Induction of anti-RA33 hnRNP and transient spread to U1-A snRNP complex of spliceosome by idiotypic manipulation with anti-RA33 antibody preparation in mice

G. Steiner1,2, O. Shovman3, K. Skriner1,2, B. Gilburd4, P. Langevitz5, M. Miholits1,2, R. Hoet6, Y. Levy 3, G. Za n d m a n - G o dd a r d3, E. Hoefle r 7,J.S . Smolen1, 7 , Y. Shoenfel d 3, 4 , 5

1Division of Rheumatology, Department of Internal Medicine III and 2Institute of Biochemistry, University of Vienna, Austria; 3Department of Medicine B, 4Center for Autoimmune Diseases, and 5Division of Rheumatology, Sheba Medical Center, Tel Hashomer and Sackler Faculty of Medicine Tel-Aviv University, Tel-Aviv; 6Department of Biochemistry, University of Nijmegen, The Netherlands; 7Second Department of Medicine, Lainz Hospital, Vienna, Austria.

Abstract Objective Anti-RA33 antibodies occur in patients with (RA), systemic erythematosus (SLE), and mixed connective tissue disease (MCTD) and target the A2/B1 protein of the heterogeneous nuclear ribonucleoprotein (hnRNP) complex 4 which forms part of the spliceosome. The aim of the present study was to evaluate the immune response and pathological features induced in mice immunized with anti-RA33 antibodies or patient-derived recombinant single-chain variable fragments (scFv) of anti-RA33 antibodies. Methods In the first set of the experiment, two strains of mice (C57BL/6J and BALB/c) were immunized with IgG preparations obtained from two patients with RA and one normal donor. One of the patients had high titer anti-RA33 antibodies; the other one showed weak borderline reactivity. In the second set of the experiment three groups of C57BL/6J mice were immunized, respectively, with affinity-purified (AP) anti-RA33 antibodies, scFv of anti-RA33 antibodies and normal human IgG. The immunological response induced in immunized mice was studied by immunoblotting and line immunoassay (LIA). The presence of arthritis, serositis or myositis was assessed six-months following initial immunization. Results While anti-RA33 antibodies developed in only two of the mice immunized with different human IgG fractions, anti- RA33 antibodies were clearly detected in 7 sera of 13 mice immunized with AP anti-RA33 antibodies three months after the boost immunization and, moreover, also in 2 sera of 13 mice immunized with scFv of anti-RA33 antibodies. In contrast, mice immunized with normal human IgG did not develop anti-RA33 antibodies. Interestingly, transient production against another nuclear autoantigen, U1 snRNP, was observed in 3 C57BL/6J mice immu- nized with scFv and in 1 mouse immunized with AP autoantibodies. However, these immunological responses were not associated with pathological findings. Conclusions Active immunization of naive mice with AP anti-RA33 antibodies and scFv of anti-RA33 antibodies resulted on the one hand in the production of murine anti-RA33 antibodies and led, on the other hand, to transient “autoantibody spread” to snRNP component of the spliceosome and other nuclear autoantigens. This “autoantibody spread” probably reflected disregulation of the idiotypic anti-idiotypic cascade. Key words Affinity purified anti-RA33 antibodies, recombinant single-chain variable fragments of anti-RA33 antibodies, rheumatoid arthritis, systemic lupus erythematosus, mixed connective tissue disease, hnRNP complex.

Clinical and Experimental Rheumatology 2002; 20: 517-524. Anti-RA33 autoantibodies in mice / G. Steiner et al.

