immunofluorescence study rheumatoid

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

AN IMMUNOFLUORESCENCE STUDY OFRHEUMATOID FACTOR*

BY

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

The rheumatoid factor group of macroglobulinsis a characteristic concomitant of rheumatoidarthritis and can be detected readily in the serumof most rheumatoid patients by a variety of sero-logical techniques employing as reactant eitherhuman gamma globulin (Heller, Jacobson, Kolodny,and Kammerer, 1954; Singer and Plotz, 1956;Epstein, Johnson, and Ragan, 1956) or erythrocytessensitized with rabbit antibody (Waaler, 1940;Rose, Ragan, Pearce, and Lipman, 1948). Variousworkers applying the immunofluorescence techniqueof Coons and Kaplan (1950) have studied thedistribution of rheumatoid factor in the tissues ofrheumatoid patients (Kaplan and Vaughan, 1959;Mellors, Heimer, Corcos, and Korngold, 1959;Kaplan, Suchy, and Meyeserian, 1960; Mellors,Nowoslawski, Korngold, and Sengson, 1961b;Mellors, Nowoslawski, and Korngold, 1961a).Employing aggregated human gamma globulin andrabbit immune complexes labelled with fluorescenttracers and known to react with rheumatoid factorin serological studies (Edelman, Kunkel, andFranklin, 1958; Christian, 1958), Mellors and hisassociates located rheumatoid factor reactive withboth types of reagent in plasma cells in synoviumand lymph nodes as well as in the intrinsic cells ofgerminal centres in lymphoid follicles. From theresults of their mixed staining experiments withcontrastingly labelled reagents (1961b), they con-cluded that there were at least two types of cellularrheumatoid factor which occur separately or simul-taneously in individual plasma cells, but that onlyone or the other was present in germinal centres.Kaplan and his colleagues reported that rabbitgamma globulin as well as aggregated humangamma globulin reacted with plasma cells andgerminal centre cells in rheumatoid lymph nodes(Kaplan and others, 1960) and that rabbit gamma

* Based on a paper read at a meeting of the Heberden Societyon September 28, 1962.

globulin apparently located rheumatoid factor atsites in perivascular collagenous connective tissue inrheumatoid synovium (Kaplan and Vaughan, 1959).More recently, Hess and Ziff (1961) have usedfluorescent aggregated human gamma globulin todetect rheumatoid factor on leucocytes and plateletsfrom patients with juvenile and adult rheumatoidarthritis.The present investigation has been concerned

mainly with the distribution of rheumatoid factorin lymph nodes and synovium and, although similartechniques were employed, the results obtained arenot entirely in agreement with those of Mellors andhis colleagues. In addition, further observationshave been made which may allow the elaborationof an hypothesis concerning the nature and bio-logical function of rheumatoid factor.

Methods and MaterialsPreparation of Conjugates

In earlier experiments, the protein reagents wereconjugated with fluorescein isocyanate by the method ofCoons and Kaplan (1950), but at a later stage, con-jugation with fluorescein isothiocyanate, as described byRiggs, Seiwald, Burckhalter, Downs, and Metcalf (1958),was used routinely. Conjugation with lissamine rhod-amine B 200 was carried out by modification of thetechnique of Chadwick, McEntegart, and Nairn (1958).In each case, a protein concentration of 10 mg./ml. wasemployed. Free dye was removed by absorption on acolumn of Dowex ion-exchange resin (1 x 10, 200-400mesh) equilibrated in 0O 1M phosphate buffered isotonicsaline (pH-7 -2) followed by dialysis against the samebuffer for 24 hours. All conjugates were absorbed twicewith acetone-dried guinea-pig liver powder and threetimes with washed human AB cells, and in addition allexcept the anti-human globulin conjugates were absorbedwith lyophilized whole human spleen.

Fluorescent Aggregated Human Gamma GlobulinThe human gamma globulin (HGG) used routinely

was obtained from a pool prepared by the method of

copyright. on D

ecember 13, 2021 by guest. P

rotected byhttp://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.22.1.1 on 1 January 1963. Dow

nloaded from

http://ard.bmj.com/

ANNALS OF THE RHEUMATIC DISEASES

Kekwick and Mackay (1954). A small quantity of humanCohn fraction 11 was available for comparison withthis preparation and was found to give similar results.The absorbed conjugates were heated in a water bath at63- C. for 5-10 minutes to aggregate the HGG and werecentrifuged in the cold for 30 minutes at 16,000 r.p.m.before use. Sodium sulphate precipitation as used byMellors and others (1959) was not employed routinely asthe redissolved precipitates tended to be unstable.Similarly, the aggregates were unstable if the proteinconcentration exceeded 10 mg./mi. or if the heating wasprolonged. The reactivity of the heated HGG con-jugates with rheumatoid factor was assessed by theirability in high dilution (e.g. 1/1,000) to inhibit theagglutination of sensitized sheep cells by two agglu-tinating doses of a standard rheumatoid serum. Heataggregation of the labelled HGG was not essential forthe subsequent staining reactions, however, as even theunheated conjugates produced convincing although lessintense staining. Presumably this was due to somedenaturation of the HGG during conjugation anddialysis, etc.

Fluorescent Rabbit Globulin Reagents

In preliminary experiments where the distribution ofglobulin in rheumatoid tissues was being studied, fluores-cent rabbit anti-human globulin antiserum exhaustivelyabsorbed with normal human serum was used as a controlof specificity. Although this absorbed conjugate did notreact with normal human tissues, it was found that somestaining persisted in rheumatoid tissues. When rheuma-toid serum was used to absorb the antiserum, the residualstaining was reduced but not abolished. It was thoughtthat traces of soluble antigen-antibody complexes in theabsorbed antiserum might be responsible for the reactionwith rheumatoid tissues but, as fluorescent normal rabbitglobulin produced similar staining, it appeared thataggregation of the rabbit globulin in an immune complexwas not essential for this reaction.

