Fibronectin in Rheumatoid and Non-Rheumatoid Arthritic Synovial Fluids and in Synovial Fluid Cryoproteins

Transcription

1 ANNALS OF CLINICAL AND LABORATORY SCIENCE, Vol. 12, No. 3 Copyright 1982, Institute for Clinical Science, Inc. Fibronectin in Rheumatoid and Non-Rheumatoid Arthritic Synovial Fluids and in Synovial Fluid Cryoproteins MARGARET LU-STEFFES, M.S., ALBERT J. IAMMARTINO, M.D., FRANK R. SCHMID, M.D., C. WILLIAM CASTOR, M.D., LYMAN DAVIS, M.S., RUTH ENTWISTLE, B.S., and BYRON ANDERSON, Ph.D.* Departments of Biochemistry, Medicine (Section of Arthritis-Connective Tissue Diseases), and Otolaryngology and Maxillofacial Surgery, Northwestern University Medical and Dental Schools, Chicago, IL ABSTRACT Concentrations of fibronectin, immunoglobulins G, M, and A, and C3 and C4 components of complement, and other plasma proteins were determ ined in synovial fluids from patients with rheumatoid arthritis (RA) and other diseases (non-ra). Fibronectin concentrations were two to three times greater in all synovial fluids than in plasma, and RA synovial fluids had a significantly higher mean concentration than non-ra fluids (883 p,g per ml vs. 588 (j.g per ml, respectively, p < 0.01). The mean concentrations of other synovial fluid constituents were less than their mean plasma concentrations. These results suggest that unlike other plasma constituents, either plasma fibronectin is concentrated in synovial fluids or that a substantial portion of synovial fluid fibronectin may be derived from synovial tissue cells. Both the C3 and C4 complement components were present in lower concentrations in RA than in non-ra synovial fluids. The C3 contents showed a statistically significant negative correlation with the fibronectin contents. Fibronectin was also found in all synovial fluid cryoprotein fractions tested, although its content varied greatly as a percent of the total cryoprotein protein (0.01 to 43 percent). The data show that fibronectin is a consistent constituent of synovial fluid cryoproteins in agreement with our previously reported finding that fibronectin is found in all serum cryoglobulin fractions tested. * Address reprint requests to: Dr. Byron Anderson, Department of Biochemistry, Northwestern University Medical School, Chicago, Illinois /82/ $01.20 Institute for Clinical Science, Inc.

2 Introduction Fibronectin (FN) is a large molecular weight glycoprotein (440,000 daltons) found as a normal plasma constituent. The FN exhibits binding properties with respect to a num ber of molecules including collagen, heparin, fibrin-fibrinogen, mucopolysaccharides, and gangliosides.34,37 Despite the many studies on FN, the biologic function(s) of the molecule remain uncertain, although it has been shown that FN exhibits an opsonic activity.17,21,23,26 It has also been suggested that FN and collagen syntheses are coordinated8,16 and that FN may play a role in matrix organization in tissue repair processes.16,24 In our laboratories it has been shown that FN is a component of serum cryoglobulins4 and other experimental evidence suggests that FN may interact with immunoglobulin14,27,36 and the Clq component of complement.2,11 There are many reports recording measurements of various constituents of synovial fluids.15,33 Normal synovial fluids contain many of the lower molecular weight components of plasma, predominantly albumin, and lesser quantities of large molecular weight proteins and those plasma molecules of asymmetric shape, e.g., fibrinogen. Synovial fluids from inflamed joints contain the larger molecular weight plasma proteins owing to the increased vascular permeability resulting from activations of the kinin and complement systems. Synovial fluids from individuals with rheumatoid arthritis (RA) compared to synovial fluids from other connective tissue diseases usually contain increased quantities of immunoglobulins and decreased amounts of the complem ent proteins. Immune complexes are also present and both RA and other inflammatory synovial fluids exhibit cryoprecipitate formation.19,20,39,40 In these studies the increased quantities of FN were reported in synovial fluids obtained from individuals with various connective tissue diseases. It was also shown that the SYNOVIAL FLU ID FIBRONECTIN 179 cryoprotein fractions of the synovial fluids have variable amounts of FN. Materials and Methods Of the synovial fluids examined, 10 were from individuals who satisfied established criteria25 