#G. Steiner, MD; #O. Shovman, MD; Introduction N-terminus, and hnRNP B2 which is K. Skriner, MD; B. Gilburd, MD; Most connective tissue diseases are assumed to be another altern at ive ly P. Langevitz, MD; M. Miholits, MD; characterized by the presence of auto- spliced variant of hnRNP-A2/B1 (13, R. Hoet, MD; Y. Levy, MD; E. Hoefler, antibodies to nucleic acid binding pro- 14). Accordingly, the RA33 autoanti- MD; J.S. Smolen, MD; Y. Shoenfeld,MD. # These two authors contributed equally teins. The targets of these autoantibod- gen is actually composed of thre e to the work. ies are, in general, protein components hnRNP pro t e i n s : A 2 , B l , B 2 , wh i ch This study was supported by a grant from of multimolecular ri b o nu cl e o p ro t e i n m ay be designated the “RA33 com- the Austrian National Bank and by the ICP (RNP) structures (1,2). These proteins p l ex”. Subsequently, anti-RA33 anti- Programme of the Austrian Federal Min- f re q u e n t ly play an important role in bodies were also detected in sera from istry for Education, Science and Culture. various vital cellular functions such as patients with SLE and MCTD (14, 15). Please address correspondence to: D NA rep l i c ation and tra n s c ri p t i o n , Similar observations were made with Dr. Yehuda Shoenfeld, Department of RNA processing, or messenger RNA the closely related hnRNP Al protein Medicine ‘B’, Sheba Medical Center, (mRNA) translation (3). For example, (15, 16). Sackler Faculty of Medicine, Tel-Aviv p roteins of the small nu clear RNP The most interesting fact is that this University, Tel-Hashomer 52621, Israel. (snRNP) complexe s , wh i ch fo rm the s u b fa m i ly of hnRNP proteins rep re- E-mail: [email protected] major constituent of the spliceosome, sents the only known spliceosome as- Received on October 8, 2001; accepted are characteristically targeted by auto- s o c i ated autoantigen wh i ch is re c og- in revised form on April 17, 2002. antibodies from patients with mixe d nized by patients with RA, emphasiz- © Copyright CLINICAL AND connective tissue disease (MCTD) and ing the immu n o l ogical re l at i o n s h i p EXPERIMENTAL RHEUMATOLOGY 2002. systemic lupus ery t h e m atosus (SLE): between RA, SLE and MCTD (14, 16). thus, Ul snRNP specific antigens are H oweve r, in SLE and MCTD anti- re c og n i zed by virt u a l ly all pat i e n t s RA33 antibodies usually occur togeth- with MCTD and by 20-50% of patients er with antibodies to other components with SLE, whereas the Sm antigen (i. e. of the spliceosome - Ul snRNP or Sm the common snRNP proteins) is a very proteins, respectively (14, 16, 17). A si- s p e c i fic autoantigen in patients with milar association has been recently ob- SLE (4, 5). Other major components of served in lupus prone MRL/lpr mice, the spliceosome are heteroge n e o u s wh e reas in NZB/NZW mice neither nuclear RNP (hnRNP) particles which anti-RA33 nor anti-snRNP antibodies are much less well characterized than could be detected. (18). Remarkably, in snRNP (3). These hnRNP are compos- MRL/lpr mice antibodies to short pep- ed of newly transcribed pre-mRNA and tide epitopes of hnRNP-A2/Bl and Sm- ~30 different proteins (6). Six of these - D protein developed very early and pre- Al, A2, Bl, B2, Cl, and C2 - have been ceded antibodies to the whole proteins termed hnRNP “core” proteins. They and, even more importantly to dsDNA are structurally related and have molec- by several weeks. Therefore, in SLE ular weights (MW) of 34-43 kd (7). (and MCTD) anti-hnRNP and anti- In contrast to snRNPs, proteins of the snRNP (i.e. anti-spliceosomal) reactiv- hnRNP complex have long been ne- ities appear to be early and closely link- glected as potential autoantigens in ed immune responses (14, 17) wh i ch c o n n e c t ive tissue diseases (8, 9). In might represent some of the first auto- 1989 a new IgG autoantibody was de- immune events, at least in the particular tected in sera from patients with rheu- disease model of SLE. On the other matoid arthritis (RA), which was di- hand, the more restrictive response to rected against a nuclear protein with the spliceosome in RA patients may MW of 33kD, therefore this antibody reflect a different type of autoimmune was designated as anti-RA33 (10). Sev- response indicative of a probable insult eral experimental studies demonstrated of anti-RA33 in the pat h ogenesis of the RA33 antigen to be indistinguish- RA. Our objective in the present study able from the hnRNP-A2 pro t e i n , was to eva l u ate the effi c a cy of anti- thereby identifying a hnRNP protein as RA33 antibodies to induce immune re- an important autoantigen in rheumatic sponses and, p o s s i bly, p at h o l ogi c a l diseases (11, 12). A dditional studies symptoms in naive mice. showed that anti-RA33 antibodies rec- ognized also hnRNP Bl, an alternative- Materials and methods ly spliced va riant diffe ring from A 2 Materials only by a 12 amino acid insertion at the Sera were obtained from two patients

