Tissue Diseases with Scleroderma and Other Connective Microfibrillar Protein, Fibrillin-1, in Patients Autoantibodies to The

Autoantibodies to the Extracellular Matrix Microfibrillar Protein, Fibrillin-1, in Patients with Scleroderma and Other Connective Tissue Diseases This information is current as of September 25, 2021. Filemon K. Tan, Frank C. Arnett, Stephan Antohi, Shinichiro Saito, Adriana Mirarchi, Harry Spiera, Takeshi Sasaki, Ozaki Shoichi, Ken Takeuchi, Janardan P. Pandey, Richard M. Silver, Carwile LeRoy, Arnold E. Postlethwaite and Constantin A. Bona Downloaded from J Immunol 1999; 163:1066-1072; ; http://www.jimmunol.org/content/163/2/1066 http://www.jimmunol.org/ References This article cites 22 articles, 5 of which you can access for free at: http://www.jimmunol.org/content/163/2/1066.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 25, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Autoantibodies to the Extracellular Matrix Microﬁbrillar Protein, Fibrillin-1, in Patients with Scleroderma and Other Connective Tissue Diseases1

Filemon K. Tan,2* Frank C. Arnett,* Stephan Antohi,† Shinichiro Saito,† Adriana Mirarchi,† Harry Spiera,† Takeshi Sasaki,‡ Ozaki Shoichi,§ Ken Takeuchi,¶ Janardan P. Pandey,ͦͦ Richard M. Silver,** Carwile LeRoy,ͦͦ Arnold E. Postlethwaite,# and Constantin A. Bona†

A duplication in the fibrillin-1 gene has been implicated as the cause of the tight skin 1 (tsk1) phenotype, an animal model of scleroderma or systemic sclerosis (SSc). In addition to the production of abnormal fibrillin-1 protein, the tsk1 mouse also produces autoantibodies to fibrillin-1. Among a population of Choctaw Native Americans with the highest prevalence of SSc yet described,

a chromosome 15q haplotype containing the fibrillin-1 gene has been strongly associated with SSc. With a recombinant human Downloaded from fibrillin-1 protein, autoantibodies to fibrillin-1 were detected in the sera of Native American SSc patients that correlated significantly with disease. Abs to fibrillin-1 also were detected in sera from Japanese, Caucasian, and African-American SSc patients. Compared with other ethnic groups, Japanese and Native American SSc patients had significantly higher frequencies of anti- fibrillin-1 Abs. Sera from patients with diffuse SSc, calcinosis, Raynaud’s, esophageal dysmotility, sclerodactyly, and telangiectasias syndrome and mixed connective tissue disease also had significantly higher frequencies of anti-fibrillin-1 Abs than sera from

controls or patients with other non-SSc connective tissue diseases (lupus, rheumatoid arthritis, and Sjo¨gren’s syndrome). Ab http://www.jimmunol.org/ specificity for fibrillin-1 was demonstrated by the lack of binding to a panel of other purified autoantigens. The results presented demonstrate for the first time the presence of high levels of anti-fibrillin-1 Abs in a significant portion of patients with SSc. The Journal of Immunology, 1999, 163: 1066–1072.

