Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

Ann. rheum. Dis. (1963), 22, 1.

AN IMMUNOFLUORESCENCE STUDY OF RHEUMATOID FACTOR*

BY

J. N. McCORMICK Oxford Regional Rheumatic Diseases Research Centre, Stoke Mandeville Hospital, Aylesbury, Bucks.

The rheumatoid factor group of macroglobulins globulin apparently located rheumatoid factor at is a characteristic concomitant of rheumatoid sites in perivascular collagenous connective tissue in arthritis and can be detected readily in the serum rheumatoid synovium (Kaplan and Vaughan, 1959). of most rheumatoid patients by a variety of sero- More recently, Hess and Ziff (1961) have used logical techniques employing as reactant either fluorescent aggregated human gamma globulin to human gamma globulin (Heller, Jacobson, Kolodny, detect rheumatoid factor on leucocytes and platelets and Kammerer, 1954; Singer and Plotz, 1956; from patients with juvenile and adult rheumatoid Epstein, Johnson, and Ragan, 1956) or erythrocytes arthritis. sensitized with rabbit antibody (Waaler, 1940; The present investigation has been concerned

Rose, Ragan, Pearce, and Lipman, 1948). Various mainly with the distribution of rheumatoid factor copyright. workers applying the immunofluorescence technique in lymph nodes and synovium and, although similar of Coons and Kaplan (1950) have studied the techniques were employed, the results obtained are distribution of rheumatoid factor in the tissues of not entirely in agreement with those of Mellors and rheumatoid patients (Kaplan and Vaughan, 1959; his colleagues. In addition, further observations Mellors, Heimer, Corcos, and Korngold, 1959; have been made which may allow the elaboration Kaplan, Suchy, and Meyeserian, 1960; Mellors, of an hypothesis concerning the nature and bio-

Nowoslawski, Korngold, and Sengson, 1961b; logical function of rheumatoid factor. http://ard.bmj.com/ Mellors, Nowoslawski, and Korngold, 1961a). Employing aggregated human gamma globulin and Methods and Materials rabbit immune complexes labelled with fluorescent Preparation of Conjugates tracers and known to react with rheumatoid factor In earlier experiments, the protein reagents were in serological studies (Edelman, Kunkel, and conjugated with fluorescein isocyanate by the method of Franklin, 1958; Christian, 1958), Mellors and his Coons and Kaplan (1950), but at a later stage, con- associates located rheumatoid factor reactive with jugation with fluorescein isothiocyanate, as described by both types of reagent in plasma cells in synovium Riggs, Seiwald, Burckhalter, Downs, and Metcalf (1958), on September 27, 2021 by guest. Protected and lymph nodes as well as in the intrinsic cells of was used routinely. Conjugation with lissamine rhod- germinal centres in lymphoid follicles. From the amine B 200 was carried out by modification of the results of their mixed staining experiments with technique of Chadwick, McEntegart, and Nairn (1958). con- In each case, a protein concentration of 10 mg./ml. was contrastingly labelled reagents (1961b), they employed. Free dye was removed by absorption on a cluded that there were at least two types of cellular column of Dowex ion-exchange resin (1 x 10, 200-400 rheumatoid factor which occur separately or simul- mesh) equilibrated in 0O 1 M phosphate buffered isotonic taneously in individual plasma cells, but that only saline (pH-7-2) followed by dialysis against the same one or the other was present in germinal centres. buffer for 24 hours. All conjugates were absorbed twice Kaplan and his colleagues reported that rabbit with acetone-dried guinea-pig liver powder and three gamma globulin as well as aggregated human times with washed human AB cells, and in addition all gamma globulin reacted with plasma cells and except the anti-human globulin conjugates were absorbed germinal centre cells in rheumatoid lymph nodes with lyophilized whole human spleen. (Kaplan and others, 1960) and that rabbit gamma Fluorescent Aggregated Human Gamma Globulin * Based on a paper read at a meeting of the Heberden Society The human gamma globulin (HGG) used routinely on September 28, 1962. was obtained from a pool prepared by the method of Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

2 ANNALS OF THE RHEUMATIC DISEASES Kekwick and Mackay (1954). A small quantity of human Fluorescent Anti-human Globulin Antisera Cohn fraction 11 was available for comparison with (I) A potent anti-human globulin antiserum pool was this preparation and was found to give similar results. obtained from rabbits immunized with alum-precipitated The absorbed conjugates were heated in a water bath at human globulin. 63- C. for 5-10 minutes to aggregate the HGG and were centrifuged in the cold for 30 minutes at 16,000 r.p.m. (2) An anti-human gamma globulin (7S) antiserum before use. Sodium sulphate precipitation as used by was raised in rabbits against a pure preparation of 7S Mellors and others (1959) was not employed routinely as gamma globulin obtained by fractionation on DEAE the redissolved precipitates tended to be unstable. cellulose. Immunoelectrophoresis showed no evidence Similarly, the aggregates were unstable if the protein of other antibodies. concentration exceeded 10 mg./mi. or if the heating was (3) An anti-19S gamma globulin antiserum was prolonged. The reactivity of the heated HGG con- obtained from a rabbit injected with cryoglobulin from jugates with rheumatoid factor was assessed by their a patient with macroglobulinaemia and was absorbed ability in high dilution (e.g. 1/1,000) to inhibit the with pure 7S gamma globulin and the serum of a patient agglutination of sensitized sheep cells by two agglu- with hypogammaglobulinaemia. This antiserum gave a tinating doses of a standard rheumatoid serum. Heat reaction of identity with an antiserum raised against aggregation of the labelled HGG was not essential for normal 19S gamma globulin and was regarded as specific the subsequent staining reactions, however, as even the for this class of globulin. unheated conjugates produced convincing although less The labelled anti-human globulin antiserum (1) was intense staining. Presumably this was due to some used routinely for general assessment of the globulin denaturation of the HGG during conjugation and content of tissues, while the more immunospecific anti-7S dialysis, etc. and anti-19S gamma globulin antisera were reserved for detailed examination of selected tissues. Fluorescent Rabbit Globulin Reagents