518 Anti-RA33 autoantibodies in mice / G. Steiner et al. with RA; one of them had high titer i m mu n i zed intra d e rm a l ly in the hind data obtained by immunoblotting the anti-RA33 antibodies, the other one foot pads with different IgG prepara- recently developed line immunoassay showed borderline reactivity, and from tions 10 µg/mouse) diluted in Freund’s ( I n n oge n e t i c s , G h e n t , B e l gium) wa s one additional healthy donor. None of complete adjuvant. Group 1 was immu- used which employs a panel of immo- the sera exhibited any reactivity to the nized with normal human IgG, group 2 bilized (mostly recombinant) antigens U1-A snRNP antigen. with anti-RA33 positive IgG and group including Sm, Ul snRNP antigens, Ro Female C57BL/6J (art h ritis pro n e 3 with IgG preparations obtained from (SS-A), La(SS-B), Topo-1, Jo-1, rRNP, mice) and BALB/c were obtained from an RA patient with bord e rline anti- CENP-B, histones and DNA (20). the Te l - Av iv Rep o s i t o ry. The mice RA33 antibodies. Th ree weeks lat e r were 14 weeks old at the initiation of boost injections wer e administered with Determination of anti-DNA and the experiments. the same amounts of antibody diluted anti-cardiolipin in PBS. The anti-dsDNA and anti-card i o l i p i n Immunoglobulin preparations and Three groups of C57BL/6J mice were antibody activity in murine sera were affinity purification of antibodies immunized intradermally into the hind evaluated by ELISA as previously de- Human IgG fractions we re puri fi e d footpads with AP anti-A2/RA33 hu- scribed by us (21). At the termination f rom the sera of RA patients and man antibodies in CFA (13 mice), re- of the experiment (6 months following healthy donor by affinity chromatogra- combinant scFv fragment of anti-A2/ initial immunization) the mice were sa- phy on protein A Sepharose Pharmacia, RA33 antibodies (15 mice) and with c ri ficed and eva l u ated histologi c a l ly Uppsala, Sweden) according to the ma- normal human IgG (10 mice), 10 µg for the presence of myositis, arthritis nufacture’s instructions. Concentration per mouse. A boost injection of the and serositis in the Pat h o l ogical De- of purified IgG was 2.9-4 mg/ml. Affin- same products was given three weeks partment of Tel-Hashomer Hospital. ity purification of anti-RA33 antibod- later.The mice were bled monthly after ies was performed by blot elution using the boost. Results p u ri fied recombinant hnRNP-A2 fu- Specificity of anti-RA33 enriched sion protein as described (17). Determination of anti-RA33 antibody human IgG preparations and AP reactivity of human IgG fractions and anti-RA33 antibodies Recombinant single chain variable mouse sera As outlined in Figure 1a only the IgG fragments (scFv) of anti-ARA33 The anti-RA33 antibody reactivity of p rep a ration obtained from an anti- antibodies enriched IgG fractions and mouse sera RA33 positive patient with RA strong- P rep a ration of the V gene libra ri e s we re eva l u ated by immu n o bl o t t i n g ly reacted with the 68kD band corre- f rom pat i e n t s ’ ly m p h o cytes or bone using a HeLa nuclear extract as well as sponding to human re c o m b i n a n t m a rrow has been described (19). p u ri fied human recombinant hnRNP- hnRNP-A2/RA33 fusion protein (lane Libraries obtained from anti-RA33 po- A2 (RA33) antigen expressed as a 68 B). Borderline reaction can be seen in sitive patients with SLE or RA were kD fusion protein as prev i o u s ly de- lane C in which IgG from the weakly screened for reactivity with recombi- scribed (14, 17). All sera were tested re a c t ive patient was probed wh e re a s nant hnRNP-A2. Selection and cloning both with the natural and with the re- n o rmal human IgG was not re a c t ive of scFv was performed essentially as combinant antigen. Sera were diluted (lane A). Immunoblotting and ELISA described. Reactive scFv were then ex- 1:50 in assay buffer (20 mM Tris-Hcl, detected no antibody reactivity to other pressed in soluble form and tested by pH 7.4, 150 mM NaCl, 0.1% Triton-X- nuclear antigens or cardiolipin, respec- ELISA for reactivity against recombi- 100) and incubated on a rocking plat- t ive ly (not shown). A ffi n i t y - p u ri fi e d nant antigen and by immu n o bl o t t i n g form for 40 min at room temperature. anti-RA33 antibodies obtained by blot for reactivity against the natural anti- After washing the nitrocellulose strips elution from a strongly reactive serum gen and potential crossreactivities with were incubated with alkaline phospha- and scFv of anti-RA33 antibodies other nu clear antigens. Similar to t a s e - c o n j u gated go at anti-human or (cloned from a patient-derived V gene serum antibodies the scFv recognized anti-mouse IgG secondary antibodies l i b ra ry) we re also highly re a c t ive in hnRNP-A2 as well as its alternatively diluted 1:2000 (Accurate Sciences, this assay (Fig. 1b) and did not crossre- spliced forms Bl and B2 and showed Westbury, NY, USA). The blots were act with other nuclear proteins includ- modest crossreactivity with hnRNP-Al d eveloped with BCIP/NBT alkaline ing hnRNP-Al (not shown). but not with other nuclear antigens in- phosphatase substrate (Sigma). cluding snRNP associated pro t e i n s . Results of autoantibody evaluation These data and a detailed analysis of Detection of other antinuclear in sera of mice immunized with the molecular pro p e rties of the scFv antibodies anti-RA33 enriched human IgG will be published elsewhere. D e t e rm i n ation of mouse and human preparations autoantibodies to other nu clear auto- The results of the first bleeding sug- Immunization of mice a n t i gens was perfo rmed by immu n o- gested that none of the sera obtained Three groups of mice (15 BALB/c and blotting as described ab ove using in one month after boost immu n i z at i o n 15 C57BL/6J mice in each group) were house reference sera. To confirm the exhibited any specifi c - re a c t ivity to

519 Anti-RA33 autoantibodies in mice / G. Steiner et al.

(a) (b)

Fig. 1. (a) Immunoblot determination of anti-RA33 reactivity of human IgG fractions using recombinant hnRNP-A2 (RA33) antigen expressed as 68 kD fusion protein. Lane A, normal human IgG that was not reactive with the RA33 antigen. Lane B, IgG preparation obtained from an anti-RA33 positive patient with RA, that exhibits strong anti-RA33 reactivity. Lane C, IgG preparation from the weakly reactive anti-RA33 patient. (b) Immunoblot determination of anti-RA33 reactivity of affinity-purified and scFv anti-RA33 antibodies using recombinant hnRNP-A2 (RA33). (1) Patient’s serum; (2) Affinity-purified anti-RA33 antibody; (3) scFv anti-RA33 antibody. Affinity-purified anti-RA33 antibodies and scFv anti-RA33 antibody were highly reactive.