cleroderma or systemic sclerosis (SSc) 3 is a human mul- topoisomerase I (Scl-70), centromeric proteins, RNA polymerases, tisystem disease characterized by prominent cutaneous and others (2, 3). These autoantibodies appear to be T cell depen- S and visceral fibrosis (1). Although the etiopathogenesis of dent and show strong associations with certain MHC or HLA class SSc is unknown, both genetic and environmental contributions are II alleles (4–6). suspected, and an autoimmune component is evidenced by the The tight skin 1 (tsk1) mouse is a potential genetic animal model by guest on September 25, 2021 presence of highly disease-specific circulating autoantibodies of human SSc in which thickened skin and visceral fibrotic lesions which target nuclear and nucleolar autoantigens, including DNA arise from the accumulation of extracellular matrix (ECM) components (7). Moreover, as in human SSc serum, autoantibodies to *Division of Rheumatology and Clinical Immunogenetics, Department of Internal DNA topoisomerase I and RNA polymerase I have been demon- Medicine, University of Texas Health Science Center, Houston, TX 77030; †Depart- strated in tsk1 animals (3). The mutation causing tsk1 recently has ment of Microbiology, Mount Sinai Medical School, New York, NY 10029; ‡Second Department of Internal Medicine, Tohoku University School of Medicine, Sendai, been shown to be a 30-kb duplication in the mouse fibrillin-1 gene Japan; §Department of Medicine and Clinical Science, Kyoto University, Graduate fbn1 (8). Mice homozygous for the fbn1 gene duplication die in School of Medicine, Kyoto, Japan; ¶Department of Rheumatology, Juntenko School ͉͉ utero, and the tsk1 phenotype results from the heterozygous state. of Medicine, Tokyo, Japan; Department of Immunology and Microbiology, and **Division of Rheumatology and Immunology, Department of Medicine, Medical Fibrillin-1 is the major structural glycoprotein of connective tis- University of South Carolina, Charleston, SC 29425; and #Division of Connective sue microfibrils, which are important components of elastic fibers Tissue Diseases, University of Tennessee, Memphis, TN 38163 widely distributed throughout the body (9). Mutations of fibrillin-1 Received for publication November 16, 1998. Accepted for publication May 10, 1999. in humans cause the Marfan syndrome, a single gene autosomal The costs of publication of this article were defrayed in part by the payment of page dominant disease of connective tissue (10). Recent histomorpho- charges. This article must therefore be hereby marked advertisement in accordance logical studies of the ECM in tsk1 animals have demonstrated two with 18 U.S.C. Section 1734 solely to indicate this fact. distinct populations of microfibrils, one representing abnormally 1 This work was supported by the National Institute of Arthritis and Musculoskeletal assembled microfibrils containing a larger mutant fibrillin-1 and and Skin Diseases Specialized Center of Research in Scleroderma (P50AR44888 F.C.A.), National Institutes of Health CAP Award (3 M01 RR02558-12S1 to F.K.T.), the other normal microfibrils and wild-type fibrillin-1 (11). as well as grants from the RGK Foundation (to F.C.A.), Scleroderma Foundation (to Using the tsk1 mutation as a candidate genetic region for sus- F.C.A.), National Institute of Allergy and Infectious Diseases (R01-AI24671-11 to C.B.), and the U.S. Department of Energy (DE-FC02-98CH10902 to J.P.P. and ceptibility to human SSc, we have recently studied polymorphic R.M.S.). microsatellite markers on human chromosome 15q, which contains 2 Address correspondence and reprint requests to Dr. Filemon K. Tan, Division of the region homologous to murine tsk1 (fbn1), in a Native American Rheumatology and Clinical Immunogenetics, University of Texas Medical population with a high prevalence of SSc (12, 13). A 2-cM hap- School, 6431 Fannin, MSB 5.626, Houston, TX 77030. E-mail address: [email protected] lotype containing the human fibrillin-1 gene (FBN1) showed a 3 Abbreviations used in this paper: SSc, scleroderma or systemic sclerosis; tsk1, tight striking association with SSc in this population. Southern blotting skin 1 mouse; ECM, extracellular matrix; fbn1, fibrillin-1 gene; CREST, calcinosis, of FBN1 in affected individuals showed no gross gene duplication Raynaud’s, esophageal dysmotility, sclerodactyly, and telangiectasias; MCTD, mixed connective tissue disease; SLE, systemic lupus erythematosus; RA, rheumatoid ar- (13), and nucleotide sequencing of FBN1 in affected individuals is thritis; rFbn1, recombinant fibrillin-1 fragment. currently ongoing (14). Subsequent pulse-chase experiments of

Table I. Binding of puriﬁed anti-ﬁbrillin-1 Abs to various Ags

Ag Coating Microtiter CREST Plates SSc (cpm) (cpm)

