Journal of Autoimmunity (1999) 13, 83–87Article No. jaut.1999.0290, available online at http://www.idealibrary.com on

A New Antigenic Epitope Localized Within Human ê LightChains Specific for Rheumatoid Arthritis and SystemicLupus Erythematosus

P. von Landenberg 1, C. von Landenberg 2, M. Grundl 1, G. Schmitz 2, J Scho lmerich 1,and I. Melchers 3

1Klinik und Poliklinik fur Innere Medizin,Universitat Regensburg, Regensburg,Germany2Institut fur Klinische Chemie undLaboratoriumsmedizin, UniversitatRegensburg, Regensburg, Germany3Klinische Forschergruppe furRheumatologie,Albert-Ludwigs-Universitat Freiburg,Freiburg, Germany

Human B cell hybridomas were established to define new autoantigens ofimportance for autoimmune diseases such as rheumatoid arthritis (RA). OnelgG1, ë monoclonal antibody (FKN-E12), was derived from synovial Blymphocytes of a patient with sero-negative RA. The purified lg was used toselect specifically binding peptides from a random peptide phage displaylibrary. Only one epitope with the heptamer sequence HLTFGPG was detectedand named RASFp1. Very similar and partly identical sequences are found inthe variable region of lg ê light chains in position 96–101, at the junction offramework 2 and the J-region. The antibody FKN-E12 was shown to detect theepitope RASFp1 also on human lgG ê chains, but only in a specific conforma-tion. The aim of the present study was to analyse human sera from patientswith autoimmune diseases, non-autoimmune inflammatory diseases andhealthy blood donors for the presence of lgG binding to RASFp1. For thispurpose a 15-mer-peptide was synthesized containing RASFp1 withinVk-derived flanking regions, and an ELISA assay established. Sera of 142individuals were studied. Only <5% of the control sera including sera frompatients with non-autoimmune inflammations were positive. In contrast,45% of sera from patients with RA or SLE contained RASFp1-bindingantibodies. Within the 40 RA sera analysed so far, rheumatoid factors andRASFp1-binding antibodies have shown no correlation with each other.

© 1999 Academic Press

Key words: autoantibodies,autoimmunity, rheumatoidfactors, rheumatoid arthritis,synovial fluid

Correspondence to: Philipp von Landenberg, Klinik und Poliklinik furInnere Medizin I, Universitat Regensburg, 93042 Regensburg,Germany. Fax: 0049 (0) 941 944 7073.

Introduction

Rheumatoid arthritis (RA) is a chronic systemicinflammatory disease of the connective tissue of yetunknown etiology preferentially involving joints.There are several lines of evidence indicating anautoimmune process. Infiltration of macrophages aswell as lymphocytes into the inflammatory lesion canbe detected. Most of these infiltrating lymphocytes areT cells, but B cells are also present [1–7]. Moreover,there is evidence that B cells entering the synovialtissue via capillaries find a microenvironment wherethey may be activated and can differentiate intoplasma cells [1, 3]. In the synovial tissue germinalcentre-like structures were described, containing Blymphocytes as well as follicular dendritic cellsarranged in a typical pattern [8–10]. Dual immuno-fluorescent staining of cell surface antigens revealed

830896–8411/99/050083+05 $30.00/0

an increased percentage of B cells and plasma cells aswell as a predominant CD4+ T cell population insynovial tissue derived from RA patients [11]. Thus,all evidence is compatible with the idea that synovialB cells are activated by synovial antigens, differentiatewithin the synovial tissue and produce antibodiesspecific for local autoantigens [3]. The same antigensmay also be responsible for the T cell activation in theB cell neighbourhood. The characterization of locallyproduced antibodies thus may provide valuable toolsto identify locally relevant autoantigens in general.We, and others, therefore started to establish hybrid-omas and monoclonal antibodies derived fromsynovial B cells of RA patients. The epitope specificityof one IgG1, lambda antibody of this series wasanalysed using the peptide phage display technique[12]. The peptide detected (RASFp1) had the sequenceHLTFGPG, which corresponds to a sequence in thevariable region of certain human lg ê light chains(rearrangement of any Vk-gene with Jk3, amino acids96–101). Using complete antibodies we could verifybinding of FKN-E12 to ê light chains in IgGs ofvarious species, if these were themselves bound to a

© 1999 Academic Press

84 P. von Landenberg et al .

plastic surface [12]. In the present study we used asynthetic peptide NLPLTLTFGPGTKVD in an ELISAto assess whether antibodies with a similar specificityas FKN-E12 occur in human sera.