(a)

Fig. 2. (a) Immunoblot analysis of autoantibod- ies to nuclear antigens in mice immunized with IgG preparations one month after boost. Sera f rom immu n i zed mice we re inve s t i gated by immunoblotting using a HeLa nuclear extract as antigenic source and human patients’sera as ref- erences. I g G - C o n t rol Ð IgG prep a ration from healthy control person IgG-RA1 Ð IgG preparation from RA patient 10 IgG-RA2 Ð IgG preparation from RA patient 2 (1) anti-RA33 re fe rence serum re c og n i z i n g hnRNP-A2/B1/B2, (2) anti-U1 RNP reference serum recognizing the U1 RNP specific proteins U1-70K, U1-A and U1-C, (3) anti-Sm reference serum recognizing the Sm proteins B/B« and D. (b) The positions of autoantigens stained by the ref- erence sera are indicated. Strong reactivities to the U1-A antigen can be clearly seen in lanes 3, 4,and 7 of the IgG-RA1 group and in lanes 5 and 8 of the IgG-RA2 group. (b) Reactivity of anti-U1-A positive sera with recombinant antigens as determined by line im- munoassay (LIA). To confirm the data obtained with the nuclear blot reactive sera were addition- ally assayed by the LIA, which employs recom- binant autoantigens immobilized on a PVDF membrane. Human sera served again as refer- ences. (1-4) Mouse sera reactive with U1-A on nuclear blots; (5) Anti-Ro/La reference serum; (6) Anti-Sm/U1 RNP reference serum; (7) Anti- topoisomerase I reference serum; (8) Anti-Jo-1 reference serum (recognizing Ro52 in addition). All 4 mouse sera we re cl e a rly re a c t ive with RNP-A (i.e. U1-A) but with none of the other autoantigens on the strip demonstrating the high specificity of this autoimmune reaction.