In order to compare our findings with those of Mellorsand others (1961a), however, soluble complexes of rabbitanti-bovine albumin and labelled bovine albumin (BSA)were prepared according to their technique. In all,four types of rabbit globulin reagent were used:

(l) Labelled rabbit anti-human globulin antiserumabsorbed with normal human serum;

(2) Labelled pooled normal rabbit globulin;

(3) Rabbit anti-BSA/labelled BSA complexes;

(4) Labelled rabbit anti-BSA globulin, absorbedwith normal human serum to remove cross-reacting antibodies (presumably traces of solubleantigen-antibody complexes may have beenpresent).

The staining reactions described subsequently for therabbit globulin reagents apply equally to all four.

Fluorescent Anti-human Globulin Antisera(I) A potent anti-human globulin antiserum pool was

obtained from rabbits immunized with alum-precipitatedhuman globulin.

(2) An anti-human gamma globulin (7S) antiserumwas raised in rabbits against a pure preparation of 7Sgamma globulin obtained by fractionation on DEAEcellulose. Immunoelectrophoresis showed no evidenceof other antibodies.

(3) An anti-19S gamma globulin antiserum wasobtained from a rabbit injected with cryoglobulin froma patient with macroglobulinaemia and was absorbedwith pure 7S gamma globulin and the serum of a patientwith hypogammaglobulinaemia. This antiserum gave areaction of identity with an antiserum raised againstnormal 19S gamma globulin and was regarded as specificfor this class of globulin.The labelled anti-human globulin antiserum (1) was

used routinely for general assessment of the globulincontent of tissues, while the more immunospecific anti-7Sand anti-19S gamma globulin antisera were reserved fordetailed examination of selected tissues.

Tissues Examined(1) Rheumatoid Biopsy Material. Fourteen lymph

nodes (3 epitrochlear, 10 axillary, I cervical) from ninepatients with sero-positive rheumatoid arthritis; threesubcutaneous nodules; two specimens of muscle;eighteen specimens of synovium obtained at arthrotomyfrom sixteen rheumatoid patients, one of whom was sero-negative although the diagnosis of rheumatoid arthritiswas unequivocal.

(2) Rheumatoid Post-mortenm Material. 41 lynphnodes obtained from two subjects with sero-positivearthritis.

(3) Control Material. Eighteen assorted lymph nodesfrom five normal subjects, two mesenteric nodes from acase of rheumatic carditis, and two hilar nodes fromtwo subjects with bronchial carcinoma were obtained atautopsy; ten specimens of normal synovium from tenmeniscectomies; two specimens from osteo-arthriticknees and one from the knee of a patient with psoriaticarthritis were obtained at exploratory arthrotomy.

Small blocks of tissue were quick-frozen at 70- C.in a mixture of dry ice and acetone. The specimensof synovium were suspended in 5 per cent. bufferedgelatin during freezing, which facilitated subsequentsectioning of the blocks. All blocks were stored in thedeep-freeze at -20- C. until required. In addition,portions of each biopsy specimen and a selection of theautopsy material were fixed in formalin and embedded inparaffin for routine histological staining with haematoxy-lin and eosin and with methyl green-pyronin. Whereindicated, sections from the frozen blocks were alsostained by routine histological procedures for comparisonwith adjacent fluorescent sections. The biopsied lymphnodes were generally hyperplastic and showed well-marked follicle formation with prominent germinal

2

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

IMMUNOFLUORESCENCE STUDY OF RHEUMATOID FACTOR

centres, sinus hyperplasia, and numerous plasma cells.The rheumatoid nodes obtained at autopsy showedchanges ranging from atrophy to moderate hyperplasiaand were comparable to the control material. Therheumatoid synovial specimens exhibited various degreesof rheumatoid synovitis.

Fluorescence Staining and Inhibition Procedures

As required, sections approximately 4u in thicknesswere cut from the frozen tissue blocks in a cryostat at- 180 C. and allowed to dry on slides at room tempera-ture. Fixation was not employed routinely.

(1) Single Staining Procedures.-In most cases thesections were exposed to one or two drops of the appro-priate conjugate for 1 hour at room temperature, washedin three changes of buffered saline, and then mountedin buffered glycerol (90 per cent.).

(2) Mixed-Staining Procedures.-In direct mixed-staining procedures, the sections were exposed to amixture of equal volumes of two contrastingly labelledreagents for 1 hour. In sequential mixed-stainingprocedures, where the contrastingly labelled reagentswere applied consecutively, one was allowed to reactwith the sections for 18 hours before the second reagentwas applied for 30 to 60 minutes, while on adjacentsections the sequence was reversed. Usually the sectionswere washed and examined under the microscope beforethe second conjugate was applied.

In parallel experiments the direct and sequential mixedstaining procedures were repeated with the fluorescentlabels reversed. This step was included to ensure thatthe reactivity of each globulin preparation was notinfluenced by the particular method of conjugationemployed.

(3) Direct Inhibition Procedures.-One drop of con-

jugate was mixed on the sections with one drop of thecorresponding unlabelled protein solution which hadbeen applied 18 hours previously, and staining was

allowed to proceed for 15 to 30 minutes. Under theseconditions complete or substantial inhibition was

obtained, and the staining observed on control sectionswhere buffered saline was substituted for the unlabelledprotein was accepted as specific. By these criteria thefluorescence staining reactions reported in this studywere specific.

(4) Cross-Inhibition Procedures. In conjunction withthe direct and sequential mixed staining procedures, eachof the conjugates under study was replaced in turn by thecorresponding unlabelled protein and applied to thesection 18 hours before, or as a mixture with, the otherconjugate. In each case the sections were exposed to theconjugate for 30 minutes and compared with controlsections on which buffered saline replaced the unlabelledprotein.

Fluorescence MicroscopyThe fluorescence microscope used in this department

has been described previously by Scott (1960).