Tissues Examined copyright. In preliminary experiments where the distribution of (1) Rheumatoid Biopsy Material. Fourteen lymph globulin in rheumatoid tissues was being studied, fluores- nodes (3 epitrochlear, 10 axillary, I cervical) from nine cent rabbit anti-human globulin antiserum exhaustively patients with sero-positive rheumatoid arthritis; three absorbed with normal human serum was used as a control subcutaneous nodules; two specimens of muscle; of specificity. Although this absorbed conjugate did not eighteen specimens of synovium obtained at arthrotomy react with normal human tissues, it was found that some from sixteen rheumatoid patients, one of whom was sero- staining persisted in rheumatoid tissues. When rheuma- negative although the diagnosis of rheumatoid arthritis toid serum was used to absorb the the residual antiserum, was unequivocal. http://ard.bmj.com/ staining was reduced but not abolished. It was thought that traces of soluble antigen-antibody complexes in the (2) Rheumatoid Post-mortenm Material. 41 lynph absorbed antiserum might be responsible for the reaction nodes obtained from two subjects with sero-positive with rheumatoid tissues but, as fluorescent normal rabbit arthritis. globulin produced similar staining, it appeared that (3) Control Material. Eighteen assorted lymph nodes aggregation of the rabbit globulin in an immune complex from five normal subjects, two mesenteric nodes from a was not essential for this reaction. case of rheumatic carditis, and two hilar nodes from In order to compare our findings with those of Mellors

two subjects with bronchial carcinoma were obtained at on September 27, 2021 by guest. Protected and others (1961a), however, soluble complexes of rabbit autopsy; ten specimens of normal synovium from ten anti-bovine albumin and labelled bovine albumin (BSA) meniscectomies; two specimens from osteo-arthritic were prepared according to their technique. In all, knees and one from the knee of a patient with psoriatic four types of rabbit globulin reagent were used: arthritis were obtained at exploratory arthrotomy. Small blocks of tissue were quick-frozen at 70- C. (l) Labelled rabbit anti-human globulin antiserum in a mixture of dry ice and acetone. The specimens absorbed with normal human serum; of synovium were suspended in 5 per cent. buffered normal rabbit gelatin during freezing, which facilitated subsequent (2) Labelled pooled globulin; sectioning of the blocks. All blocks were stored in the (3) Rabbit anti-BSA/labelled BSA complexes; deep-freeze at -20- C. until required. In addition, portions of each biopsy specimen and a selection of the (4) Labelled rabbit anti-BSA globulin, absorbed autopsy material were fixed in formalin and embedded in with normal human serum to remove cross- paraffin for routine histological staining with haematoxy- reacting antibodies (presumably traces of soluble lin and eosin and with methyl green-pyronin. Where antigen-antibody complexes may have been indicated, sections from the frozen blocks were also present). stained by routine histological procedures for comparison with adjacent fluorescent sections. The biopsied lymph The staining reactions described subsequently for the nodes were generally hyperplastic and showed well- rabbit globulin reagents apply equally to all four. marked follicle formation with prominent germinal Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

IMMUNOFLUORESCENCE STUDY OF RHEUMATOID FACTOR 3 centres, sinus hyperplasia, and numerous plasma cells. Fluorescence Microscopy The rheumatoid nodes obtained at autopsy showed The fluorescence microscope used in this department changes ranging from atrophy to moderate hyperplasia has been described previously by Scott (1960). and were comparable to the control material. The rheumatoid synovial specimens exhibited various degrees of rheumatoid synovitis. Results Fluorescent Aggregated Human Gamma Globulin. Fluorescence Staining and Inhibition Procedures -In sections of rheumatoid lymph nodes stained As required, sections approximately 4u in thickness with the fluorescent aggregate, bright specific were cut from the frozen tissue blocks in a cryostat at fluorescence was seen mainly in scattered cells - 180 C. and allowed to dry on slides at room tempera- lying in the medullary cords, commonly in a peri- ture. Fixation was not employed routinely. follicular distribution, and sometimes in the sinuses. These cells were in all respects similar to those (1) Single Staining Procedures.-In most cases the described in detail and illustrated by Mellors and sections were exposed to one or two drops of the appro- others (1959). Many of these appeared to be priate conjugate for 1 hour at room temperature, washed various types of plasma cell, Russell body cells in three changes of buffered saline, and then mounted being the most striking and relatively numerous in in buffered glycerol (90 per cent.). some sections. In most sections there were also (2) Mixed-Staining Procedures.-In direct mixed- ring type cells with a narrow rim of fluorescent staining procedures, the sections were exposed to a cytoplasm and a central unstained nucleus, resem- mixture of equal volumes of two contrastingly labelled bling small and large lymphocytes. These may be reagents for 1 hour. In sequential mixed-staining immature plasma cells, as suggested by the previous procedures, where the contrastingly labelled reagents workers, or similar to the "lymphocytoid plasma were applied consecutively, one was allowed to react cells" found in Waldenstrom's macroglobulinaemia copyright. with the sections for 18 hours before the second reagent (Dutcher and Fahey, 1959). was applied for 30 to 60 minutes, while on adjacent In the more hyperplastic lymph nodes, but rarely sections the sequence was reversed. Usually the sections in others, bright specific fluorescence was seen in were washed and examined under the microscope before germinal centres. On the few the second conjugate was applied. whole, comparatively In parallel experiments the direct and sequential mixed germinal centres reacted with the fluorescent staining procedures were repeated with the fluorescent aggregate, although abundant macroglobulin could