520 Anti-RA33 autoantibodies in mice / G. Steiner et al.

RA33 antigen. Interestingly, however, reactivity to the U1-A snRNP antigen was detected in 4 of the 15 C57BL/6J mice immunized with anti-RA33 posi- tive IgG preparations and in 4 of 15 C57BL/6J mice immunized with IgG from the anti-RA33 borderline patient, whereas such reactivities were not seen clearly in the control group immunized with IgG from a normal human donor (Fig. 2). Three months after the boost, 3 of 15 C57BL/6J mice immunized with anti- Fig. 3. Detection of anti-RA33 autoantibodies by Western blotting in mice immunized with AP or scFv anti-RA33 antibodies or normal human IgG 3 months after the boost. Sera were investigated by RA33 positive IgG fractions and one immunoblotting using purified recombinant hnRNP-A2 fusion protein as antigen (see also Fig. 1). The mouse immunized with anti-RA33 bor- position of the fusion protein is indicated by an arrow. Bands visible below the major band represent d e rline IgG still had autoantibodies degradation products. (R) Human anti-RA33 reference serum.In sera from several animals immunized against U1-A snRNP. Reactivities to with AP anti-RA33 antibody pronounced reactivities are clearly seen (lanes 1, 2,4,9), weaker reactiv- ities are visible in lanes 3, 5, and 10. Sera from two mice immunized with scFv RA33 antibody show RA33 antigen were seen only in sera also strong reactivity (lanes 1 and 11), while weak reactivities can be seen in lanes 2, 12 and 13. from one BALB/c mouse immunized with bord e rline anti-RA33 IgG and weak re a c t ivity was detected in one these sera demonstrating the transient ies. In add i t i o n , re a c t ivity to RA33 C57BL/6J mouse immunized with high nature of those additional responses. antigen was seen in the sera of 2 mice titer anti-RA33 IgG; this mouse devel- immunized with scFv of anti-RA33 an- oped also antibodies to U1-A snRNP. Histopathological findings tibodies. It was found that these reac- No distinct histopathological findings tivities appeared only 3 months after Development of autoantibodies in were observed in the mice immunized the boost immunization. Therefore the C57BL/6J mice immunized with AP with the different IgG preparations,AP major immune response was noted anti-RA33 antibodies and scFv anti-RA33 antibodies and scFv of anti- against RA33 antigen (hnRNP-A2). fragments of anti-RA33 antibodies RA33 antibodies. These observat i o n s These murine anti-RA33 autoantibod- Since immunization with IgG prepara- included the absence of synovitis, sero- ies are likely to be representative of an- tions had not led to the induction of sitis or myositis in all the groups of tibodies populations produced in one of anti-RA33 antibodies, we next employ- mice. the steps of the idiotypic cascade. ed high titer affinity-purified antibodies The other interesting data observed in as well as scFv anti-RA33 antibodies Discussion the current study is a tendency to gen- (cloned from a V gene library derived This study describes an attempt to in- eration of an immune response against from a anti-RA33 positive patient with vestigate the nature of the serological U1-A snRNP antigen in mice immu- RA). Similar to the previously obtained and pathological response, induced in nized with anti-RA33 enriched prepa- results one out of 13 mice immunized mice following immunization with dif- rations. Immunization with scFv frag- with AP anti-RA33 antibodies and 3 of fe rent types of anti-RA33 antibodies ments of anti-RA33 also induced pro- the mice immu n i zed with scFv dis- such as anti-RA33 positive IgG prepa- duction of autoantibodies to U1-A sn- p l ayed re a c t ivities to U1-A snRNP rat i o n , AP anti-RA33 antibodies and R N P, p rior to production of mu ri n e antigen one month following the boost. scFv of anti-RA33 antibodies. anti-RA33 antibodies. On the basis of Anti-RA33 autoantibodies were found The results of the study show that anti- above-listed findings we suggested that in sera of one mouse immunized with bodies recognizing RA33 antigen were im mu n i z a tion of mice with autoanti- AP anti-RA33 antibodies (Table I). detected in 7 sera from 13 mice immu- bodies against one component of the However, 3 months after the boost anti- nized with purified anti-RA33 antibod- spliceosome (hnRNP) resulted in tran- RA33 reactivity was clearly seen in the sera of 7 out of the 13 mice immunized Table I. Autoantibody profiles in sera of C57BL/6J mice immunized with AP anti-RA33 with AP anti-RA33 antibodies and in 2 antibodies and scFv with anti-RA33 antibodies (one month after the boost immunization). mice immunized with scFv of RA33 antibodies with weak reactivities being Antigens Mice immunized Mice immunized with Mice immunized with d e t e c t able in 3 additional animals. with AP A2/RA33 ScFv A2/RA33 normal human IgG (n = 13) (n = 15) (n = 10) Mice immunized with normal human Number positive, #mice Number positive, #mice Number positive, #mice IgG did not show re a c t ivity aga i n s t human recombinant RA33 antige n RA33 1 (#8) Not detected Not detected (Fig. 3). Interestingly, no reactivities to RNP-U1-A 1 (#10) 3 (#4,7,8) Not detected other nuclear antigens were found in