Results

Fluorescent Aggregated Human Gamma Globulin.-In sections of rheumatoid lymph nodes stainedwith the fluorescent aggregate, bright specificfluorescence was seen mainly in scattered cellslying in the medullary cords, commonly in a peri-follicular distribution, and sometimes in the sinuses.These cells were in all respects similar to thosedescribed in detail and illustrated by Mellors andothers (1959). Many of these appeared to bevarious types of plasma cell, Russell body cellsbeing the most striking and relatively numerous insome sections. In most sections there were alsoring type cells with a narrow rim of fluorescentcytoplasm and a central unstained nucleus, resem-bling small and large lymphocytes. These may beimmature plasma cells, as suggested by the previousworkers, or similar to the "lymphocytoid plasmacells" found in Waldenstrom's macroglobulinaemia(Dutcher and Fahey, 1959).

In the more hyperplastic lymph nodes, but rarelyin others, bright specific fluorescence was seen ingerminal centres. On the whole, comparatively fewgerminal centres reacted with the fluorescentaggregate, although abundant macroglobulin couldbe demonstrated in the majority of germinal centreswhen sections previously exposed to the fluorescentaggregate were counterstained with the specificanti-19S antiserum labelled contrastingly. Rheu-matoid factor reactivity in germinal centres waslocated in the cytoplasm of the large intrinsic cellsand in many sections the fluorescence was mostintense in the cells in the central portion of thegerminal centre, tending to tail off peripherally ina fine cytoplasmic or pericytoplasmic reticulumwhich ended abruptly at the collar of maturelymphocytes. Rheumatoid factor was not foundin the peripheral rim of lymphocytes except in anoccasional cell identified as a plasma cell. In someinstances, plasma cells including Russell body typesexhibiting rheumatoid factor activity were located ingerminal centres where the intrinsic cells themselveswere devoid of rheumatoid factor. Although insome sections where this was observed the plasmacells may have been imprints on the germinal centre,in others counterstained with contrastingly labelledanti-human globulin antiserum the plasma cellsappeared to be an integral part of the reactivefollicle.

3

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


In sections of rheumatoid synovium examined,germinal centres were invariably absent, and rheu-matoid factor activity was found mainly in matureand immature plasma cell types lying in the sub-synovial connective tissue, occasionally in a peri-vascular distribution. In the tissues showing themost florid rheumatoid synovitis, with well-markedfollicle formation, large cellular aggregates ofrheumatoid factor were sometimes seen. As mightbe expected, in sections showing more chronicrheumatoid changes with less inflammatory cellinfiltration cellular rheumatoid factor was oftenvery scanty, and in all instances cells showingrheumatoid factor reactivity constituted only asmall proportion of the total number of plasmacells present. Even so, in the one case of sero-negative rheumatoid arthritis examined, a fewfaintly staining plasma cells were found in thesubsynovial connective tissue. In addition tostaining of plasma cell types, dull but specificfluorescence was seen in the cytoplasm of the synovialcells in several specimens, and this reactive materialcould not be removed by prior washing of the sec-tions or by irrigation of the tissue with bufferedsaline before quick-freezing. It is not clear whetherthis was a diffusion artifact or perhaps representedmaterial containing rheumatoid factor globulin andabsorbed by the phagocytic action of the synovialcells.Only limited information is available from the

examination of the chronic subcutaneous nodules.On one occasion rheumatoid factor was detected ina few plasma cells in the peripheral zone of a nodulebut in most sections dull specific fluorescence wasseen in the amorphous fibrinoid material in thecentral region. In all other tissues examined,rheumatoid factor detectable by the fluorescent

aggregate was entirely intracellular apart from fainttraces in the lumen of blood vessels, and globulesand amorphous aggregates which may have beenextruded by plasma cells or resulted from celldisintegration. Such extracellular deposits werehaphazard in distribution and did not appear to berelated specifically to any connective tissue element.

Fluorescent Rabbit Globulin Reagents.-When thesame tissues were studied with the various rabbitglobulin reagents, with some reservations, a similardistribution of rheumatoid factor reactivity wasencountered. Rheumatoid factor was found insimilar plasma cell types, but these were usuallyfewer in number than detected by the fluorescentaggregate in adjacent sections. On the other hand,more germinal centres were stained by the rabbitglobulins than by the fluorescent aggregate. Insome sections the number of germinal centresstained was comparable to the number seen inadjacent sections exposed to the labelled anti-19Santiserum.

In addition to the two categories of cellularrheumatoid factor, rheumatoid factor reactivitywas detected by the rabbit globulin reagents atvarious extracellular sites the lumen, media, andadventitia of blood vessels, lining the walls ofsinuses, and as streaks applied to connective tissueand among bundles of collagen. On the wholethese extracellular reactions were rather faint andless striking than cellular staining, but they parallel-led to some extent the extracellular distribution ofmacroglobulin.

Inhibition Procedures.-In cross-inhibition pro-cedures, staining by the fluorescent aggregate wasusually inhibited to some extent, but never com-pletely by the rabbit anti-BSA globulin, the immune

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 plasmacells stained by fluorescein-labelled aggregate ( 160).

Fig. 2.-Germinal centre in rheumatoid lymph node. Fluorescein-labelled aggregate followed by rhodamine-labelled anti-humanglobulin antiserum. The intrinsic cells have not reacted with thelabelled aggregate. The yellow-green plasma cells (including Russell

body types) contain rheumatoid factor ( x 160).Fig. 3.-Germinal centre in rheumatoid lymph node. Rhodamine-labelled aggregate followed by fluorescein-labelled anti-19S antiserum.Yellow cells have reacted with both reagents. Mainly green

fluorescence towards the periphery. (x 160.)Fig. 4.-Central region of large ovoid germinal centre in rheumatoid

lymph node stained by rhodamine-labelled aggregate (x 160).