BSA 104 Ϯ 14 109 Ϯ 13 Fibrillin-1 2774 Ϯ 28 3268 Ϯ 85 Collagen I 95 Ϯ 16 171 Ϯ 72 Collagen III 82 Ϯ 20 105 Ϯ 24 Collagen IV 98 Ϯ 3 145 Ϯ 13 Collagen VI 106 Ϯ 6 144 Ϯ 51 Actin 101 Ϯ 7 212 Ϯ 20 Histone 2A 97 Ϯ 16 226 Ϯ 8 Histone 2B 90 Ϯ 14 163 Ϯ 52 Histone 4 79 Ϯ 13 147 Ϯ 4 Intrinsic factor 94 Ϯ 0 179 Ϯ 1 Transferrin 98 Ϯ 25 128 Ϯ 25 Sm (Smith) 137 Ϯ 18 126 Ϯ 25 Cardiolipin 98 Ϯ 18 134 Ϯ 6 Thyroglobulin 180 Ϯ 17 279 Ϯ 38 Myelin basic protein 265 Ϯ 16 238 Ϯ 28 TEPC 183a 115 Ϯ 1 152 Ϯ 28 MOPC 21 37 Ϯ 627Ϯ 0 Downloaded from Leu 4 31 Ϯ 17 47 Ϯ 8 TEPC 1017 202 Ϯ 21 255 Ϯ 1 MOPC 870 118 Ϯ 10 98 Ϯ 69

a TEPC 183, MOPC 21, and TEPC 1017 are murine myeloma proteins with unknown binding activity, Leu-4 is a mAb speciﬁc for TSH receptor, and MOPC 870 is a myeloma protein speciﬁc for Salmonella tranaroa lipopolysaccharide. http://www.jimmunol.org/