521 Anti-RA33 autoantibodies in mice / G. Steiner et al. sient “ a u t o a n t i b o dy spre a d ” aga i n s t nization with PPGMRPP which is the and snRNP components of the spliceo- other component of spliceosome (sn- major B-cell epitope of Sm autoantigen some in RA patients as compared to RNP). These results may provide an resulted in the production of the anti- SLE and MCTD patients (17). An alter- additional support for the hypothesis of bodies against PPGMRPP, but did not native explanation for the appearance “autoantibody spread” which proposed induce autoantibodies against nat ive of anti-UlA antibodies prior to anti- similar to the novel concept of "epitope forms of Sm and Ul RNP antigens. RA33 would be that the anti-UlA anti- spread". Mechanisms underlying “epitope” and body could actually be identical with The concepts of the “epitope spread” “autoantibody spread” are still under the anti-anti-RA33 antibody (i.e. Ab2) have been noticed following immuni- debate. However, it has been suggested mimicking part of the antigen’s struc- z ation with diffe rent autoantigens. It by several authors (23-26) that B cells ture. Given that hnRNP-A2 and U1-A has been shown that immunization of bearing specificities to distinct compo- may directly interact in vivo (at the 5’ animals with peptides or antigens can nents on determinants within cert a i n splice site), binding of Ab2 to U1-A lead to the production of autoantibod- molecules (e.g. snRNP complexes, Ro/ might actually mimic pro t e i n - p ro t e i n ies reactive with different epitope with- La particles) are capable of capturing interaction between U1-A and hnRNP- in the same molecule (intramolecular and presenting the endogenously dri- A2. Such direct interaction between the spread) and sometimes to other mole- ven “products” to T helper cells which two proteins has not been demonstrated cules (intermolecular spread) (22-26). could, in turn, stimulate B cells which yet but it is known that both hnRNP-Al Similar to this novel concept of “epi- possess diffe rent autoantibody speci- and A2 are in close neighborhood to Ul tope spread” the hypothesis of “autoan- ficities. snRNP in the pre-splicing complex (37, t i b o dy spre a d ” was proposed by us We used for immunization anti-RA33 38). (27). It states that immunization with antibodies which were obtained from In this study we were not able to show respective antibodies results in the pro- the sera of patients with RA, these in- h i s t o p at h o l ogical manife s t ations con- duction of autoantibodies with different terspliceosomal “autoantibody spread” sistent with those of RA in humans by a n t i genic specifi c i t i e s , although re- is of a particular interest. As known, i m mu n i z ation with diffe rent types of stricted to the disease (21, 27-33). Fur- anti-RA33 patients have a more restric- anti-RA33 antibodies. Investigations of thermore, from our previous observa- tive response to the spliceosome, they etiology and pathogenesis of RA have tions and those of others, we suggested react only to hnRNP proteins, particu- suggested that RA is a multifactorial, that active immunization with autoanti- larly to the “RA33 complex” (14-17). i m mu n o l ogi c a l ly mediated disease, b o d i e s , c a rrying diffe rent idiotypes In contra s t , p atients with SLE and although the original initiating stimulus could lead both to the production of MCTD may exhibit reactivity to auto- has not yet been characterized.The role “autoantibody spread” with disease re- a n t i gens of both hnRNP and snRNP of anti-RA33 antibodies in pathogene- stricted pattern (28, 29, 32) or predomi- complexes of spliceosome (14, 15, 17). sis and their influence on clinical mani- nant autoantibody response to antigens In addition, the detailed epitope map- fe s t ations has not been known ye t . from a target complex (34). ping study with sera from patients with Thus, in patients with RA no correla- The diversification of the immune re- R A , SLE and MCTD revealed that tion between anti-RA33 reactivity and sponse to components of the spliceo- major autoepitopes of hnRNP-A2 loca- clinical or serological features (includ- some following immunization with an- ted in the RNA binding regions are dif- ing rheumatoid factor) was seen (10, tigens was exemplified in several mur- ferent in these groups of the patients 14). In contrast, presence of anti-RA33 ine models. In 1994, Mamula and col- (17). Thus, the complete RNA binding antibodies in SLE patients was associ- l e agues (22) have demonstrated that domain (RBD II), which is composed ated with erosive artropathy (39, 40). mice immunized with Ul snRNP A pro- of 95 amino acids, was shown to be Absence of histopathological findings tein generated autoantibodies directed required for binding of the anti-RA33 in immunized mice can be explained by against individual proteins of the Ul autoantibodies from patients with RA needs of additional factors in propaga- s n R N P, in addition to anti-UlA anti- and SLE. This epitope region was not tion of RA. Thus, the antibodies are bodies. They had previously shown that recognized by antibodies from patients presumably unable to recognize their such mice developed snRNP specific with MCTD wh i ch targeted another nuclear target under normal physiolog- a u t o re a c t ive T-cells (23). James and lager epitope consisted of RBD I and ical conditions. However, aberrant ex- colleagues (35) have shown that immu- two thirds of RBD II (17). This differ- p ression of hnRNP-A2 under condi- n i z ation of normal rabbits and mice ential epitope and antigen recognition tions of ch ronic infl a m m ation as re- with a peptide fragment of Sm-B can patterns observed for RA (A2-RBD II cently observed in human synovial tis- induce an anti-Sm-B antibody re- only), SLE (A2-RBD II, Sm, and Ul sue (unpublished observation) might sponse, which then spreads to induce snRNP), and MCTD (A2-RBD I+II, Ul render the antigen accessible to the im- antibodies against other snRNP poly- snRNP) suggested highly specifi c mune system leading to loss of toler- p eptides (including Ul-70kD) and a u t o i m mune re s p o n s e, e s p e c i a l ly in ance and subsequently to pathological against dsDNA, but not to hnRNP-A2. RA patients (17). These findings prob- e ffects. A l t e rn at ive ly, p roduction of In contra s t , V l a ch oyiannopoulos and ably may exclude possibility of close- murine anti-RA33 antibodies without c o l l e agues (36) rep o rted that immu- linked reactivity between anti-hnRNP clinical findings compat i ble with the