Fig. 5.-Same germinal centre as in Fig. 4. Rhodamine-labelledaggregate followed by rabbit anti-BSA/fluorescein-labelled BSAcomplex. Central region of germinal centre at right. Transitionfrom orange-red to yellow-orange, yellow, yellow-green, and green

from centre towards periphery. (x 160.)

Fig. 6.-Germinal centre in rheumatoid lymph node. Rhodamine-labelled aggregate followed by fluorescein-labelled rabbit anti-BSA.Orange-red, green, and traces of yellow fluorescence shown. Green

mainly at periphery. ( - 160.)Fig. 7.-Germinal centre in rheumatoid lymph node. Fluorescein-labelled anti-19S antiserum followed by rhodamine-labelled anti-7Santiserum. Yellow mixture of fluorescence at periphery and mainly

green towards the central region. ( x 160.)Fig. 8.-Periphery of germinal centre in rheumatoid lymph node.Fluorescein-labelled anti-19S antiserum preceded rhodamine-labelledanti-7S antiserum. Green and yellow mixed fluorescence shown

together in individual intrinsic cells. ( x 240.)Fig. 9.-Rheumatoid lymph node. Fluorescein-labelled anti-19S

antiserum followed by rhodamine-labelled anti-7S antiserum.Unusually large plasma cell shows yellow fluorescence in cytoplasm

with traces of orange-red at periphery. The cell membrane hasapparently ruptured above, releasing a shower of yellow and orangeglobules. Two larger Russell bodies on the right show yellow fluores-cence. Cytoplasm of smaller plasma cell is mainly green, apart fromyellow specks at upper pole and yellow rim around nucleus. ( > 400.)

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


2

7

Facing p. 4]

c.

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

IMMUNOFLUORESCENCE STUDY OF RHEUMATOID FACTORcomplex, absorbed rabbit anti-human globulinantiserum, normal rabbit globulin, and a varietyof immune rabbit globulins, including anti-humanglomerulus, anti-human A cells, anti-sheep cells,anti-ox cells, and anti-dextran (Leuconostoc mesen-teroides). It was not possible to produce con-vincing inhibition by pre-treatment of sections withthe rabbit anti-19S antiserum because the briefapplication of the fluorescent aggregate apparentlycompleted a fluorescent sandwich which reproducedthe pattern of staining seen in adjacent sectionstreated directly with the labelled anti-19S antiserum.This difficulty was encountered with labelled aggre-gated Cohn fraction II as well as with the aggregateused routinely and, as the anti-19S antiserum hadbeen absorbed free of cross-reactivity with 7Sgamma globulin, the staining was presumably dueto a small content of labelled macroglobulin in thegamma globulin preparations. It should be noted,however, that when the anti-19S antiserum wasomitted, staining produced by the fluorescentaggregate was more intense, although limited indistribution, and it seems likely that the anti-19Santiserum produced relative inhibition despite thesandwich effect. In view of this difficulty and thefact that "non-specific" rabbit globulins inhibitedstaining, any inhibition produced by the specificanti-19S antiserum could not be interpreted directlyas demonstrating the macroglobulin nature of thecellular proteins reacting with the fluorescentaggregate.

Staining by the fluorescent rabbit globulins wasusually inhibited by pre-treatment of the sectionswith unlabelled aggregated human gamma globulin,but the degree of inhibition was variable and,although often virtually complete, extracellularstaining was not inhibited and in some sectionsplasma cells and germinal centres were still brightlystained. On the other hand, complete inhibitionwas produced by a variety of normal and immunerabbit globulins so that again inhibition by theanti-19S antiserum was not conclusive evidence thatthe reactive tissue proteins were macroglobulins.

Mixed-Staining Procedures.-When the rhod-amine-labelled aggregate and one or other of thefluorescein-labelled rabbit globulin reagents wereapplied as a mixture or in sequence, three types ofrheumatoid factor activity could be distinguishedaccording to the colours of the resulting fluorescence(Table).These results were not influenced by reversing the

sequence in which the conjugates were applied or byinterchanging the fluorescent labels. In lymph

TABLE

COLOUR OF FLUORESCENCE RELATED TO TYPEOF RHEUMATOID FACTOR ACTIVITY

Rheuma-toid Fluorescence Specificity

Factor

TYPE I Orange-red Reaction with aggregatedHGG only

TYPE II Mixture varying from Reaction with aggregatedyellow-green, through HGG and rabbit globu-yellow to yellow- lin (RGG)orange

TYPE III Green Reaction with rabbitglobulin only (RGG)

nodes and synovium, all three types of reactivitywere seen in plasma cells. The Type I (HGG)reaction was most common, but the Type II (mixed)reactions were found in moderate numbers of plasmacells, while Type III (RGG) reactivity was foundin relatively few cells. In germinal centres Type Istaining was occasionally seen alone, but it was morecommon to find Type I with various combinationsof the other two. Type III reactivity was seen inmany germinal centres in the absence of eitherType I or II and was also found exclusively atextracellular connective tissue sites. On severaloccasions, in germinal centres where all three typesof staining reaction were present, zoning of fluores-cence was observed-Type I being present in thecentral portion of the germinal centre and Type IIin the intermediate zone with gradual transitionto Type III at the periphery. Similarly, variouscombinations of all three types were seen in indivi-dual germinal centre cells as well as in plasmacells where even Russell bodies in the same celloccasionally showed contrasting staining reactions.