Materials and Methods

Monoclonal antibodies and cell culture

The monoclonal antibody FKN-E12 was generated asdescribed [12]. Briefly, synovial fluid B lymphocytesfrom a 62-year-old male patient suffering from RAwere isolated by Ficoll density centrifugation. Thelymphocytes were electrofused in a helical fusionchamber in a hypoosmolar (75 mosm) fusion mediumwith the heteromyeloma cell line CB-F7 [13]. Cellswere adjusted to a total concentration of 3×106 cells in200 ìl and fusion was performed in a weakly inhom-ogenous alternating electrical field. After recloningtwice we obtained the monoclonal hybridoma cell lineFKN-E12. Cells were cultured in RPMI 1640 completemedium (supplemented with 10% FCS, 2 mMglutamine, 100 units/ml penicillin and 100 mg/mlstreptomycin).

Peptide-ELISA

The peptide NLPLTLTFGPGTKVD was synthesizedby Prof S. Modrow, Institute for Microbiology,University Hospital Regensburg, Germany anddiluted in H2O to a concentration of 1.2 mg/ml. Coat-ing of an ELISA-plate was performed in PBS in aconcentration of 100 ng/ml for 12 h at 4°C. Blockingwas achieved with PBS Tween 1% for 1 h at 37°C. Seratested were diluted 1:100 in PBS Tween 0.1%. Onehundred microlitres per well was used in doubletesting. Bound human serum IgG was detected with aperoxidase-coupled goat-anti-human IgG (Dianova)1:10.000 in PBS Tween 0.1%, 100 ìl/well for 1 h at37°C. As enzyme substrate ABTS, 50 ìl/well wasused.

RF-test

The presence of RF in the sera of RA and SLE patientswas detected with a standard RF-Latex-assay inthe Central Laboratory Unit at the University Hospitalof Regensburg, Regensburg, Germany.

Antibody purification

The supernatants of the hybridoma cell line werepurified by using Protein G Sepharose (Sigma). Thebound antibodies were eluated with glycin buffer 0.1M, pH 2.2, neutralized with 1 M Tris pH 9.0 anddialysed against PBS. In the ELISA assays the purifiedantibody was used as control at a concentration of1 mg/ml.

Patients

Sera from 40 patients with RA and 40 patients withSLE (diagnosed according to the ACR criteria) wereused in the ELISA. As controls, 36 sera from healthyblood donors were used. To determine the relevanceof these autoantibodies in autoimmune diseases, serafrom patients with elevated inflammation parametersbut no autoimmune disease were used as control.

Results

The monoclonal antibody FKN-E12

The production of the hybridoma clone FKN-E12 wasdescribed in detail elsewhere [12]. Briefly, B cells fromthe synovial fluid of a male patient with sero-negativeRA were fused with CB-F7 cells [13], cloned andselected for the stable production of lgG. The purifiedlgG was used to select peptides from a heptamerrandom peptide library displayed on a phage surfaceprotein [14]. Only one sequence was found(HLTFGPG) and the epitope was named RASFp1. Thehexamer LTFGPG was detected in three human lg êlight chains, the pentamer TFGPG in all human ê lightchains created by rearrangement of a Vk-gene withJk3, and similar sequences in all ê light chains ofseveral species [12]. FKN-E12 bound to all lgGs testedin an ELISA, as long as they contained ê light chainsand were bound to a plastic surface, but did not bindto ê light chains in solution or to B cell receptors.Later, the V regions of both heavy and light chainswere sequenced, and revealed a normal ë light chain,but an unsusual VH region, containing an insertion of13 amino acids within the FR2/CDR2 region (Schaferet al., manuscript in preparation).