labels reversed. This step was included to ensure that be demonstrated in the majority of germinal centres http://ard.bmj.com/ the reactivity of each globulin preparation was not when sections previously exposed to the fluorescent influenced by the particular method of conjugation aggregate were counterstained with the specific employed. anti-19S antiserum labelled contrastingly. Rheu- matoid factor reactivity in germinal centres was (3) Direct Inhibition Procedures.-One drop of con- located in the cytoplasm of the large intrinsic cells jugate was mixed on the sections with one drop of the and in many sections the fluorescence was most corresponding unlabelled protein solution which had intense in the cells in the central portion of the

been applied 18 hours previously, and staining was on September 27, 2021 by guest. Protected allowed to proceed for 15 to 30 minutes. Under these germinal centre, tending to tail off peripherally in conditions complete or substantial inhibition was a fine cytoplasmic or pericytoplasmic reticulum obtained, and the staining observed on control sections which ended abruptly at the collar of mature where buffered saline was substituted for the unlabelled lymphocytes. Rheumatoid factor was not found protein was accepted as specific. By these criteria the in the peripheral rim of lymphocytes except in an fluorescence staining reactions reported in this study occasional cell identified as a plasma cell. In some were specific. instances, plasma cells including Russell body types exhibiting rheumatoid factor activity were located in (4) Cross-Inhibition Procedures. In conjunction with germinal centres where the intrinsic cells themselves the direct and sequential mixed staining procedures, each were devoid of rheumatoid factor. Although in of the conjugates under study was replaced in turn by the some sections where this was corresponding unlabelled protein and applied to the observed the plasma section 18 hours before, or as a mixture with, the other cells may have been imprints on the germinal centre, conjugate. In each case the sections were exposed to the in others counterstained with contrastingly labelled conjugate for 30 minutes and compared with control anti-human globulin antiserum the plasma cells sections on which buffered saline replaced the unlabelled appeared to be an integral part of the reactive protein. follicle. Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

ANNALS OF THE RHEUMATIC DISEASES In sections of rheumatoid synovium examined, aggregate was entirely intracellular apart from faint germinal centres were invariably absent, and rheu- traces in the lumen of blood vessels, and globules matoid factor activity was found mainly in mature and amorphous aggregates which may have been and immature plasma cell types lying in the sub- extruded by plasma cells or resulted from cell synovial connective tissue, occasionally in a peri- disintegration. Such extracellular deposits were vascular distribution. In the tissues showing the haphazard in distribution and did not appear to be most florid rheumatoid synovitis, with well-marked related specifically to any connective tissue element. follicle formation, large cellular aggregates of rheumatoid factor were sometimes seen. As might Fluorescent Rabbit Globulin Reagents.-When the be expected, in sections showing more chronic same tissues were studied with the various rabbit rheumatoid changes with less inflammatory cell globulin reagents, with some reservations, a similar infiltration cellular rheumatoid factor was often distribution of rheumatoid factor reactivity was very scanty, and in all instances cells showing encountered. Rheumatoid factor was found in constituted only a similar plasma cell types, but these were usually rheumatoid factor reactivity fewer in number than detected by the fluorescent small proportion of the total number of plasma aggregate in adjacent sections. On the other hand, cells present. Even so, in the one case of sero- rabbit negative rheumatoid arthritis examined, a few more germinal centres were stained by the faintly staining plasma cells were found in the globulins than by the fluorescent aggregate. In subsynovial connective tissue. In addition to some sections the number of germinal centres specific stained was comparable to the number seen in staining of plasma cell types, dull but adjacent sections exposed to the labelled anti-19S fluorescence was seen in the cytoplasm of the synovial antiserum. cells in several specimens, and this reactive material In addition to the two categories of cellular could not be removed by prior washing of the sec- factor reactivity tions or by irrigation of the tissue with buffered rheumatoid factor, rheumatoid

was detected by the rabbit globulin reagents at copyright. saline before quick-freezing. It is not clear whether media, and this was a diffusion artifact or perhaps represented various extracellular sites the lumen, factor globulin and adventitia of blood vessels, lining the walls of material containing rheumatoid sinuses, and as streaks applied to connective tissue absorbed by the phagocytic action of the synovial and among bundles of collagen. On the whole cells. these extracellular reactions were rather faint and Only limited information is available from the less striking than cellular staining, but they parallel- examination of the chronic subcutaneous nodules. led to some extent the extracellular distribution of On one occasion rheumatoid factor was detected in http://ard.bmj.com/ a few plasma cells in the peripheral zone of a nodule macroglobulin. but in most sections dull specific fluorescence was Inhibition Procedures.-In cross-inhibition pro- seen in the amorphous fibrinoid material in the cedures, staining by the fluorescent aggregate was central region. In all other tissues examined, usually inhibited to some extent, but never com- rheumatoid factor detectable by the fluorescent pletely by the rabbit anti-BSA globulin, the immune on September 27, 2021 by guest. Protected Dark-ground illumination and a water immersion objective were employed routinely for colour photography. All fluorescence photo- micrographs were taken on Super Anscochrome daylight type 35-mm. film. Exposure times ranged from 2.-10 minutes. Fig. 1.-Rheumatoid synovium. Collection of immature plasma Fig. 6.-Germinal centre in rheumatoid lymph node. Rhodamine- cells stained by fluorescein-labelled aggregate ( 160). labelled aggregate followed by fluorescein-labelled rabbit anti-BSA. Orange-red, green, and traces of yellow fluorescence shown. Green Fig. 2.-Germinal centre in rheumatoid lymph node. Fluorescein- mainly at periphery. ( - 160.) labelled aggregate followed by rhodamine-labelled anti-human Fig. 7.-Germinal centre in rheumatoid lymph node. Fluorescein- globulin antiserum. The intrinsic cells have not reacted with the labelled anti-19S antiserum followed by rhodamine-labelled anti-7S labelled aggregate. The yellow-green plasma cells (including Russell antiserum. Yellow mixture of fluorescence at periphery and mainly body types) contain rheumatoid factor ( x 160). green towards the central region. ( x 160.) Fig. 3.-Germinal centre in rheumatoid lymph node. Rhodamine- Fig. 8.-Periphery of germinal centre in rheumatoid lymph node. labelled aggregate followed by fluorescein-labelled anti-19S antiserum. Fluorescein-labelled anti-19S antiserum preceded rhodamine-labelled Yellow cells have reacted with both reagents. Mainly green anti-7S antiserum. Green and yellow mixed fluorescence shown fluorescence towards the periphery. (x 160.) together in individual intrinsic cells. ( x 240.) Fig. 9.-Rheumatoid lymph node. Fluorescein-labelled anti-19S Fig. 4.-Central region of large ovoid germinal centre in rheumatoid antiserum followed by rhodamine-labelled anti-7S antiserum. lymph node stained by rhodamine-labelled aggregate (x 160). Unusually large plasma cell shows yellow fluorescence in cytoplasm Fig. 5.-Same germinal centre as in Fig. 4. Rhodamine-labelled with traces of orange-red at periphery. The cell membrane has aggregate followed by rabbit anti-BSA/fluorescein-labelled BSA apparently ruptured above, releasing a shower of yellow and orange complex. Central region of germinal centre at right. Transition globules. Two larger Russell bodies on the right show yellow fluores- from orange-red to yellow-orange, yellow, yellow-green, and green cence. Cytoplasm of smaller plasma cell is mainly green, apart from from centre towards periphery. (x 160.) yellow specks at upper pole and yellow rim around nucleus. ( > 400.) Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