In parallel mixed-staining experiments, where thelabelled rabbit globulin reagents were followed bythe contrastingly labelled anti-19S antiserum, plasmacells and germinal centre cells previously stainedby the non-specific rabbit globulins showed mixedfluorescence, indicating the macroglobulin natureof the Type III and presumably Type II factor.Similarly, when the labelled anti-19S antiserumfollowed contrastingly labelled aggregate, plasmacells and germinal centre cells previously stainedby the aggregate now showed mixed fluorescence.This cannot be interpreted directly as demon-strating the macroglobulin character of the factorreacting with the aggregate, as the mixed stainingmight be due to a Type 1I reaction with HGG and"non-specific" RGG in the rabbit antiserum. Itmay be significant, however, that it was commonlythe central portion of the germinal centre whichreacted with the aggregate, the peripheral zone being

2

5

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

ANNALS OF THE RHEUMATIC DISEASESstained only by the anti-19S antiserum, while thereverse was sometimes true when a labelled rabbitglobulin reagent preceded the contrastingly labelledanti-19S antiserum. Although the macroglobulinnature of the Type I factor could not be confirmedby the techniques employed in this investigation,this can perhaps be assumed from the knowndistribution of macroglobulin in the tissues studied.

Distribution of 7S and 19S Gamma Globulins.-As fairly limited supplies of the immunospecificanti-19S and anti-7S antisera were available, onlythe most hyperplastic lymph nodes and floridexamples of rheumatoid synovitis were examinedand compared with an appropriate selection ofcontrol tissues. The labelled anti-7S antiserumproduced a wide distribution of staining in themajority of rheumatoid sections where high con-centrations of globulin were found in most germinalcentres and many plasma cells, while traces wereseen in the sinuses, walls, and lumen of vessels, andat most connective tissue sites. The labelled anti-19S antiserum produced a similar distribution ofextracellular staining, but on the whole fewer plasmacells were stained. Most, if not all, of the germinalcentres present in many sections reacted with thisantiserum. The distribution of 7S gamma globulinalone was determined by staining sections withfluorescent anti-7S antiserum after prior treatmentwith unlabelled anti-19S to block the cross-reactionof the anti-7S antiserum with macroglobulin. Therewas very little overall difference in the distributionof staining produced by this technique comparedwith the direct staining procedure and approxi-mately the same number of germinal centres andplasma cells were stained. It was significant, inview of subsequent findings, that there was apparentinhibition of staining in the central portion of somegerminal centres. In sequential mixed-stainingprocedures where sections of lymph nodes andsynovium were pre-treated with fluorescein labelledanti-19S antiserum followed by rhodamine labelledanti-7S antiserum, the resulting fluorescencereactions were green in some germinal centres andplasma cells and orange-red in others, but manyshowed a mixture of the two colours. Extracellularfluorescence was generally of the mixed variety.In some germinal centres, zoning of contrastingfluorescence was again seen, with bright yellow andtraces of orange-red at the periphery and gradualtransition through yellow-green in the intermediatezone to green in the central region. Similarly,various combinations of green, orange-red, andmixed fluorescence were seen in individual germinalcentre cells and plasma cells. The cells illustrated

in Fig. 9 were originally stained uniformly green bythe fluorescein labelled anti-19S antiserum but,when identified again after exposure to the rhod-amine labelled anti-7S antiserum, it was found thatorange-red and various mixed hues had been addedto the original fluorescence. In comparison, whenthe sequence was reversed, no green fluorescencewas observed and cells or parts of cells were stainedeither yellow or orange-red.From these findings it appeared that, in many

plasma and germinal centre cells, 7S gamma globulinwas present in association or in combination withmacroglobulin, although in some instances itappeared that the two classes of protein weresegregated in the cytoplasm in different parts of thesame cell. It seems unlikely that the mixed stainingwas due to a cross-reaction between the anti-7Santiserum and the previously stained macroglobulin,because in these circumstances there would havebeen no green cells at all, and it is perhaps reasonableto suppose that bound anti-19S antibody blockedany subsequent cross-reaction. Another possibleexplanation of this phenomenon is that, in cellscontaining rheumatoid factor macroglobulin reactivewith RGG and previously stained by the fluoresceinlabelled anti-19S antiserum, the appearance ofcontrasting or mixed fluorescence after exposure tothe rhodamine labelled anti-7S antiserum might bedue to a reaction between rheumatoid factor and thecontrastingly labelled "non-specific" rabbit globulincomponent in the anti-7S antiserum. Against thisexplanation is the observation that prior treatmentof the sections with unlabelled anti-19S antiserumcompletely inhibited staining by the routine rabbitglobulin reagents. Assuming that the antigenic andantibody sites on the rheumatoid factor moleculeare separate and there is no steric hindrance, thespecific antibody and the "non-specific" rabbitglobulin components of the anti-19S antiserum mayboth be involved in this inhibition. It seems likely,therefore, that the mixed fluorescence observed withcontrastingly labelled anti-19S and anti-7S antiseraindicates the coexistence of macroglobulin and 7Sgamma globulin in germinal centre cells and plasmacells as well as at various extracellular connectivetissue sites.

Control Materials.-With one exception, thelymph nodes from the controls showed no specificreaction with the fluorescent aggregate or rabbitglobulin reagents. In three nodes from one normalsubject, plasma cell types and reticulo-endothelialcells lining the sinuses reacted specifically with thesereagents, although there was no macroscopicevidence of rheumatoid arthritis and no previous

6

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


history of rheumatic symptoms. No serology wasavailable, but in most subsequent cases sensitizedsheep cell and latex-fixation tests were performedon fresh or post mortem serum, all of which werenegative. In one specimen of osteo-arthritic syno-vium (sero-negative) a reaction was obtained betweenfibroblasts and the fluorescent aggregate, but noplasma cells were seen. The remainder of the tissuesgave no reaction with the rheumatoid factorreagents. Globulin, especially macroglobulin, wasscanty in most control tissues and germinal centresrarely contained more than a trace if any at all.One of the two specimens of osteo-arthritic synoviumand the one psoriatic synovium showed moderateamounts of globulin bound at connective tissuesites, but this lacked rheumatoid factor activity.

