). (Hematoxylin and eosin stained, magnification x 400.)

MlLJRINE ARTHRITIS FLARE 673

Another important aspect which should be mentioned is in situ immune complex formation, which particularly may occur in the arthritic joint. In rabbit studies, Cooke and Jasin (21) have shown that 40410 of the plasma cells in the chronic infiltrate of arthritic joints manufacture antibodies specific for the arthritis-inducing antigen, thereby generating a high local antibody concentration. Plasma cells are found in artliritic joints of mice (3); however, their numbers weire fewer when compared with our findings in rabbits (unpublished observations), and nothing is known about their specificity. Whether the difference in anti- gen handling in the arthritic joint, compared with the contralateral, control joint, is related to differences in loc,al immune complex formation remains to be clari- fiedl.

Small quantitative differences in antigen pas- sage across the vessel wall may exist for arthritic and noninflamed joints. Between 30 minutes and 6 hours after antigen administration, a significant decline in the antigen content was found in the noninflamed joint, and only a small decline was noted in the arthritic joint (Table 2). Over this period, inflammation with abun- dant granulocytes developed in the arthritic joint, and inci-eased passage of vessel wall-associated antigen into the joint may be generated by a granulocyte- mediated increase in vascular permeability (22). In addition, the influx of granulocytes carrying radiolab- eled antigen may also contribute. The localization of single, heavily radiolabeled cells between numerous unlabeled granulocytes in the synovium (Figures 6B and 7A) makes local uptake of antigen in the synovium unlikely and suggests that these cells already con- tained antigen before entering. Whether this granulo- cyte-associated antigen contributes to the occurrence of the flare reaction remains to be clarified.

The flare of chronic arthritis occurred after intravenous injection of the antigen, without affecting the contralateral, nonarthritic joint. Earlier studies indicated that local hyperreactivity mediated by anti- gen-reactive T cells in the chronically inflamed synlovium is of prime importance in the arthritis flare phenomenon ( 5 ) . The present data show qualitative, but no large quantitative, differences in antigen arrival at the arthritic and noninflamed joints; thus, it is unlikely that increased antigen entrance in the arthritic joint is an important principle in the flare reaction of chronic arthritis. The amounts of antigen (<400 ng of mBSA) associated with the joints after intravenous injection of antigen (Table 2) were intermediate be- tween the minimal dose of intraarticularly injected antigen needed to induce arthritis in normal joints (1 pg) and that needed to elicit a flare in chronically

inflamed joints (10 ng) (4). This seems compatible with the induction of a flare after intravenous injection of antigen without affecting the contralateral, non- arthritic joint. It remains questionable, however, whether antigen associated with the synovium and vessel walls is equally effective in eliciting local in- flammation, as compared with antigen injected directly into the joint cavity.

Recently, we have reported that similar flare reactions were seen in mice with unilateral, chronic mBSA-induced arthritis after oral challenge with anti- gen (2), indicating the possibility that macromolecules can pass through the gastrointestinal mucosa and reach the joint tissue. This principle may have rele- vance for the perpetuation of some forms of human chronic arthritis.

ACKNOWLEDGMENTS The authors wish to thank Liduine van den Bers-

selaar for technical assistance, G. J. P. Grutters and P. B. Spaan for care of the animals, the Isotope Laboratory of the Department of Internal Medicine, and Marion Jansen for secretarial assistance.

1.

2.

3.

4,

5 .

6.