RASFp1 is an epitope recognized by antibodiesin many sera of RA and SLE patients

Our analysis is based on the assumption that epitopesdetected by individual monoclonal antibodies derivedfrom synovial B cells are relevant for the B cellrepertoire of the individual patient. They are evenmore interesting, if also in other patients, antibodies—and B cells—with the same specificity can be detected.We therefore established an ELISA to screen sera ofhealthy donors and patients for IgGs binding toRASFp1. The peptide synthesized for this purposecontained RASFp1 within Vk-derived flankingregions (NLPLTLTFGPGTKVD) and allowed simplehandling. FKN-E12 itself bound to this peptide (datanot shown). We then screened 142 sera, from eitherhealthy donors (36) patients with RA or SLE (40 each)or non-autoimmune inflammation (26), for lgG bind-ing to the peptide (Figure 1). The sera derived fromHD served to define the normal distribution (mean±SD=0.147±0.073) and a threshold (t=mean+3SD=0.367), above which sera were regarded as beingpositive. Within the healthy population, one serum

Autoantibodies to light chains in RA 85

was above this threshold, whereas 45% of bothpatients with RA or SLE, exhibited values above thethreshold (18 of 40 in each group). Most interestingly,only two out of 26 non-autoimmune inflammatorysera were above the threshold, suggesting that theepitope is not generally recognized by antibodiesarising during inflammation. The means and standarddeviations calculated for the four groups of serumdonors showed a significant difference between RAand SLE patients on the one hand and betweenhealthy donors and donors with inflammations on theother (Figure 2). Selected positive sera were shown tobind to the peptide in a dose-dependent manner (notshown).

Antibodies binding to RASFp1 and rheumatoidfactors

FKN-E12 originated from a sero-negative RA patient[12]. Nevertheless, RASFp1 is an epitope, situated in ahuman lg molecule, close to but outside of the antigenbinding site. The antibodies best characterized in RApatients are the rheumatoid factors, also binding tolgG, but classically to the heavy chain Fc-region [15].The lgG molecule carries several epitopes in theFc-region of the heavy chain, and also in the lightchain. One might therefore expect a correlationbetween the titers of antibodies binding to theserelated antigenic determinants.

Titers of RF were measured in the same samplesof RA derived sera which were also used for thepeptide binding assay (Table 1). These titers and thecorresponding ELISA data are plotted against eachother in Figure 3. There was no correlation detectedbetween the two (Pearson rank correlation coefficientr=0.228).

1.0

0

OD

(40

5 n

m) 0.8

0.6

0.4

0.2

Inflammation sera (n = 26)1.0

0

0.8

0.6

0.4

0.2

Healthy controls (n = 36)

1.0

0

OD

(40

5 n

m) 0.8

0.6

0.4

0.2

RA patients (n = 40)1.0

0

0.8

0.6

0.4

0.2

SLE patients (n = 40)

Figure 1. Results of the peptide ELISA. The synthetic peptide NLPLTLTFGPGTKVD was used to coat an ELISA plate(100 ng/ml). Patient and control sera were tested at a 1:100 dilution. Bound IgG was detected with a peroxidasecoupled goat-anti-human-IgG-antibody. Results are given in OD values. Each value represents the mean of duplicatedeterminations.

0.6

0.0Normal

OD

(40

5 n

m)

0.5

0.4

0.3

0.2

0.1

InflammationSLERA

Figure 2. Results of the anti-peptide ELISA. There is asignificant difference between the mean OD in the anti-peptide ELISA between RA and SLE sera compared tohealthy controls and non-autoimmune sera. The P value is<0.0001, determined by the t-test.

Discussion

The presence of RF in the serum of RA patients is usedas as diagnostic criterium for the definition of thedisease and of subgroups of patients (sero-positiveand sero-negative). For a long time it has been knownthat RF are produced by synovial B cells [1, 3, 16]. Thisresults in the formation of immune complexes withinthe synovium which are thought to contribute to thepathogenesis of RA [15, 17, 18]. RFs may be of anyisotype (lgM, lgG, lgA), and their specificity may beone of several. Most antigenic determinants werelocalized to the Fc portion of the lgG molecules in both

86 P. von Landenberg et al .

the CH2 and CH3 domains. These determinants in-clude cross-reactive epitopes shared by human andanimal lgG, species-specific epitopes, subclass-specificepitopes, allotypes and neoantigens expressed ondenatured, aglycosylated or aggregated lgG [15].Occasionally, antibodies were also described thatreacted with epitopes on the Fab region of human IgG[19]. We here describe a new epitope (RASFp1) whichis localized in human lg ê chains. lgG binding athigher titers to a synthetic peptide of 15 amino acidscontaining the RASFp1 epitope was detected in 45% ofthe tested sera derived from patients with RA or SLE,compared to <5% of sera derived from healthy blooddonors and patients with other inflammatory diseases.