DiscussionOn the basis of results obtained from the cross-

inhibitions and mixed staining procedures withcontrastingly labelled aggregated human gammaglobulin and the rabbit globulin reagents, it has beenestablished that many plasma cells and fewergerminal centre cells reacted with HGG alone(Type 1), more germinal centre cells and fewerplasma cells reacted with RGG alone (Type I1),and moderate numbers of each reacted with bothHGG and RGG (Type II). In addition, the Type IIIreaction occurred at scattered connective tissue sites.

Mellors and others (1961b) suggested that thethree types of fluorescence found in their mixed-staining experiments indicated that some plasmacells formed a rheumatoid factor detected only withfluorescent aggregate, some formed a seconddemonstrable with fluorescent immune complex aswell as with fluorescent aggregate, and still othersformed both. Earlier (1961a), these authors hadsuggested that there might be several (two or more)factors directed against different antigenic com-ponents of aggregated human gamma globulin andthat some of these components were present in therabbit immune complex. From the results of thepresent investigation, however, where cells showinggreen fluorescence were found in sections exposedto the fluorescein labelled rabbit globulin reagentsfollowing pre-treatment with rhodamine labelledaggregate, it appears that there must also be a factorwhich reacts only with rabbit (heterologous) gammaglobulin. Furthermore, it is likely that the mixedfluorescence observed in these circumstances mustbe due either to the simultaneous presence of atleast two factors with individual specificity for HGGor RGG or to a single factor capable of reactingwith separate components in HGG and RGG. Itwould be difficult to explain this phenomenon

otherwise, and it does not seem logical to suggestthat mixed staining is due to a factor which reactsprimarily with HGG and cross-reacts with RGG,as it would be expected that such cross-reactionswould be blocked by prior exposure of the sectionsto the labelled aggregate.The mixed-staining experiments which suggested

that 7S globulin might be associated in varyingdegrees with macroglobulin in germinal centrecells and plasma cells might provide an alternativeexplanation of the three types of rheumatoid factorreactivity observed at similar sites. If the zoning ofcontrasting fluorescence which was seen in somegerminal centres in mixed staining with the anti-19Sand anti-7S antisera can be correlated with thesimilar zoning of staining with the contrastinglylabelled aggregate and rabbit globulin reagents, itwould seem that, as the concentration of 7S gammaglobulin decreased from the periphery inwards,so the reactivity of rheumatoid factor with thefluorescent aggregate decreased from the centreoutwards. This might suggest that the Type Ireaction was progressively inhibited as the con-centration of cellular 7S gamma globulin increased,while there was gradual transition through Type IIreactivity in the intermediate zone to Type III atthe periphery in direct proportion to the concen-tration of 7S gamma globulin. Similarly, thismight apply to different sites in individual plasmacells. If this assumption is valid and the varyingreactivity of cellular rheumatoid factor dependsupon the ratio within the cell of 7S gamma globulinto a "stem" 19S rheumatoid factor whose inherentspecificity is directed entirely against HGG, thecombination of the 7S gamma globulin with thereactive sites on the 19S nucleus may confer uponthe rheumatoid factor additional affinity for rabbit(heterologous) gamma globulin. This might occurif such 7S gamma globulin itself possessed aninherent affinity for heterologous gamma globulinand at the same time blocked sites on the rheumatoidfactor nucleus reactive with further HGG. Alter-natively, the reaction of the 7S gamma globulinwith the 19S nucleus might alter the configurationof either component so that sites reactive withheterologous gamma globulin were exposed. Najjarand Robinson (1959) have indicated that antigen-antibody complexes may have antigenic deter-minants not originally present in either component,and it is conceivable that rheumatoid factorreactivity with heterologous gamma globulin may beacquired in a similar fashion. If this were so,depending on the degree of aggregation with cellular7S gamma globulin, the resulting partially-saturatedcomplexes might react with both HGG and hetero-

7

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


logous gamma globulin, while further saturationmight produce a factor reactive with only hetero-logous globulin. In cells where "stem" 19S rheu-matoid factor was present in the absence of 7Sgamma globulin, r'activity only with HGG maybe encountered. Thus rheumatoid factor moleculescould be produced with various sedimentationcoefficients depending on the degree of aggregationand with corresponding transition of reactivity.It is well known that high molecular weight com-plexes occur in rheumatoid sera, possibly withdifferent molar ratios of 7 and 19S gamma globulin(Franklin, Holman, Muller-Eberhard, and Kunkel,1957), and it is conceivable that a proportion of thesecomplexes may be formed intracellularly. Anattempt such as this to equate the various typesof cellular rheumatoid factor activity with the sizeof the molecular complexes would not necessarilybe valid for the standard serological procedureswhere a dynamic equilibrium might exist, allowingdissociation and re-association of complexes whichcould not occur at cellular level in tissue sections.Further, Epstein and Tan (1961) have observeddissociation of a 22S complex on anionic columnchromatography and, if it were necessary for thishypothesis to account for the sensitized sheep cellagglutinating activity of 19S rheumatoid factorisolated by such techniques (Lospalluto and Ziff,1959), one might postulate that some hetero-specificity had survived in partially dissociatedcomplexes not resolved from "stem" 19S rheumatoidfactor in the ultracentrifuge.The possibility that rheumatoid factor may be an

antibody-antibody complex has been considered byNajjar and Robinson (1959); Dubiski (1958) sug-gested that rheumatoid factor was an anti-antibodyand others indicated that it might represent anantibody to altered human gamma globulin (Glynn,Holborow, and Johnson, 1927). The presence ofrheumatoid factor in plasma cells and the intrinsiccells of germinal centres suggests though it doesnot confirm its antibody nature and it may bepossible to extend the interpretation of the resultsof this investigation by postulating that "stem"rheumatoid factor arises in one cell or germinalcentre in response to an antigenic stimulus from anabnormal 7S gamma globulin being produced atanother site which may be distant or in the samecell or germinal centre. If such abnormal gammaglobulins were, for instance, autoantibody directedagainst connective tissue antigens, then rheumatoidfactor might play a protective role by complexingwith such autoantibodies as they are produced.When pre-formed complexes are eventually releasedfrom the cell, the rheumatoid factor may prevent or