Van de Putte LBA, Lens JW, van den Berg WB: Exa- cerbations of antigen-induced arthritis after challenge with intravenous antigen. Immunology 49: 161-167, 1983 Lens JW, van den Berg WB, van de Putte LBA, van den Bersselaar L: Flare-up of antigen-induced ar- thritis after challenge with oral antigen. Clin Exp Im- munol 58:364371, 1984 Lens JW, van den Berg WB, van de Putte LBA: Flare- up of antigen-induced arthritis: the role of retained inflammatory cells in local hyperreactivity. Agents Ac- tions 15:99-101, 1984 Lens JW, van den Berg WB, van de Putte LBA: Flare- up of antigen-induced arthritis after challenge with intra- venous antigen. Clin Exp Immunol 55:287-294, 1984 Lens JW, van den Berg WB, van de Putte LBA: Flare- up of antigen-induced arthritis in mice after challenge with intravenous antigen: effects of pretreatment with cobra venom factor and antilymphocyte serum. Clin Exp Immunol57:520-528, 1984 Hunter HM, Greenwood FC: Preparation of iodine-13 1 labeled growth hormone of high specific activity. Nature 194:495496, 1962

I . Brackertz D, Mitchell GF, MacKay IR: Antigen- induced arthritis in mice. I. Induction of arthritis in various strains of mice. Arthritis Rheum 20:841-850, 1977

8. Wish L, Furth J , Storey RH: Direct determinations of plasma, cell and organ blood volumes in normal and hypervolemic mice. Proc SOC Exp Biol Med 74:644648, 1950

674 LENS ET AL

9.

10.

1 I.

12.

13.

14.

15.

16

Udall JN, Bloch KJ, Fritze L, Walker WA: Binding of exogenous protein fragments to native proteins: possible explanation for the overestimation of uptake of extrin- sically labeled macromolecules from the gut. Immunol-

Hudson L, Hay FC, editors: Practical Immunology. London, Blackwell Scientific, 1980, pp 8-9 Liew M, Dick WC: The anatomy and physiology of blood flow in a diarthrodial joint. Clin Rheum Dis

Schumacher HR: Fate of particulate material arriving at the synovium via the circulation: an ultrastructural study. Ann Rheum Dis 32:212-218, 1973 Mitnick H, Hoffstein S, Weissmann G: Fate of antigen after intravenous and intraarticular injection into rab- bits. Arthritis Rheum 21:918-929, 1978 Hamanishi C, Nagai J, Ito T: Electron microscopic study of the permeability of the synovial small vessels to protein molecules (abstract). Arthritis Rheum 19:800, 1976 Thorbecke GJ, Maurer PH, Benacerref B: The affinity of the reticulo endothelial system for various modified serum proteins. Br J Exp Pathol 41:190-193, 1960 Buys C: Uptake and degradation of circulating proteins by the liver (thesis). Veenstra Visser, Groningen, The Netherlands, 1976

ogy 42:251-257, 1981

7:131-148, 1981

17. Border WA, Ward HJ, Kamil ES, Cohen AH: Induction of membranous nephropathy in rabbits by administra- tion of an exogenous cationic antigen: demonstration of a pathogenetic role for electrical charge. J Clin Invest 69:451461, 1982

18. Van den Berg WB, van Beusekom HJ, van de Putte LBA, Zwarts WA, van der Sluis M: Antigen handling in antigen-induced arthritis in mice: an autoradiographic and immunofluorescence study using whole joint sec- tions. Am J Pathol 108:9-16, 1982

19. Vanden Berg WB, vande Putte LBA, Zwarts WA, Joosten LAB: Electrical charge of the antigen deter- mines intraarticular antigen handling and chronicity of arthritis in mice. J Clin Invest 74:1850-1859, 1984

20. Vogt A, Rohrbach R, Shirnizu F, Takamiya H, Batsford S: Interaction of cationized antigen with rat glomerular basement membrane: in situ immune complex forma- tion. Kidney Int 22:27-35, 1982

21. Cooke TD, Jasin HE: The pathogenesis of chronic inflammation in experimental antigen-induced arthritis. I. The role of antigen on the local immune response. Arthritis Rheum 15:327-337, 1972

22. Wedmore CV, Williams TJ: Control of vascular perme- ability by polymorphonuclear leukocytes in inflamma- tion. Nature 289:64&650, 1981