for definite or classical RA; two synovial fluids were from individuals with gout, two with ankylosing spondylitis, one each with degenerative joint disease, psoriasis and sarcoid disease, and three non-ra specimens were monoarticular joint swelling which were either of traumatic or unknown causes. All the synovial fluids studied were obtained by joint aspiration for diagnostic purposes. At the time of joint aspiration, synovial fluid specimens were placed into tubes containing EDTA in order to prevent clotting, an aliquot removed for determination of the cell counts, and the specimen was carefully maintained at 37 until cellular components and debris were removed by centrifugation. A portion of the supernatant was then stored at -20 for subsequent study. Another portion of the supernatant was placed in an 80 water bath such that the temperature would rise to 56 in 5 to 10 minutes. The 56 temperature was then maintained for three minutes, and the tube placed immediately in an ice bath. The fibrinogen containing precipitate that formed was removed by centrifugation and the supernatant was kept at 4 for 48 hours to allow for cryoprecipitate formation. The cryoprotein fraction was then collected by centrifugation and washed twice with 0.01 M phosphate buffer, ph 7.4, containing 0.15 M NaCl (phosphate-buffered saline, PBS). For analyses, the cryoprotein was dissolved in 1M Nal in PBS. The FN in the original supernatants and in cryoprotein fractions was determ ined by a radioimmunoassay (RIA) as previously described.1,4 The reproducibility of the FN determinations was within 8 per

3 180 LU-STEFFES, IAMMARTINO, SCHMID, CASTOR, DAVIS, ENTW ISTLE, AND ANDERSON cent for the synovial fluid specimens assayed at different times. Total protein was determ ined by the method of Lowry et al.18 The concentrations of the other synovial fluid constitutents were determined by radial immunodiffusion (RID) * Rheumatoid factor (RF) activities were assayed for the synovial fluids diluted ten-fold in PBS by agglutination of IgG coated latex beads. Results In table I are given the individual synovial fluid FN quantities together with data on the cryoprotein fractions, and in table II are listed the concentration ranges and the mean values ± SD of FN and other synovial fluid constituents divided into RA and non-ra categories. The mean FN value reported for plasma is 330 fig per * Behring Diagnostics Tri-Partigen plates were used. ml,18 and for plasma specimens examined in our laboratory was 255 fig per ml. As seen in table I, all individual synovial fluid FN concentrations were greater than the expected plasma value. Mean values for RA and non-ra synovial fluids (table I) were 2.7 and 1.8 times the reported plasma value, respectively (table II). Six of the synovial fluid RA values were greater and only one RA value was less than the mean value for the 10 non-ra synovial fluids. The difference of the means of the RA versus the non-ra FN concentrations was statistically significant (p < 0.01, Student s t test). Both C3 and C4 concentrations were significantly lower in RA than in non-ra synovial fluids, a finding consistent with previous reports.15,33 Fibrinogen concentration was also lower in RA synovial fluids (p < 0.05). All other component concentrations were not significantly different in these two groups of fluids. TABLE I F ib r o n e c tin C o n ten ts o f S y n o v ia l F lu id s and S y n o v ia l F lu id C ry o p ro te in F r a c tio n s Fibronectin and Protein (vg) in Percent Percent Fibronectin Cryoprotein per ml of Synovial Fibronectin Synovial Content Leukocytes of Synovial Fluid: Fluid Fibronectin of Cyro- Fluid \ig/ml RF Cells/ml Fibronectin, Protein in Cryoprotein protein RA 1, , RA ND RA ND 4.7 1,' RA , , RA 1, ,250 ND ND RA 940 ND 1, RA < RA ,000 ND ND RA ND ND ND RA 1, , , Gout ND Gout , ND < 0.01 AS ND AS ND ND DJD ,900 ND ND Psoriasis , Sarcoid 440 ND 2, ND 0.05 Non-RA ND 0.04 Non-RA 900 ND 9, Non-RA ,100 ND ND RF = Rheumatoid factor activity RA = Rheumatoid arthritis AS = Ankylosing spondylitis DJD Degenerative joint disease Non-RA = Non-rheumatoid arthritis monoarticular joint effusions ND = Not determined