modify the reaction with the antigen against whichthe autoantibody is directed, perhaps by preventingfixation of complement. It is also implicit in thishypothesis that the reaction between rheumatoidfactor and autoantibody in the parent cell could notresult in damage, as such cells would not survive.Not all 7S gamma globulin present in circulatingrheumatoid factor complexes need be autoantibodyin view of the indication that isolated 19S rheuma-toid factor can complex with normal gamma globulin(Edelman, Kunkel, and Franklin, 1958). It islikely that 19S rheumatoid factor liberated fromType I cells would be capable of forming equi-librium complexes with normal 7S gamma globulinin the serum, and that partially saturated complexesof rheumatoid factor and autoantibody releasedfrom Type II cells may become saturated by aggrega-tion with normal 7S gamma globulin withoutdisplacement of the autoantibody. In addition, theratio of normal to abnormal 7S in the complexesmay vary according to activity of the disease andthe degree of antigenic stimulation.The results of further work around this theme

(McCormick, to be published) indicate that absorp-tion of rheumatoid factor from selected rheumatoidsera results in loss of antinuclear factor activity,which suggests that rheumatoid factor-autoantibodycomplexes are a possibility. Kaplan and Vaughan(1959) suggested that rheumatoid factor andreactant gamma globulin may occur bound in thetissues, possibly in the form of complexes, and, onthe basis of the assumptions outlined above, it ispossible that the extracellular Type IIl reactionsnoted in this study represent saturated rheumatoidfactor complexes bound specifically to connectivetissue antigens. As a corollary, it is likely in inves-tigations of this kind that fluorescent aggregatedhuman gamma globulin would not detect speci-fically bound rheumatoid factor.

Finally, the coexistence of 7S gamma globulinand macroglobulin in individual cells may indicatemerely a maturation process in which macroglobulinis being produced by the aggregation of 7S mole-cules. Deutsch and Morton (1957), from theirstudies on the dissociation of macroglobulins bymercapto-ethanol treatment, suggested that macro-globulins might be formed from normal 7S sub-units.Presumably, therefore, such aggregation could occurintracellularly but, on the whole, the hypothesis,however tenuous, which allows a protective bio-logical role for rheumatoid factor seems moreattractive.

SummaryFluorescent aggregated human gamma globulin

and normal or immune rabbit globulins have been

8

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


used to trace the distribution of rheumatoid factorin rheumatoid lymph nodes and synovium. Rheu-matoid factor was detected by each class of reagentin plasma cells in lymph nodes and synovium, andin the intrinsic cells of reactive follicles. Thelabelled rabbit globulin reagents also reacted withrheumatoid factor at various extracellular con-nective tissue sites. Mixed-staining experimentswith contrastingly labelled aggregated human gammaglobulin and the rabbit globulins showed that someplasma cells and intrinsic cells reacted with eitherthe aggregate or the rabbit globulins and that othersreacted with both. Various combinations of eachfluorescence reaction were found in individual cellsand germinal centres. Mixed-staining procedureswith contrastingly labelled anti-7S and a specificanti-19S antisera indicated that 7S gamma globulinwas associated with macroglobulin at similar sites,and it is suggested that the varying reactivity ofrheumatoid factor with the aggregate and rabbitglobulins depends upon the ratio of native 7S gammaglobulin to a "stem" 19S rheumatoid factor inrheumatoid factor complexes. In addition, it waspostulated that rheumatoid factor might arise asan antibody to autoantibody with which it com-plexes intracellularly, preventing or modifyingsubsequent reactions with connective tissue antigens.Finally, the possibility that macroglobulins may beformed from intracellular aggregation of 7S gammaglobulin was considered.

I wish to thank A.G.S. Hill and Dr. D. G. Scott formuch helpful discussion and criticism.

I am grateful to Dr. J. F. Soothill, University of Bir-mingham, for providing the anti-7S and anti-19S gammaglobulin antisera, and to Dr. W. d'A. Maycock, ListerInstitute of Preventive Medicine, for supplying thehuman gamma globulin.

Dr. C. L. Greenbury kindly supplied the Cohn fractionII and various rabbit antisera.The lissamine rhodamine B 200 was supplied by

Imperial Chemical Industries, Dyestuffs Division.I am indebted to the Nuffield Foundation and the

Empire Rheumatism Council whose grants to theRheumatism Research Centre enabled this work to bepursued.

REFERENCESChadwick, C. S., McEntegart, M. G., and Nairn, R. C.

(1958). Lancet, 1, 412.Christian, C. L. (1958). J. exp. Med., 108, 139.Coons, A. H., and Kaplan, M. H. (1950). Ibid., 91, 1.Deutsch, H. F., and Morton, J. I. (1957). Science,

125, 600.Dubiski, S. (1958). Fol. biol., 6, 47.Dutcher, T. F., and Fahey, J. L. (1959). J. nat. Cancer

Inst., 22, 887.

Edelman, G. M., Kunkel, H. G., and Franklin, E. C.(1958). J. exp. Med., 108, 105.

Epstein, W., Johnson, A., and Ragan, C. (1956). Proc.Soc. exp. Biol. (N.Y.), 91, 235.

Epstein, W. E., and Tan, M. (1961). Arthr. and Rheum.,4, 414 (Abstract).

Franklin, E. C., Holman, H. R., Muller-Eberhard, H. J.,and Kunkel, H. G. (1957). J. exp. Med., 105,425.

Glynn, L. E., Holborow, E. J., and Johnson, G. D. (1957).Proc. roy. Soc. Med., 50, 469.

Heller, G., Jacobson, A. S., Kolodny, M. H., andKammerer, W. H. (1954). J. Immunol., 72, 66.