RFs are usually measured by the latex agglutinationtest. Human lgG is absorbed to latex beads, which can

be agglutinated mainly by lgM binding to the lgG.Thus, the RF titer given in this and other publicationsmainly corresponds to the titer of lgM RF. In theELISA we detected lgG binding to the epitopeRASFp1; other isotypes were not tested. The results ofboth assays showed no correlation, neither positivenor negative. Future assays will also show whetherlgM anti-RASFp1 is enhanced in RA sera, and whetherthis specificity is included within the RF populationdetected using the latex test (which we would expect).The structure of the original antibody FKN-E12 wasfound to be rather unusual (Schafer et al., manuscriptin preparation). Its binding specificity may thereforediffer from the binding specificity of other antibodiesrecognizing the same epitope. We have not yet provenwhether RA-derived serum lgG specific for RASFp1does bind to lg ê light chains in solution or not. If itdoes, it is able to form immune complexes with justone other lg ê molecule similar to classical RF. If theydo not, but behave as FKN-E12, they may formimmune complexes with lg ê molecules which havechanged their conformation, e.g. by binding them-selves to a surface or, even more interesting, bybinding to antigen. In this latter case immune com-plexes will form, containing three different molecularspecies: (1) the antigen, (2) the antigen-specific lgG, êmolecule and (3) the RASFp1-specific antibody. Inaddition, anti-RASFp1 antibodies may bind to a sub-population of ê chains, carrying a special sequence inposition 96–101, or even activate a B cell populationusing such receptors.

Immune complexes can be formed in various ways,including idiotypic interactions [20]. Many RF onlybind to lgG which is changed in one of severaldifferent ways, e.g. by denaturation or deglycosyla-tion. In the case of RASFp1-binding antibodies, subtlechanges in conformation may convert a normal lgG êmolecule into a recognized autoantigen. We suggestthat such a conversion might occur by binding of thelgG to its antigen.

Table 1. RF serum titers determined with a routine RF Latetest (cut off 18 ì/ml)

Serum No. RF units Serum No. RF units

1 1153.3 21 108.72 1710.5 22 74.23 715 23 1354 384.4 24 82.75 18 25 6606 16.8 26 18807 18 27 15.28 16.8 28 21.19 146.9 29 181.2

10 63.5 30 259011 20 31 9.412 9.4 32 15.213 16.5 33 33.414 38.1 34 47.415 89.6 35 016 23.6 36 17.517 17.1 37 247.718 17.5 38 15.219 166.8 39 211.320 231.4 40 23.4

0.8

3,000

00.0

Peptide OD

RF

un

its

2,500

2,000

1,500

1000

500

0.2 0.4 0.6

Figure 3. Correlation curves. There is no correlation betweenthe RF titers in the RA sera and the OD values in theanti-peptide ELISA.

References1. Natvig J.B., Randen I., Thompson K., Forre O., Munthe

E. 1989. The B cell system in the rheumatoidinflammation. New insights into the pathogenesis ofrheumatoid arthritis using synovial B cell hybridomaclones. Springer Semin. Immunopathol. 11: 301

2. Rudolphi U., Hohlbaum A., Lang B., Peter H.H.,Melchers I. 1993. The B cell repertoire of patients withrheumatoid arthritis. Frequencies and specificities ofperipheral blood B cells reacting with human IgG,human collagens, a mycobacterial heat shock proteinand other antigens. Clin. Exp. Immunol. 92: 404

3. Rudolphi U., Rzepka R., Batsford S., Kaufmann S.H.,von der Mark K., Peter H.H., Melchers I. 1997. The Bcell repertoire of patients with rheumatoid arthritis. II.Increased frequencies of IgG+ and IgA+ B cells specificfor mycobacterial heat-shock protein 60 or human typeII collagen in synovial fluid and tissue. ArthritisRheum. 40: 1409

4. Panayi G.S. 1993. The pathogenesis of rheumatoidarthritis: from molecules to the whole patient. Br. J.Rheumatol. 32: 533