IMMUNOFLUORESCENCE STUDY OF RHEUMATOID FACTOR

2 copyright.

c. http://ard.bmj.com/ on September 27, 2021 by guest. Protected

7

Facing p. 4] Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

IMMUNOFLUORESCENCE STUDY OF RHEUMATOID FACTOR 5 complex, absorbed rabbit anti-human globulin TABLE antiserum, normal rabbit globulin, and a variety COLOUR OF FLUORESCENCE RELATED TO TYPE of immune rabbit globulins, including anti-human OF RHEUMATOID FACTOR ACTIVITY glomerulus, anti-human A cells, anti-sheep cells, Rheuma- anti-ox cells, and anti-dextran (Leuconostoc mesen- toid Fluorescence Specificity teroides). It was not possible to produce con- Factor vincing inhibition by pre-treatment of sections with TYPE I Orange-red Reaction with aggregated the rabbit anti-19S antiserum because the brief HGG only TYPE II Mixture varying from Reaction with aggregated application of the fluorescent aggregate apparently yellow-green, through HGG and rabbit globu- completed a fluorescent sandwich which reproduced yellow to yellow- lin (RGG) the pattern of staining seen in adjacent sections orange TYPE III Green Reaction with rabbit treated directly with the labelled anti-19S antiserum. globulin only (RGG) This difficulty was encountered with labelled aggre- gated Cohn fraction II as well as with the aggregate used routinely and, as the anti-19S antiserum had nodes and synovium, all three types of reactivity been absorbed free of cross-reactivity with 7S were seen in plasma cells. The Type I (HGG) gamma globulin, the staining was presumably due reaction was most common, but the Type II (mixed) to a small content of labelled macroglobulin in the reactions were found in moderate numbers of plasma gamma globulin preparations. It should be noted, cells, while Type III (RGG) reactivity was found however, that when the anti-19S antiserum was in relatively few cells. In germinal centres Type I omitted, staining produced by the fluorescent staining was occasionally seen alone, but it was more aggregate was more intense, although limited in common to find Type I with various combinations distribution, and it seems likely that the anti-19S of the other two. Type III reactivity was seen in antiserum produced relative inhibition despite the many germinal centres in the absence of either sandwich effect. In view of this difficulty and the Type I or II and was also found exclusively at fact that "non-specific" rabbit globulins inhibited extracellular connective tissue sites. On several copyright. staining, any inhibition produced by the specific occasions, in germinal centres where all three types anti-19S antiserum could not be interpreted directly of staining reaction were present, zoning of fluores- as demonstrating the macroglobulin nature of the cence was observed-Type I being present in the cellular proteins reacting with the fluorescent central portion of the germinal centre and Type II aggregate. in the intermediate zone with gradual transition Staining by the fluorescent rabbit globulins was to Type III at the periphery. Similarly, various usually inhibited by pre-treatment of the sections combinations of all three types were seen in indivi- http://ard.bmj.com/ with unlabelled aggregated human gamma globulin, dual germinal centre cells as well as in plasma but the degree of inhibition was variable and, cells where even Russell bodies in the same cell although often virtually complete, extracellular occasionally showed contrasting staining reactions. staining was not inhibited and in some sections In parallel mixed-staining experiments, where the plasma cells and germinal centres were still brightly labelled rabbit globulin reagents were followed by stained. On the other hand, complete inhibition the contrastingly labelled anti-19S antiserum, plasma was produced by a variety of normal and immune cells and germinal centre cells previously stained on September 27, 2021 by guest. Protected rabbit globulins so that again inhibition by the by the non-specific rabbit globulins showed mixed anti-19S antiserum was not conclusive evidence that fluorescence, indicating the macroglobulin nature the reactive tissue proteins were macroglobulins. of the Type III and presumably Type II factor. Similarly, when the labelled anti-19S antiserum followed contrastingly labelled aggregate, plasma Mixed-Staining Procedures.-When the rhod- cells and germinal centre cells previously stained amine-labelled aggregate and one or other of the by the aggregate now showed mixed fluorescence. fluorescein-labelled rabbit globulin reagents were This cannot be interpreted directly as demon- applied as a mixture or in sequence, three types of strating the macroglobulin character of the factor rheumatoid factor activity could be distinguished reacting with the aggregate, as the mixed staining according to the colours of the resulting fluorescence might be due to a Type 1I reaction with HGG and (Table). "non-specific" RGG in the rabbit antiserum. It These results were not influenced by reversing the may be significant, however, that it was commonly sequence in which the conjugates were applied or by the central portion of the germinal centre which interchanging the fluorescent labels. In lymph reacted with the aggregate, the peripheral zone being 2 Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