522 Anti-RA33 autoantibodies in mice / G. Steiner et al. respective human disease may reflect rheumatoid arthritis. Arthritis Rheum 1989; 23. FATENEJAD S, MAMULA MJ, CRAFT J: Role only epiphenomena. 32: 1515-20. of intermolecular/intrastructural B- and T-cell 11. STEINER G, H A RTMUTH K, SKRINER K e t determinants in the diversification of autoan- In summary, our studies have shown al.: Purification and partial sequencing of the tibodies to ribonucleoprotein particles. Proc that active immunization of naive mice nuclear autoantigen RA33 shows that it is in- Natl Acad Sci USA 1993; 90: 12010-4. with AP anti-RA33 antibodies and with distinguishable from the A2 protein of the he- 24. TOPFER F, GORDON T, McCLUSKEY J: Intra- recombinant scFv fragments of anti- terogeneous nuclear ribonucleoprotein com- and intermolecular spreading of autoimmuni- plex. J Clin Invest 1992; 90: 1061-6. ty involving the nuclear self-antigens La (SS- RA33 antibodies resulted on the one 12. SMOLEN JS, HASSFELD W, GRANINGER W, B) and Ro (SS-A). Proc Natl Acad Sci USA hand in the production of murine anti- STEINER G: Antibodies to antinuclear subsets 1995; 92: 875-9. RA33 autoantibodies. On the other in systemic lupus erythematosus and rheuma- 25. LEHMANN PV, FORSTHUBER T, MILLER A, hand, mice immunized with the differ- toid arthritis. Clin Exp Rheumatol 1990; 8: SERCARZEE: Spreading of T-cell autoimmu- 41-4. nity to cryptic determinants of an autoanti- ent types of anti-RA33 antibody prepa- 13. BURD CG, S WANSON MS, G O R L ACH M, gen. Nature 1992; 358: 155-7. ration developed an early but transient DREYFUSS G: Primary structures of the hete- 26. KAUFMAN DL, CLARE-SALZLER M, TIAN J: in t e r spliceosomal “a u t o a n t i b o d y sprea d ” rogeneous nuclear ribonucleoprotein A2, B1, Spontaneous loss of T-cell tolerance to glu- and C2 proteins:A diversity of RNA binding tamic acid decarboxylase in murine insulin- to snRNP complex of the spliceosome. proteins is generated by small peptide inserts. dependent diabetes. Nature 1993; 366:69-72. These results may provide evidences to Proc Natl Acad Sci USA 1989; 86: 9788-92. 27. GEORGE J, GILBURD B, SHOENFELD Y: Au- the diversification in autoantibody pro- 14. HASSFELD W, STEINER G, S T U D N I C K A - toantibody spread may explain multiple anti- duction and probably is caused by dis- BENKE A et al.: Autoimmune response to the bodies. The Immunologist 1999; 7: 193-8. spliceosome. An immunologic link between 28. MENDLOVIC S, BROCKE S, SHOENFELD Y et regulation of idiotypic network. rheumatoid arthritis, mixed connective tissue al.: Induction of a systemic lupus erythema- disease, and systemic lupus erythematosus. tosus-like disease in mice by a common hu- References Arthritis Rheum 1995; 38: 777-85. man anti-DNA idiotype. Proc Natl Acad Sci 1.TAN EM,CHAN EK,SULLIVAN KF,RUBINRL: 15. M E Y E R O, TAUXE F, FA B R E G A S D et al. : USA 1988; 85: 2260-4. Antinuclear antibodies (ANAs):diagnostical- Anti-RA 33 antinuclear autoantibody in rheu- 29. BLANK M, M E N D E L OVIC S, MOZES E, ly specific immune markers and clues toward matoid arthritis and mixed connective tissue SHOENFELDY: Induction of SLE-like disease the understanding of systemic autoimmunity. d i s e a s e : C o m p a rison with antike ratin and in naive mice with a monoclonal anti-DNA Clin Immunol Immunopathol 1988; 47: 121- antiperinuclear antibodies. Clin Exp Rheuma - antibody derived from a patient with poly- 41. tol 1993; 11: 473-8. myositis carrying the 16/6 ID. JAutoimmun 2.TAN EM: Antinuclear antibodies: Diagnostic 16. ASTALDI RICOTTI GC, BESTAGNO M, CERI- 1988; 1: 683-91. markers for autoimmune diseases and probes NO A et al.: Antibodies to hnRNP core pro- 30. TOMER Y, GILBURD B, BLANK M et al.: Cha- for cell biology. Adv Immunol 1989; 44: 93- tein A1 in connective tissue diseases. J Cell racterization of biologically active antineu- 151. Biochem 1989; 40: 43-7. t rophil cytoplasmic antibodies induced in 3.CHOI YD, GRABOWSKI PJ, SHARP PA, DREY- 17. SKRINER K, SOMMERGRUBER WH, TREM- mice. Pathogenetic role in experimental vas- FUSS G: Heterogeneous nuclear ribonucleo- MEL V et al.: Anti-A2/RA33 autoantibodies culitis. Arthritis Rheum 1995; 38: 1375-81. p ro t e i n s : Role in RNA splicing. S c i e n c e are directed to the RNA binding region of the 31. SWISSA M, LERNER A, SASAKI T, SELA E, 1986; 231: 1534-9. A2 protein of the heterogeneous nuclear ri- BLANK M, SHOENFELD Y:The diverse pa- 4.LERNER MR, STEITZ JA: Antibodies to small b o nu cl e o p rotein complex. Diffe rential ep i- t h ogenic potential of anti-DNA antibodies nuclear RNAs complexed with proteins are tope recognition in rheumatoid arthritis, sys- from various sources to induce experimental p roduced by patients with systemic lupus temic lupus erythematosus,and mixed conne- systemic lupus erythematosus. Pathobiology e ry t h e m atosus. P roc Natl Acad Sci USA ctive tissue disease. J Clin Invest 1997; 100: 1996; 64: 32-9. 