Autoantibodies to light chains in RA 87

5. Panayi G.S. 1993. The immunopathogenesis ofrheumatoid arthritis. Br. J. Rheumatol. 32(Suppl): 4

6. Panayi G.S. 1994. The involvement of autoantigens inrheumatoid arthritis. Br. J. Rheumatol. 33: 996

7. Melchers I., Peter H.H., Eibel H. 1995. The T and B cellrepertoire of patients with rheumatoid arthritis. Scand.J. Rheumatol. 101: (Suppl) 153

8. Krenn V., Molitoris R., Berek C., Sack U., Konig A.,Muller-Deubert S., Kempf V., Mosgoeller W., KerkauT., Vollmers H.P., Muller-Hermelink H.K. 1998. A novelmonospecific IgG2/lambda-autoantibody withspecificity for a mitochondrial antigen: evidence for anantigen-driven pathogenetic B-cell response inrheumatoid synovial tissue, induced by tissuealteration. Lab. Invest. 78: 485

9. Krenn V., Schedel J., Doring A., Huppertz H.I., GohlkeF., Tony H.P., Vollmers H.P., Muller-Hermelink H.K.1997. Endothelial cells are the major source ofsICAM-1 in rheumatoid synovial tissue. Rheumatol. Int.17: 17

10. Krenn V., Schalhorn N., Greiner A., Molitoris R., KonigA., Gohlke F., Muller-Hermelink H.K. 1996.Immunohistochemical analysis of proliferating andantigen-presenting cells in rheumatoid synovial tissue.Rheumatol. In. 15: 239

11. Nakao H., Eguchi K., Kawakami A., Migita K., OtsuboT., Ueki Y., Shimomura C., Tezuka H., Matsunaga M.,Maeda K. 1990. Phenotypic characterization oflymphocytes infiltrating synovial tissue from patientswith rheumatoid arthritis: analysis of lymphocytesisolated from minced synovial tissue by dualimmunofluorescent staining. J. Rheumatol. 17: 142

12. von Landenberg P., Rzepka R., Melchers I. 1999.Monoclonal autoantibodies from patients withautoimmune diseases: synovial fluid B lymphocytes ofa patient with rheumatoid arthritis produced an IgG

lambda antibody recognizing J-sequences of Ig kappachains in a conformation dependant way.Immunobiology (in press)

13. Grunow R., Jahn S., Porstmann T., Kiessig S.S.,Steinkellner H., Steindl F., Mattanovich D., Gurtler L.,Deinhardt F., Katinger H. 1988. The high efficiency,human B cell immortalizing heteromyeloma CB-F7.Production of human monoclonal antibodies to humanimmunodeficiency virus. J. Immunol. Methods 106: 257

14. Devlin J.J., Panganiban L.C., Devlin P.E. 1990. Randompeptide libraries: a source of specific protein bindingmolecules. Science 249: 404

15. Carson D.A. 1989. Rheumatoid factor. In Kelley J.,Hains G., Ruddy A., Sledge B., eds. Textbook ofRheumatology, 3rd edn. WB Saunders Co.Philadelphia, London, Toronto

16. Munthe E., Natvig J.B. 1972. Immunoglobulin classes,subclasses and complexes of lgG rheumatoid factor inrheumatoid plasma cells. Clin. Exp. Immunol. 12: 55–70

17. Mageed R.A., Kirwan J.R., Holborow E.J. 1991.Localization of circulating immune complexes frompatients with rheumatoid arthritis in murine spleengerminal centres. Scand. J. Immunol. 34: 323

18. Jarvis J.N., Diebold M.M., Chadwell M.K., Iobidze M.,Moore H.T. 1995. Composition and biologicalbehaviour of immune complexes isolated fromsynovial fluid of patients with juvenile rheumatoidarthritis (JRA). Clin. Exp. Immunol. 100: 514

19. Nasu H., Chia D.S., Knutson D.W., Barrett E.V. 1980.Naturally occurring human antibodies to the F(ab)2portion of lgG. Clin. Exp. Immunol. 42: 378

20. Rajnavolgyi E., Reth M., Uher F., Miklos K., Gergely J.,Rajewsky K. 1984. Fc-dependent effector functions ofidiotype-anti-idiotype immune complexes. Mol.Immunol. 21: 127