6 ANNALS OF THE RHEUMATIC DISEASES stained only by the anti-19S antiserum, while the in Fig. 9 were originally stained uniformly green by reverse was sometimes true when a labelled rabbit the fluorescein labelled anti-19S antiserum but, globulin reagent preceded the contrastingly labelled when identified again after exposure to the rhod- anti-19S antiserum. Although the macroglobulin amine labelled anti-7S antiserum, it was found that nature of the Type I factor could not be confirmed orange-red and various mixed hues had been added by the techniques employed in this investigation, to the original fluorescence. In comparison, when this can perhaps be assumed from the known the sequence was reversed, no green fluorescence distribution of macroglobulin in the tissues studied. was observed and cells or parts of cells were stained either yellow or orange-red. Distribution of 7S and 19S Gamma Globulins.- From these findings it appeared that, in many As fairly limited supplies of the immunospecific plasma and germinal centre cells, 7S gamma globulin anti-19S and anti-7S antisera were available, only was present in association or in combination with the most hyperplastic lymph nodes and florid macroglobulin, although in some instances it examples of rheumatoid synovitis were examined appeared that the two classes of protein were and compared with an appropriate selection of segregated in the cytoplasm in different parts of the control tissues. The labelled anti-7S antiserum same cell. It seems unlikely that the mixed staining produced a wide distribution of staining in the was due to a cross-reaction between the anti-7S majority of rheumatoid sections where high con- antiserum and the previously stained macroglobulin, centrations of globulin were found in most germinal because in these circumstances there would have centres and many plasma cells, while traces were been no green cells at all, and it is perhaps reasonable seen in the sinuses, walls, and lumen of vessels, and to suppose that bound anti-19S antibody blocked at most connective tissue sites. The labelled anti- any subsequent cross-reaction. Another possible 19S antiserum produced a similar distribution of explanation of this phenomenon is that, in cells extracellular staining, but on the whole fewer plasma containing rheumatoid factor macroglobulin reactive cells were stained. Most, if not all, of the germinal with RGG and previously stained by the fluoresceincopyright. centres present in many sections reacted with this labelled anti-19S antiserum, the appearance of antiserum. The distribution of 7S gamma globulin contrasting or mixed fluorescence after exposure to alone was determined by staining sections with the rhodamine labelled anti-7S antiserum might be fluorescent anti-7S antiserum after prior treatment due to a reaction between rheumatoid factor and the with unlabelled anti-19S to block the cross-reaction contrastingly labelled "non-specific" rabbit globulin of the anti-7S antiserum with macroglobulin. There component in the anti-7S antiserum. Against this was very little overall difference in the distribution explanation is the observation that prior treatment http://ard.bmj.com/ of staining produced by this technique compared of the sections with unlabelled anti-19S antiserum with the direct staining procedure and approxi- completely inhibited staining by the routine rabbit mately the same number of germinal centres and globulin reagents. Assuming that the antigenic and plasma cells were stained. It was significant, in antibody sites on the rheumatoid factor molecule view of subsequent findings, that there was apparent are separate and there is no steric hindrance, the inhibition of staining in the central portion of some specific antibody and the "non-specific" rabbit germinal centres. In sequential mixed-staining globulin components of the anti-19S antiserum may procedures where sections of lymph nodes and both be involved in this inhibition. It seems likely, on September 27, 2021 by guest. Protected synovium were pre-treated with fluorescein labelled therefore, that the mixed fluorescence observed with anti-19S antiserum followed by rhodamine labelled contrastingly labelled anti-19S and anti-7S antisera anti-7S antiserum, the resulting fluorescence indicates the coexistence of macroglobulin and 7S reactions were green in some germinal centres and gamma globulin in germinal centre cells and plasma plasma cells and orange-red in others, but many cells as well as at various extracellular connective showed a mixture of the two colours. Extracellular tissue sites. fluorescence was generally of the mixed variety. In some germinal centres, zoning of contrasting Control Materials.-With one exception, the fluorescence was again seen, with bright yellow and lymph nodes from the controls showed no specific traces of orange-red at the periphery and gradual reaction with the fluorescent aggregate or rabbit transition through yellow-green in the intermediate globulin reagents. In three nodes from one normal zone to green in the central region. Similarly, subject, plasma cell types and reticulo-endothelial various combinations of green, orange-red, and cells lining the sinuses reacted specifically with these mixed fluorescence were seen in individual germinal reagents, although there was no macroscopic centre cells and plasma cells. The cells illustrated evidence of rheumatoid arthritis and no previous Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