Hess, E., and Ziff, M. (1961). Arthr. and Rheum., 4, 574.Kaplan, M. H., Suchy, M. L., and Meyeserian, M. (1960).

Ibid., 3, 272 (Abstract).and Vaughan, J. H. (1959). Ibid., 2, 356 (Abstract).

Kekwick, R. A., and Mackay, M. E. (1954). M.R.C.Spec. Rep. Ser. No. 286.

Lospalluto, J., and Ziff, M. (1959). J. exp. Med.,110, 169.

Mellors, R. C., Heimer, R., Corcos, J., and Korngold, L.(1959). Ibid., 110, 875.Nowoslawski, A., and Korngold, L. (1961a).Amer. J. Path., 39, 533.

- ~,, , and Sengson, B. L. (1961b). J. exp.Med., 113, 475.

Najjar, V. A., and Robinson, J. P. (1959). In "Immunityand Virus Infection", ed. V. A. Najjar, p. 71.Wiley, New York.

Riggs, J. L., Seiwald, R. J., Burckhalter, J. H., Downs,C. M., and Metcalf, T. G. (1958). Amer. J.Path., 34, 1081.

Rose, H. M., Ragan, C., Pearce, E., and Lipman, M. 0.(1948). Proc. Soc. exp. Biol. (N. Y.), 68, 1.

Scott, D. G. (1960). Immunology, 3, 226.Singer, J. M., and Plotz, C. M. (1956). Amer. J. Med.,

21, 888.Waaler, E. (1940). Acta path. microbial. scand., 17, 172.

Discussion.-Dr. J. T. Scorr (Hammersmith): MayI ask Dr. McCormick about the correlation of thesefindings with the presence of rheumatoid factor in theserum? Has he found rheumatoid factor by thesetechniques in patients with negative Waaler-Rose tests?DR. MCCORMICK: Of the patients I have studied, all

have had a positive Waaler-Rose test except one, whohad definite rheumatoid arthritis, but negative sheep cellagglutination and latex-fixation tests. A few plasmacells containing rheumatoid factor were found in syno-vium from this subject.

Etude immunofluorescente du facteur rhumatismalREuSMi

La globulin gamma humaine agglutinee et renduefluorescent, et des globulines de lapin, normales etimmunes, furent utilisees pour suivre la distributiondu facteur rhumatismal dans les ganglions lymphatiqueset la synovie des rhumatisants. Le facteur rhumatismalfut ddc ld par chaque type de reactif dans des plasmo-cytes, ganglions lymphatiques et dans la synovie, ainsique dans les cellules intrinseques des follicules actives.La globulin coloree de lapin reagit aussi avec le facteurrhumatismal a de differents endroits du tissu conjonctif

9

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

ANNALS OF THE RHEUMATIC DISEASESextracellulaire. Des experiences de coloration mixte,avec la globulin gamma agglutinee humaine color6edifferemment des globulines de lapin, demontrerent quecertains plasmocytes et cellules intrinseques reagirentsoit avec la globulin humaine, soit avec les globulinesde lapin, tandis que d'autres reagirent avec les deux.On trouva de differentes combinaisons de chaquereaction de fluorescence dans des cellules individuelles etdes centres germinaux. En colorant differemmentl'antiserum anti-S7 et l'antiserum specifique anti-19S,on vit que la globulin gamma 7S se trouvait associee ala macroglobuline a des endroits similaires, et on penseque la reactivite variable du facteur rhumatismal avecla globuline humaine et de lapin depend du rapport entrela globuline gamma 7S native et un facteur rhumatismal"tige" ("stem") 19S dans les complexes du facteurrhumatismal. De plus, on posture que le facteurrhumatismal pourrait se presenter comme un anticorpscontre un autoanticorps, avec sequel il formerait descomplexes intracellulaires, empechant ou modifiantles reactions subsequentes avec les antigenes du tissuconjonctif. Finalement, on a considered la possibility queles macroglobulines seraient formees par l'aggregationintracellulaire de la globulin gamma 7S.

Estudio inmmunofluorescente del factor reumatoide

SUMA10OLa globulin gamma humana, aglutinada y fluores-

cente, asi como globulinas de conejo, normales e

inmunes, fueron empleadas para investigar la distri-bucion del factor reumatoide en los ganglios linfaticosy la sinovia de enfermos reumatoides. El factor reuma-toide fue evidenciado por cada clase de reactivo enplasmocitos, ganglios linfaticos y en la sinovia, asicomo en las celulas intrinsecas de los foliculos activos.La globulin fluorescent reaccion6 tambien con elfactor reumatoide en varios sitios del tejido conectivoextracelular. Experiencias de coloracion mixta, tiniendodiferentemente la globulin gamma aglutinada humanay las globulinas de conejo, demostraron que ciertosplasmocitos y celulas intrinsecas reaccionan sea conla globulin humana sea con las globulinas de conejo,mientras que otros reaccionan con ambas globulinas.Varias combinaciones de cada reaccion de fluorescenciafueron encontradas en celulas individuales y en centrosgerminales. Con coloraciones de contrast del anti-suero anti-7S y del antisuero especifico anti-19S se pudoobservar una asociaci6n de la globulin gama 7S conla macroglobulina en sitios similares y se sugiere que lareactividad variable del factor reumatoide con lasglobulinas humana y de conejo depende de la relacionentre la globulin gamma 7S nativa y un factor reuma-toide "tallo" ("stem") 19S en los complejos del factorreumatoide. Ademas, se postula que el factor reuma-toide pudiera presentarse como un anticuerpo contra unautoanticuerpo y formar con el complejos intracelularese impedir o modificar reacciones subsiguientes conantigenos del tejido conectivo. Por fin, se considerposible que las macroglobulinas sean formadas por unaagregacion intracelular de la globulin gamma 7S.

10

copyright. on D



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

immunofluorescence study rheumatoid

Documents