4 The calculated correlation coefficients for synovial fluid FN concentration were compared to the other constituents (table II). There was a negative correlation between the FN and the C3 and the C4 concentrations for both RA and non-ra synovial fluids, the r value for C3 vs. FN being SYNOVIAL FLU ID FIBRONECTIN 181 statistically significant. There were positive and statistically significant correlations of RA and non-ra synovial fluid FN with the transferrin concentrations, whereas for IgG a positive correlation was noted for RA and a negative correlation for non-ra synovial fluids. TABLE XI C o m p arativ e C o m p o sitio n s o f R heum atoid A r t h r i t i s (RA) and Non-RA S y n o v ia l F lu id s Concentrat ion Component Range, ug/m l Mean ± SD t Value* p Value r Valuet Fibronectin Non-RA (10) , IgG Non-RA (10)+ 5,105 4,360 19,170 15,610 10,831 ± 4,586 9,939 4, * IgM ,460 1,066 1, , ÏÏ IgA ,403 4,153 1,335 1,514 ± 841 ± 1, c*2 -Macroglobulin ± Transferrin RA (10)t 1,047-2,855 1,517 ± Non-RA (9)t 997-2,054 1,472 ± Fibrinogen RA (7)J ± Non-RA (9)1 17-2, ± Total protein/ RA (10)t ± 13.8 Non-RA (10)t ± Leukocyte cell counts RA (7) ,600** 8,745 ± 10, Non-RA (8)4: ,600 11,093 ± 10, Determined by Student's t test +The r values are the correlation coefficients calculated for fibronectin and each of the other synovial fluid components $The number of synovial fluid specimens tested p < 0.05 lip < o.oi /Concentration of total protein in mg/ml **Number/ml

5 182 LU -STEFFES, IAMMARTINO, SCHMID, CASTOR, DAVIS, ENTW ISTLE, AND ANDERSON All the cryoproteins tested, whether from RA or non-ra sources, contained assayable FN (table I). The amounts of FN in the cryoproteins, calculated as a percent of the synovial fluid FN content, varied from 0.01 to 43 percent. The fibrinogen contents of the original supernatants of the synovial fluids and the fluid fraction following the 56 C heat treatment as described in the Methods section were determined, and greater than 95 percent of the fibrinogen was removed by this procedure. Fibrinogen was detected in only one cryoprotein fraction; thus, the FN in the cryoprotein was not due to an association with fibrinogen. Known quantities of purified FN were added to several of the synovial fluid original supernatants and to cryoprotein fractions, and the specimens assayed in the RIA. The FN determ ined was equal to the amounts of FN in the specimens alone plus the amounts of FN added, indicating that the other constituents of the synovial fluids and cryoproteins did not affect the accuracy of the RIA method. Discussion These results show that the FN contents of synovial fluids were approximately two to three times greater than plasma concentrations. Although concurrent plasma FN analyses were not made for the persons listed in table I, it is unlikely that the high concentrations of FN in synovial fluid are related to high concentrations in plasma. Fyrand et al9 have shown that concentrations of FN in citrated plasma specimens were not significantly different betw een normals and those of typical rheumatoid arthritis. Similar results to those of Fyrand were obtained in our laboratory; the mean plasma FN concentrations for normals and connective tissue diseases were 255 ± 17 and 245 ± 35 jli.g per ml, respectively, (p > 0.4). It has been previously shown that FN is a major constituent of synovial tissuederived cells30,31 and that those cells secrete FN into the culture medium (unpublished results). Either plasma-derived FN is concentrated in synovial fluids or a major portion of the synovial fluid FN may be derived from synthesis and secretion from synovial lining cells. Also, the RA synovial fluids contained significantly higher contents of FN than the non-ra fluids which may be related to the hyperplasia of the synovial tissue characteristic of the RA pathology. The mean value of the fibrinogen contents of RA synovial fluids was significantly less than that of non-ra fluids perhaps owing to greater fibrin formation in RA synovial tissues or fluids. If so, a portion of the FN of RA synovial fluids may be expected to be associated with the fibrin because of the known binding of FN and fibrinogen-fibrin.22 These results are similar to those reported by Carsons et al6,7 who showed that FN concentrations in RA synovial fluids were approximately three times the mean plasma concentrations; however, their mean values for non- RA synovial fluid concentrations were similar to the plasma concentrations whereas all non-ra synovial fluid FN values reported here exceeded the mean plasma value. Carsons et al6 also reported that there were direct correlations between white blood cell counts and the FN concentrations. In the data reported here, no statistically significant correlation with FN and the cell counts was observed by us for either the RA or the non-ra synovial fluids. The positive correlation between FN and IgG content of RA synovial fluids may be reflective of the hypertrophy of RA synovial tissue and the local synthesis of IgG in the tissue.29,32 The complement components C3 and C4 were both significantly less in RA synovial fluids, and both exhibited a negative correlation with FN contents. These results may suggest that FN concentrations are in some way related to complement consumption as have been shown by us in other studies that