IMMUNOFLUORESCENCE STUDY OF RHEUMATOID FACTOR 7 history of rheumatic symptoms. No serology was otherwise, and it does not seem logical to suggest available, but in most subsequent cases sensitized that mixed staining is due to a factor which reacts sheep cell and latex-fixation tests were performed primarily with HGG and cross-reacts with RGG, on fresh or post mortem serum, all of which were as it would be expected that such cross-reactions negative. In one specimen of osteo-arthritic syno- would be blocked by prior exposure of the sections vium (sero-negative) a reaction was obtained between to the labelled aggregate. fibroblasts and the fluorescent aggregate, but no The mixed-staining experiments which suggested plasma cells were seen. The remainder of the tissues that 7S globulin might be associated in varying gave no reaction with the rheumatoid factor degrees with macroglobulin in germinal centre reagents. Globulin, especially macroglobulin, was cells and plasma cells might provide an alternative scanty in most control tissues and germinal centres explanation of the three types of rheumatoid factor rarely contained more than a trace if any at all. reactivity observed at similar sites. If the zoning of One of the two specimens of osteo-arthritic synovium contrasting fluorescence which was seen in some and the one psoriatic synovium showed moderate germinal centres in mixed staining with the anti-19S amounts of globulin bound at connective tissue and anti-7S antisera can be correlated with the sites, but this lacked rheumatoid factor activity. similar zoning of staining with the contrastingly labelled aggregate and rabbit globulin reagents, it Discussion would seem that, as the concentration of 7S gamma On the basis of results obtained from the cross- globulin decreased from the periphery inwards, inhibitions and mixed staining procedures with so the reactivity of rheumatoid factor with the contrastingly labelled aggregated human gamma fluorescent aggregate decreased from the centre globulin and the rabbit globulin reagents, it has been outwards. This might suggest that the Type I established that many plasma cells and fewer reaction was progressively inhibited as the con- germinal centre cells reacted with HGG alone centration of cellular 7S gamma globulin increased, (Type 1), more germinal centre cells and fewer while there was gradual transition through Type II plasma cells reacted with RGG alone (Type I1), reactivity in the intermediate zone to Type III at copyright. and moderate numbers of each reacted with both the periphery in direct proportion to the concen- HGG and RGG (Type II). In addition, the Type III tration of 7S gamma globulin. Similarly, this reaction occurred at scattered connective tissue sites. might apply to different sites in individual plasma Mellors and others (1961b) suggested that the cells. If this assumption is valid and the varying three types of fluorescence found in their mixed- reactivity of cellular rheumatoid factor depends staining experiments indicated that some plasma upon the ratio within the cell of 7S gamma globulin cells formed a rheumatoid factor detected only with to a "stem" 19S rheumatoid factor whose inherent http://ard.bmj.com/ fluorescent aggregate, some formed a second specificity is directed entirely against HGG, the demonstrable with fluorescent immune complex as combination of the 7S gamma globulin with the well as with fluorescent aggregate, and still others reactive sites on the 19S nucleus may confer upon formed both. Earlier (1961a), these authors had the rheumatoid factor additional affinity for rabbit suggested that there might be several (two or more) (heterologous) gamma globulin. This might occur factors directed against different antigenic com- if such 7S gamma globulin itself possessed an

ponents of aggregated human gamma globulin and inherent affinity for heterologous gamma globulin on September 27, 2021 by guest. Protected that some of these components were present in the and at the same time blocked sites on the rheumatoid rabbit immune complex. From the results of the factor nucleus reactive with further HGG. Alter- present investigation, however, where cells showing natively, the reaction of the 7S gamma globulin green fluorescence were found in sections exposed with the 19S nucleus might alter the configuration to the fluorescein labelled rabbit globulin reagents of either component so that sites reactive with following pre-treatment with rhodamine labelled heterologous gamma globulin were exposed. Najjar aggregate, it appears that there must also be a factor and Robinson (1959) have indicated that antigen- which reacts only with rabbit (heterologous) gamma antibody complexes may have antigenic deter- globulin. Furthermore, it is likely that the mixed minants not originally present in either component, fluorescence observed in these circumstances must and it is conceivable that rheumatoid factor be due either to the simultaneous presence of at reactivity with heterologous gamma globulin may be least two factors with individual specificity for HGG acquired in a similar fashion. If this were so, or RGG or to a single factor capable of reacting depending on the degree of aggregation with cellular with separate components in HGG and RGG. It 7S gamma globulin, the resulting partially-saturated would be difficult to explain this phenomenon complexes might react with both HGG and hetero- Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

8 ANNALS OF THE RHEUMATIC DISEASES logous gamma globulin, while further saturation modify the reaction with the antigen against which might produce a factor reactive with only hetero- the autoantibody is directed, perhaps by preventing logous globulin. In cells where "stem" 19S rheu- fixation of complement. It is also implicit in this matoid factor was present in the absence of 7S hypothesis that the reaction between rheumatoid gamma globulin, r'activity only with HGG may factor and autoantibody in the parent cell could not be encountered. Thus rheumatoid factor molecules result in damage, as such cells would not survive. could be produced with various sedimentation Not all 7S gamma globulin present in circulating coefficients depending on the degree of aggregation rheumatoid factor complexes need be autoantibody and with corresponding transition of reactivity. in view of the indication that isolated 19S rheuma- It is well known that high molecular weight com- toid factor can complex with normal gamma globulin plexes occur in rheumatoid sera, possibly with (Edelman, Kunkel, and Franklin, 1958). It is different molar ratios of 7 and 19S gamma globulin likely that 19S rheumatoid factor liberated from (Franklin, Holman, Muller-Eberhard, and Kunkel, Type I cells would be capable of forming equi- 1957), and it is conceivable that a proportion of these librium complexes with normal 7S gamma globulin complexes may be formed intracellularly. An in the serum, and that partially saturated complexes attempt such as this to equate the various types of rheumatoid factor and autoantibody released of cellular rheumatoid factor activity with the size from Type II cells may become saturated by aggrega- of the molecular complexes would not necessarily tion with normal 7S gamma globulin without be valid for the standard serological procedures displacement of the autoantibody. In addition, the where a dynamic equilibrium might exist, allowing ratio of normal to abnormal 7S in the complexes dissociation and re-association of complexes which may vary according to activity of the disease and could not occur at cellular level in tissue sections. the degree of antigenic stimulation. Further, Epstein and Tan (1961) have observed The results of further work around this theme dissociation of a 22S complex on anionic column (McCormick, to be published) indicate that absorp- chromatography and, if it were necessary for this tion of rheumatoid factor from selected rheumatoidcopyright. hypothesis to account for the sensitized sheep cell sera results in loss of antinuclear factor activity, agglutinating activity of 19S rheumatoid factor which suggests that rheumatoid factor-autoantibody isolated by such techniques (Lospalluto and Ziff, complexes are a possibility. Kaplan and Vaughan 1959), one might postulate that some hetero- (1959) suggested that rheumatoid factor and specificity had survived in partially dissociated reactant gamma globulin may occur bound in the complexes not resolved from "stem" 19S rheumatoid tissues, possibly in the form of complexes, and, on