1979; 76: 5495-9. 127-35. 32. DANG H,OGAWA N,TAKEI M,LAZARIDIS K, 5.VAN V E N ROOIJ W J, SILLEKENS P T: S m a l l 18. D U M O RTIER H, M O N N E AUX F, JA H N - TALAL N: Induction of lupus-associated auto- nuclear RNA associated proteins: autoanti- SCHMID B et al.:B and T cell responses to the antibodies by immunization with native and gens in connective tissue diseases. Clin Exp spliceosomal heterogeneous nuclear ribonu- recombinant Ig poly p eptides ex p ressing a Rheumatol 1989; 7: 635-45. cleoproteins A2 and B1 in normal and lupus cross-reactive idiotype 4B4. J Immunol 1993; 6.P I N O L - ROMA S, CHOI Y D, M ATUNIS MJ, mice. J Immunol 2000; 165: 2297-305. 151: 7260-7. DREYFUSS G: Immunopurification of hetero- 19. DE WILDT RM,FINNERN R,OUWEHANDWH: 33. P iERANGELI SS, HARRIS E N: Induction of geneous nuclear ribonucleoprotein particles Characterization of human variable domain phospholipid-binding antibodies in mice and reveals an assortment of RNA-binding pro- antibody fragments against the U1 RNA-as- rabbits by immunization with human beta 2 teins. Genes Dev 1988; 2: 215-27. sociated A protein, selected from a synthetic glycoprotein 1 or anticardiolipin antibodies 7.KUMAR A,WILLIAMS KR, SZER W: Purifica- and patient-derived combinatorial V gene li- alone. Clin Exp Immunol 1993; 93: 269-72. tion and domain stru c t u re of core hnRNP brary. Eur J Immunol 1996; 26: 629-39. 34. GEORGE J, G I L BURD B, LEVY Y et al. : proteins A1 and A2 and their relationship to 20. MEHEUS L, VAN VENROOIJ WJ, WIIK A et al.: ‘Autoantibody dominance’ pattern following single-stranded DNA-binding proteins. J Biol Multicenter validation of recombinant, natur- idiotypic manipulation of naive mice by im- Chem 1986; 261: 11266-73. al and synthetic antigens used in a single mul- mu n i z ation with anti-U1RNP antibodies. 8.FRITZLER MJ, ALI R, TA N E M: A n t i b o d i e s tiparameter assay for the detection of specific Pathobiology 1997; 65: 204-9. from patients with mixed connective tissue anti-nuclear autoantibodies in connective tis- 35. JA M E S JA , G RO S S T, SCOFIELD RH, H A R- disease react with heterogeneous nuclear ri- sue disorders. Clin Exp Rheumatol 1999; 17: LEY JB: Immunoglobulin epitope spreading bonucleoprotein or ribonucleic acid (hnRNP/ 205-14. and after peptide immu- RNA) of the nuclear matrix. J Immunol 1984; 21. BAKIMER R,FISHMAN P, BLANKM, SREDNI nization: Sm B/B’-derived PPPGMRPP and 132: 1216-22. B, DJALDETTI M, SHOENFELD Y: Induction PPPGIRGP induce spliceosome autoimmuni- 9.GELPI C, RODRIGUEZ-SANCHEZ JL,HARDIN of pri m a ry antiphospholipid syndrome in ty. J Exp Med 1995; 181: 453-61. JA: Purification of hnRNP from HeLa cells mice by immunization with a human mono- 36. VLACHOYIANNOPOULOS PG, PETROVAS C, with a monoclonal antibody and its applica- clonal anticardiolipin antibody (H-3). J Clin TZIOUFAS AG et al.: No evidence of epitope tion in ELISA: detection of autoantibodies. Invest 1992; 89: 1558-63. spreading after immunization with the major Clin Exp Immunol 1988; 71: 281-8. 22. MAMULAMJ, FATENEJAD S, CRAFTJ: B cells Sm epitope P-P-G-M-R-P-P anchored to se- 10. HASSFELD W, STEINER G , HARTMUTH K et process and present lupus autoantigens that quential oligo p eptide carri e rs (SOCs). J al.: Demonstration of a new antinuclear anti- initiate autoimmune T cell responses. J Im - Autoimmun 2000; 14: 53-61. body (anti-RA33) that is highly specific for munol 1994; 152: 1453-61. 37. BUVOLI M, COBIANCHI F, RIVAS: Interac-

523 Anti-RA33 autoantibodies in mice / G. Steiner et al.

tion of hnRNP A1 with snRNPs and pre - KRAINER A R : Function of conserved do- thematosus. J Rheumatol 1994; 21: 1260-3. mRNAs: evidence for a possible role of A1 mains of hnRNP A1 and other hnRNP A/B 40. RICHTER COHEN M , STEINER G, S M O L E N RNA annealing activity in the first steps of proteins. EMBO J 1994; 13: 5483-95. JS, ISENBERG DA: Erosive arthritis in sys- spliceosome assembly. N u cleic Acids Res 39. IS E N B E R G DA , STEINER G, S M O L E N J S: temic lupus erythematosus:Analysis of a dis- 1992; 20: 5017-25. Clinical utility and serological connections of tinct clinical and sero l ogical subset. Br J 38. MAY E DA A , M U N ROE SH, C ACERES JF, anti-RA33 antibodies in systemic lupus ery- Rheumatol 1998; 37: 421-4.

524