6 SYNOVIAL FLU ID FIBRONECTIN 183 FN interacts with immune system components ,12,14,27,36 There was also a positive correlation betw een synovial fluid FN and transferrin concentrations. Transferrin is a growth promoter for fibroblastic cell strains,10' 13 and the positive correlation may result from transferrin s effect on synovial lining cell synthesis of FN or on the degree of hyperplasia of the lining cells. It has previously been shown by us that FN was a constituent of serum cryoglobulins, that the FN was not loosely adherent to the cryoglobulin fraction, and that the FN precipitated with other cryoglobulin components on repeated dissolution and re-precipitation of the cryoglobulins.4 It is clear from the data of table I that FN is also a consistent constituent of all the synovial fluid cryoprotein fractions tested of either RA or non-ra origin. Ludivico and Myers19 and Marcus and Townes20 examined synovial fluid cryoprecipitates and showed that they contained IgG, IgM, and Ig complexes. Marcus and Townes also showed that fibrinogen was present in all cryoprecipitates. It is important to emphasize that fibrinogen was removed from the synovial fluids in this study before cryoprotein formation was allowed to proceed, and only one of the cryoproteins contained detectable fibrinogen. Thus, the presence of FN in the cryoproteins was probably not due to an association with fibrinogen. The FN contents of the cryoproteins varied greatly, in one specimen constituting 43 percent of the total protein. There were not sufficient amounts of the samples to test for other components. Because of the relative cold-insolubility of FN,38 it will be of interest to determine whether FN is necessary or potentiates synovial fluid cryoprotein formation. In this regard, Beaulieu et al5 recently reported that the IgG and IgM contents were markedly reduced in cryoprecipitates of both serum and synovial fluids when FN was removed from the specimens by gelatin-sepharose adsorption prior to cryoprecipitation. Furthermore, the protein in cryoprecipitates of a FN-depleted SLE serum and a RA synovial fluid were increased following reconstitution with purified FN. Additional data suggest that FN may associate with immunoglobulin of the serum cryoglobulins 36 and with Clq.2,11 W hether or not a similar association pertains for the synovial fluid cryoproteins remains to be determined. Our results and those of Carsons et al6 and Beaulieu et al5 indicate that FN may be an important factor in the formation of both serum cryoglobulins and synovial fluid cryoproteins. In two papers,28,35 the distribution of FN in non-inflammed synovium was shown to be localized to blood vessels and within synovial cells, whereas RA synovium and pannus were characterized by diffuse and prominent staining for FN, especially in and around proliferating synovial cells. Thus, the marked increase of FN in inflammed tissue may be derived from the plasma exudate and from local synthesis, both contributing to the high FN concentrations in synovial fluids. It is probable that the increased FN of RA synovial tissue and fluid is not specific to RA but is a response to tissue breakdown and repair concomitant with the inflammatory processes. The FN may have a function in the formation and clearance of immune complexes21,26 particularly those that are cryoprecipitable,4,36 but the mechanisms of interaction remain to be discerned. Acknowledgments These studies were supported by grants AM , AM and AM from the National Institutes of Health and by a grant from the Illinois Chapter of the Arthritis Foundation. References 1. A n d e r s o n, B., D o n a k o w s k i, C., E n t w i s t l e, R., and D a v is, L.: Reisolation of immunoreactive radioiodinated antigens using glutaraldehyde-insolubilized antibody preparations. J. Immunological Methods 36: , 1980.