factor in the ultracentrifuge. the basis of the assumptions outlined above, it is http://ard.bmj.com/ The possibility that rheumatoid factor may be an possible that the extracellular Type IIl reactions antibody-antibody complex has been considered by noted in this study represent saturated rheumatoid Najjar and Robinson (1959); Dubiski (1958) sug- factor complexes bound specifically to connective gested that rheumatoid factor was an anti-antibody tissue antigens. As a corollary, it is likely in inves- and others indicated that it might represent an tigations of this kind that fluorescent aggregated antibody to altered human gamma globulin (Glynn, human gamma globulin would not detect speci- fically bound rheumatoid factor.

Holborow, and Johnson, 1927). The presence of on September 27, 2021 by guest. Protected rheumatoid factor in plasma cells and the intrinsic Finally, the coexistence of 7S gamma globulin cells of germinal centres suggests though it does and macroglobulin in individual cells may indicate not confirm its antibody nature and it may be merely a maturation process in which macroglobulin possible to extend the interpretation of the results is being produced by the aggregation of 7S mole- of this investigation by postulating that "stem" cules. Deutsch and Morton (1957), from their rheumatoid factor arises in one cell or germinal studies on the dissociation of macroglobulins by centre in response to an antigenic stimulus from an mercapto-ethanol treatment, suggested that macro- abnormal 7S gamma globulin being produced at globulins might be formed from normal 7S sub-units. another site which may be distant or in the same Presumably, therefore, such aggregation could occur cell or germinal centre. If such abnormal gamma intracellularly but, on the whole, the hypothesis, globulins were, for instance, autoantibody directed however tenuous, which allows a protective bio- against connective tissue antigens, then rheumatoid logical role for rheumatoid factor seems more factor might play a protective role by complexing attractive. with such autoantibodies as they are produced. Summary When pre-formed complexes are eventually released Fluorescent aggregated human gamma globulin from the cell, the rheumatoid factor may prevent or and normal or immune rabbit globulins have been Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

IMMUNOFLUORESCENCE STUDY OF RHEUMATOID FACTOR 9 used to trace the distribution of rheumatoid factor Edelman, G. M., Kunkel, H. G., and Franklin, E. C. in rheumatoid lymph nodes and synovium. Rheu- (1958). J. exp. Med., 108, 105. matoid factor was detected by each class of reagent Epstein, W., Johnson, A., and Ragan, C. (1956). Proc. in plasma cells in lymph nodes and synovium, and Soc. exp. Biol. (N.Y.), 91, 235. Epstein, W. E., and Tan, M. (1961). Arthr. and Rheum., in the intrinsic cells of reactive follicles. The 4, 414 (Abstract). labelled rabbit globulin reagents also reacted with Franklin, E. C., Holman, H. R., Muller-Eberhard, H. J., rheumatoid factor at various extracellular con- and Kunkel, H. G. (1957). J. exp. Med., 105,425. nective tissue sites. Mixed-staining experiments Glynn, L. E., Holborow, E. J., and Johnson, G. D. (1957). with contrastingly labelled aggregated human gamma Proc. roy. Soc. Med., 50, 469. globulin and the rabbit globulins showed that some Heller, G., Jacobson, A. S., Kolodny, M. H., and plasma cells and intrinsic cells reacted with either Kammerer, W. H. (1954). J. Immunol., 72, 66. the aggregate or the rabbit globulins and that others Hess, E., and Ziff, M. (1961). Arthr. and Rheum., 4, 574. reacted with both. Various combinations of each Kaplan, M. H., Suchy, M. L., and Meyeserian, M. (1960). Ibid., 3, 272 (Abstract). fluorescence reaction were found in individual cells and Vaughan, J. H. (1959). Ibid., 2, 356 (Abstract). and germinal centres. Mixed-staining procedures Kekwick, R. A., and Mackay, M. E. (1954). M.R.C. with contrastingly labelled anti-7S and a specific Spec. Rep. Ser. No. 286. anti-19S antisera indicated that 7S gamma globulin Lospalluto, J., and Ziff, M. (1959). J. exp. Med., was associated with macroglobulin at similar sites, 110, 169. and it is suggested that the varying reactivity of Mellors, R. C., Heimer, R., Corcos, J., and Korngold, L. rheumatoid factor with the aggregate and rabbit (1959). Ibid., 110, 875. globulins depends upon the ratio of native 7S gamma Nowoslawski, A., and Korngold, L. (1961a). globulin to a "stem" 19S rheumatoid factor in Amer. J. Path., 39, 533. -,~ , , and Sengson, B. L. (1961b). J. exp. rheumatoid factor complexes. In addition, it was Med., 113, 475. postulated that rheumatoid factor might arise as Najjar, V. A., and Robinson, J. P. (1959). In "Immunity an antibody to autoantibody with which it com- and Virus Infection", ed. V. A. Najjar, p. 71. plexes intracellularly, preventing or modifying Wiley, New York. copyright. subsequent reactions with connective tissue antigens. Riggs, J. L., Seiwald, R. J., Burckhalter, J. H., Downs, Finally, the possibility that macroglobulins may be C. M., and Metcalf, T. G. (1958). Amer. J. formed from intracellular aggregation of 7S gamma Path., 34, 1081. globulin was considered. Rose, H. M., Ragan, C., Pearce, E., and Lipman, M. 0. (1948). Proc. Soc. exp. Biol. (N. Y.), 68, 1. I wish to thank A.G.S. Hill and Dr. D. G. Scott for Scott, D. G. (1960). Immunology, 3, 226. Singer, J. M., and Plotz, C. M. (1956). Amer. J. Med.,