7 184 LU -STEFFES, IAMMARTINO, SCHMID, CASTOR, DAVIS, ENTW ISTLE, AND ANDERSON 2. An d e r s o n, B., E n t w i s t l e, R., P u y a t, L., D a v is, L. E., and S c h m id, F. R.: Fibronectin associated with Clq in a Clq isolation procedure. Immunol. Common. 10: , A n d e r s o n, B., J a m e s o n, A., St e f f e s, M. L., and M ARTINCIC, R. R.: Purification and quantities of immunoglobulins of rheumatoid synovial tissues. Ann. Clin. Lab. Sci. 20:432-^438, An d e r s o n, B., R u c k e r, M., E n t w i s t l e, R., Sc h m id, F. R., and W o o d, G. W.: Plasma fibronectin is a component of cryoglobulins from patients with connective tissue and other diseases. Ann. Rheum. Dis. 40: , B e a u l ie u, A. D., Va l e t, J. P., and St r e v e y, J.: The influence of fibronectin on cryoprecipitate formation in rheumatoid arthritis and systemic lupus erythematosus. Arth. Rheum. 24: , C a r s o n s, S., M o s e s s o n, M. W., and D i a m o n d, H. S.: Detection and quantitation of fibronectin in synovial fluid from patients with rheumatic disease. Arth. Rheum. 24: , C a r s o n s, S., N a t a r a ja n, C., D r e w, H., and D ia m o n d, H.: Fibronectin in human synovial fluid. Arth. Rheum. 23:661, F o id a r t, J. M., B e r m a n, J. J., P a g l ia, L., R e n n a r d, S., A b e, S., P e r a n t o n i, A., and M a r t in, G. R.: Synthesis of fibronectin, 1aminin and several collagens by a liver-derived epithelial line. Lab. Invest. 42: , F y r a n d, P., M u n t h e, E., and S o l u m, N. O.: Studies on cold insoluble globulin. I. Concentrations in citrated plasma in rheumatic disorders. Ann. Rheum. Dis. 37: , H a m i l t o n, T. A., W a d a, H. G., and S u s s m a n, H. H.: Identification of transferrin receptors on the surface of human cultured cells. Proc. Nat. Acad. Sci. 76: , H a r r is, C. A., R o t h, S., S c h m id, F. R., and ANDERSON, B.: Binding of fibronectin to Clq; Inhibition of binding by aggregated IgG. Immunol. Comm. 20: , H a r r is, C. A., Sc h m id, F. R., E n t w i s t l e, R., D a v is, L., P u y a t, L., R o t h, S. A., and An d e r s o n, B.: Fibronectin associated with Clq in a Clq isolation procedure: possible interaction of fibronectin with complement function. Arth. Rheum. 24 :S67, H u t c h i n g s, S. E. and Sa t o, G. H.: Growth and maintenance of HeLa cells in serum-free medium supplmented with hormones. Proc. Nat. Acad. Sci. 75: , IAMMARTINO, A., SCHMID, F. R., DONAKOWSKI, C., a n d A n d e r s o n, B.: F ib r o n e c tin is a c o m p o n e n t o f s e r u m c ry o g lo b u lin s a n d s y n o v ia l flu id c r y o p r o te in s p o s s ib le a s s o c ia tio n w ith im m u n o g lo b u lin s. A rth. R h e u m. 23:694, JESSA R, A. R.: The study of synovial fluid. Arthritis and Allied Conditions, 8th ed. Hollander, J. L. and McCarthy, D. J., Jr., eds. Philadelphia, Lea and Febiger, 1972, pp K u r k i n e n, M., V a h e r i, A., R o b e r t s, P. J., and St e n m a n, S.: Sequential appearance of fibronectin and collagen in experimental granulation tissue. Lab. Invest. 43:47-51, L a n s e r, M. E., Sa b a, T. M., and Sc o v i l l, W. A.: Opsonic glycoprotein (plasma fibronectin) levels after burn injury. Ann. Surg. 292: , L o w r y, O. H., R o s e b r o u g h, N. J., F a r r, A. L., a n d Ra n d a l l, R. J.: P ro te in m e a s u re m e n t w ith th e F o lin p h e n o l re a g e n t. J. B iol. C h e m. 193: , L u d iv ic o, C. L. and M y e r s, A. R.: Survey of synovial fluid cryoprecipitates. Ann. Rheum. Dis. 39: , M a r c u s, R. L. a n d T o w n e s, A. S.: T h e o c c u r re n c e o f c ry o p ro te in s in sy n o v ial flu id ; T h e a s s o c ia tio n o f a c o m p le m e n t-fix in g a c tiv ity in rh e u m a to id sy n o v ia l flu id w ith c o ld p re c ip ita - b le p ro te in. J. C lin. In v e st. 50: , M o l n a r, J., G e l d e r, F. B., L a i, M. Z., St e f r i n g, G. E., C r e d o, R. B., and L o r a n d, L.: Purification of opsonically active human and rat cold-insoluble globulin (plasma fibronectin). Biochemistry 18: , M o s e s s o n, M. W. and U m f l e e t, R. A.: The cold-insoluble globulin of human plasma I. Purification, primary characterization and relationship to fibrinogen and other cold-insoluble fraction components. J. Biol. Chem. 245: , M o s h e r, D. F. and W i l l ia m s, E. M.: Fibronectin concentration is decreased in plasma of severely ill patients with disseminated intravascularcoagulation. J. Lab. Clin. M ed.92: , P e t t e r s s o n, E. E. and COLVIN, R. B.: Coldinsoluble globulin (fibronectin, LETS protein) in normal and diseased human glomeruli: Papain-sensitive attachment to normal glomeruli and deposition in crescents. Clin. Immunol. Immunopathol. 11: , Ro d n a n, G. P.: Criteria for diagnosis and classification of rheumatic diseases. 1. Diagnostic criteria for rheumatoid arthritic. J. Amer. Med. Assoc. 224: , Sa b a, T. M. and J a f f e, E.: Plasma fibronectin (opsonic glycoprotein): Its synthesis by vascular endothelial cells and role in cardiopulmonary integrity after trauma as related to reticuloendothelial function. Amer. J. Med. 68: , Sc h m i d, F. R., W o o d, G. W., and A n d e r s o n, B.: Plasma fibronectin in cryoprecipitates; possible association with immunoglobulins and/ or Clq. Clin. Res. 27:694A, Sc o t t, D. L., D e l a m e r e, J. P., and W a l t o n, K. W.: The distribution of fibronectin in the pannus in rheumatoid arthritis. Brit. J. Exp. Path. 62: , S l iw i n s k i, A. J. and Z v a i f l e r. N. J.: In vivo synthesis of IgG by rheumatoid synovium. J. Lab. Clin. Med. 76: , Sl o a n, T. B., M a r t in c ic, R. R., and An d e r s o n, B.: Synovial cell antigens differences of antigen compositions between rheumatoid arthritic (RA) and non-ra-derived synovial cells detected with anti-synovial cell sera. Exp. Cell. Biol. 49: , 1981.

8 SYNOVIAL FLU ID FIBRONECTIN Sl o a n, T. B., M a r t in c ic, R. R., and An d e r s o n, B.: Synovial cell antigens production of heterologous anti-human synovial cell sera and general reactivities of the antisera. Exp. Cell. Biol. 49:20-33, S m il e y, J. D., S a c h s, C., and Z i f f, M.: In vitro synthesis of immunoglobulin by rheumatoid synovial membrane. J. Clin. Invest. 47: , Sw a n n, D. A.: Macromolecules of synovial fluid. The Joints and Synovial Fluid, vol. I. Sokoloff, L., ed. New York, Academic Press, 1978, pp Va h e r i, A. and M o s h e r, D.: High molecular weight, cell surface-associated glycoprotein (fibronectin) lost in malignant transformation. Biochim. Biophys. Acta 526:1-25, Va r t io, T., V a h e r i, A., V o n E s s e n, R., I s o m a k i, H., and St e n m a n, S.: Fibronectin in synovial fluid and tissue in rheumatoid arthritis. Eur. J. Clin. Invest. 11: , W o o d, G. W., Ru c k e r, M., D a v is, J. W., E n t w i s t l e, R., and A n d e r s o n, B.: Interaction of plasma fibronectin with selected cryoglobulins. Clin. Exp. Immunol. 40: , Ya m a d a, K. M. and O l d e n, K.: Fibronectins adhesive glycoproteins of cell surface and blood. Nature 275: , Ya m a d a, K. M., Sc h l e s m g e r, D. H., Ke n e d y, D. W., and P a s t a n, I.: Characterization of a major fibroblast cell surfae glycoprotein. Biochemistry 26: , Z v a i f l e r, N. J.: Rheumatoid synovitis, an extravascular immune complex disease. Arth. Rheum. 27: , Z v a i f l e r, N. J. and G r e e n b e r g, P. D.: Immunopathology of rheumatoid arthritis. Mechanisms of Immunopathology. Cohen, S., Ward, P. A., and McCluskey, R. T., eds. New York, J. Wiley and Sons, 1979, pp