much helpful discussion and criticism. http://ard.bmj.com/ I am grateful to Dr. J. F. Soothill, University of Bir- 21, 888. mingham, for providing the anti-7S and anti-19S gamma Waaler, E. (1940). Acta path. microbial. scand., 17, 172. globulin antisera, and to Dr. W. d'A. Maycock, Lister Discussion.-Dr. J. T. Scorr (Hammersmith): May Institute of Preventive Medicine, for supplying the I ask Dr. McCormick about the correlation of these human gamma globulin. findings with the presence of rheumatoid factor in the Dr. C. L. Greenbury kindly supplied the Cohn fraction serum? Has he found rheumatoid factor by these II and various rabbit antisera. techniques in patients with negative Waaler-Rose tests? The lissamine rhodamine B 200 was supplied by DR. MCCORMICK: Of the patients I have studied, all on September 27, 2021 by guest. Protected Imperial Chemical Industries, Dyestuffs Division. have had a positive Waaler-Rose test except one, who I am indebted to the Nuffield Foundation and the had definite rheumatoid arthritis, but negative sheep cell Empire Rheumatism Council whose grants to the agglutination and latex-fixation tests. A few plasma Rheumatism Research Centre enabled this work to be cells containing rheumatoid factor were found in syno- pursued. vium from this subject. Etude immunofluorescente du facteur rhumatismal REFERENCES REuSMi Chadwick, C. S., McEntegart, M. G., and Nairn, R. C. La globulin gamma humaine agglutinee et rendue (1958). Lancet, 1, 412. fluorescent, et des globulines de lapin, normales et Christian, C. L. (1958). J. exp. Med., 108, 139. immunes, furent utilisees pour suivre la distribution Coons, A. H., and Kaplan, M. H. (1950). Ibid., 91, 1. du facteur rhumatismal dans les ganglions lymphatiques Deutsch, H. F., and Morton, J. I. (1957). Science, et la synovie des rhumatisants. Le facteur rhumatismal fut ddc ld par chaque type de reactif dans des plasmo- 125, 600. cytes, ganglions lymphatiques et dans la synovie, ainsi Dubiski, S. (1958). Fol. biol., 6, 47. que dans les cellules intrinseques des follicules actives. Dutcher, T. F., and Fahey, J. L. (1959). J. nat. Cancer La globulin coloree de lapin reagit aussi avec le facteur Inst., 22, 887. rhumatismal a de differents endroits du tissu conjonctif Ann Rheum Dis: first published as 10.1136/ard.22.1.1 on 1 January 1963. Downloaded from

10 ANNALS OF THE RHEUMATIC DISEASES extracellulaire. Des experiences de coloration mixte, inmunes, fueron empleadas para investigar la distri- avec la globulin gamma agglutinee humaine color6e bucion del factor reumatoide en los ganglios linfaticos differemment des globulines de lapin, demontrerent que y la sinovia de enfermos reumatoides. El factor reuma- certains plasmocytes et cellules intrinseques reagirent toide fue evidenciado por cada clase de reactivo en soit avec la globulin humaine, soit avec les globulines plasmocitos, ganglios linfaticos y en la sinovia, asi de lapin, tandis que d'autres reagirent avec les deux. como en las celulas intrinsecas de los foliculos activos. On trouva de differentes combinaisons de chaque La globulin fluorescent reaccion6 tambien con el reaction de fluorescence dans des cellules individuelles et factor reumatoide en varios sitios del tejido conectivo des centres germinaux. En colorant differemment extracelular. Experiencias de coloracion mixta, tiniendo l'antiserum anti-S7 et l'antiserum specifique anti-19S, diferentemente la globulin gamma aglutinada humana on vit que la globulin gamma 7S se trouvait associee a y las globulinas de conejo, demostraron que ciertos la macroglobuline a des endroits similaires, et on pense plasmocitos y celulas intrinsecas reaccionan sea con que la reactivite variable du facteur rhumatismal avec la globulin humana sea con las globulinas de conejo, la globuline humaine et de lapin depend du rapport entre mientras que otros reaccionan con ambas globulinas. la globuline gamma 7S native et un facteur rhumatismal Varias combinaciones de cada reaccion de fluorescencia "tige" ("stem") 19S dans les complexes du facteur fueron encontradas en celulas individuales y en centros rhumatismal. De plus, on posture que le facteur germinales. Con coloraciones de contrast del anti- rhumatismal pourrait se presenter comme un anticorps suero anti-7S y del antisuero especifico anti-19S se pudo contre un autoanticorps, avec sequel il formerait des observar una asociaci6n de la globulin gama 7S con complexes intracellulaires, empechant ou modifiant la macroglobulina en sitios similares y se sugiere que la les reactions subsequentes avec les antigenes du tissu reactividad variable del factor reumatoide con las conjonctif. Finalement, on a considered la possibility que globulinas humana y de conejo depende de la relacion les macroglobulines seraient formees par l'aggregation entre la globulin gamma 7S nativa y un factor reuma- intracellulaire de la globulin gamma 7S. toide "tallo" ("stem") 19S en los complejos del factor reumatoide. Ademas, se postula que el factor reuma- toide pudiera presentarse como un anticuerpo contra un Estudio inmmunofluorescente del factor reumatoide autoanticuerpo y formar con el complejos intracelulares e impedir o modificar reacciones subsiguientes con SUMA10O antigenos del tejido conectivo. Por fin, se consider La globulin gamma humana, aglutinada y fluores- posible que las macroglobulinas sean formadas por una cente, asi como globulinas de conejo, normales e agregacion intracelular de la globulin gamma 7S. copyright. http://ard.bmj.com/ on September 27, 2021 by guest. Protected