ment. CREST syndrome (calcinosis, Raynaud’s, esophageal dysmotility, FIGURE 1. Lack of inhibition of binding of anti-fibrillin-1 Abs by other sclerodactyly, and telangiectasias) was diagnosed on clinical grounds and autoantigens. The inhibition of the binding of anti-fibrillin-1 Abs to mi- the presence of anti-centromere Abs. Mixed connective tissue disease crotiter plates coated with fibrillin-1 fusion protein by various Ags in a (MCTD) also was diagnosed clinically based on overlapping features of patient with SSc (A) and another with CREST syndrome (B). Microtiter scleroderma (Raynaud’s phenomenon, puffy hands), myositis and/or sys- plates were coated with various amounts of Ags, and after extensive wash- temic lupus erythematosus along with the presence of serum autoantibodies to U1-RNP (18, 19). Those MCTD patients who demonstrated unequivocal ing and blocking with BSA, the plates were incubated with CREST or SSc by guest on September 25, 2021 serum diluted 1:200. Binding was detected with goat 125I-␥ F(ab)Ј anti-human evidence of diffuse SSc (i.e., proximal skin involvement) as defined above 2 were subclassified as MCTD-SSc. The diagnosis of the localized form of IgG. The results shown are averages of triplicate experiments Ϯ SD. scleroderma, morphea, was based on clinical diagnosis, usually with skin biopsy confirmation. Systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) patients fulfilled their respective American College of Rheu- fibrillin-1 metabolism in explanted SSc fibroblasts have demon- matology criteria (20, 21), whereas the criteria for poly- and dermatomy- strated normal synthesis and secretion of the protein but abnormal ositis were those suggested by Bohan and Peter (22). Primary Sjo¨gren’s ECM incorporation (15). At the same time, Murai et al. (16) found syndrome was defined on clinical grounds. Normal sera were from healthy adult Caucasian, African-American, and that tsk1 mice spontaneously produce IgG autoantibodies to a re- Japanese medical center personnel or blood donors with no history of au- combinant protein from the C region of human fibrillin-1. Thus, toimmune diseases. In addition, sera were included from normal Choctaw because of both genetic and immunological data potentially link- Native Americans attending the same clinics and matched for blood quan- ing both the murine tsk1 phenotype and human SSc to fibrillin-1, tum as the Choctaw SSc cases from whom samples had been collected and stored previously (12). we sought to determine whether patients with SSc or other connective tissue diseases spontaneously produce autoantibodies to a recombinant human fibrillin-1 protein. The results presented dem- Autoantigens onstrate for the first time the presence of anti-fibrillin-1 Abs in a The recombinant human fibrillin-1 peptide (rFbn1) was generated from the high proportion of patients with SSc. Ab specificity for fibrillin-1 pET-3xa Escherichia coli expression vector kindly supplied by Dr. was demonstrated by the lack of binding to other purified autoan- Francesco Ramirez (Brookdale Center for Molecular Biology, Mount Sinai School of Medicine, New York, NY) and purified from bacterial lysate as tigens, including collagens I and III. previously described (16). rFbn1 is 30 kDa long and contains the proline- rich C region (aa 395–446) of fibrillin-1 (23). rFbn1 was chosen because Materials and Methods autoantibodies produced by the tsk1 mice bind to it; and it contains the Patients and controls proline-rich region characteristic of fibrillin-1. Thyroglobulin, cardiolipin, actin, intrinsic factor, transferrin, and collagens were purchased from Stored sera from patients with SSc, other connective tissue diseases and Sigma (St. Louis, MO), topoisomerase I from Life Technologies (Gaith- normal controls collected at several academic medical centers were studied ersburg, MD), and RNA polymerase I from Pharmacia (Piscataway, NJ). retrospectively for autoantibodies to fibrillin-1. Because of evidence for Human myelin basic protein was a gift from Dr. G. Lewis (Institute for association of fibrillin-1 with SSc in Choctaw Native Americans, sera from Basic Research Development Disability, Staten Island, NY; Sm (Smith) Choctaw SSc cases and normal Choctaw controls collected earlier (12) was a gift from D. H. Dang (University South of Texas Medical School at were included but were analyzed separately. All patients with SSc, except San Antonio, TX); and the histones were a gift from Dr. T. Fesy (Depart- one Choctaw subject, fulfilled the American College of Rheumatology cri- ment of Pathology, Mount Sinai School of Medicine, New York, NY). teria (formerly the American Rheumatism Association) for SSc (17). Pa- TEPC 183, TEPC 1017, MOPC 21 and MOPC 870 are murine myeloma tients with SSc were further classified into those having diffuse vs limited proteins kindly supplied by Dr. M. Porter (National Institutes of Health, cutaneous involvement, with diffuse SSc being defined as skin thickening Bethesda, MD) and Leu-4 mAb was provided by Dr. B. Erlinger (Columbia involving areas proximal to elbows and/or knees, excluding facial involve- University, New York, NY). 1068 ANTI-FIBRILLIN-1 Abs IN SYSTEMIC SCLEROSIS

FIGURE 2. Dose-response relationship between serum dilution and binding to BSA and rFbn1. The binding of Abs was assessed as described in Materials and Methods. Titer of serum Ab was considered the highest dilution that gave 50% binding. Representative sera from Choctaw SSc patients are shown. Negative control (no sera added) for

the detecting Ab indicated by “0” on the x-axis. Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

Radioimmunoassay Statistical analysis Direct binding assay for anti-fibrillin-1 Abs. Microtiter plates were In addition to standard descriptive statistics, nonparametric tests (Staview coated with BSAor various Ags 0.3–1 ␮g/ml in carbonate buffer, pH 9.6, 5.0, SAS Institute, Cary, NC) were used to determine significant differ- at 4°C overnight. After washings, the plates were blocked with 3% BSA in ences in anti-fibrillin-1 levels between scleroderma patients and controls PBS for 4 h, washed three times, and incubated for 2h with 1:100 dilution within each ethnic group (Mann-Whitney U test), and between controls ϫ 4 125 Ј of sera. After thorough washing, 5 10 cpm of I-labeled goat F(ab )2 from different ethnic groups (Kruskal-Wallis). The two-tailed Fisher’s ex- anti-human ␮ or ␥ chains (BioSource International, Camarillo, CA) in 50 act test were used to test for statistically significant differences in autoan- ␮l were added and incubated for 2 h. After extensive washing, bound tibody frequencies. EPI-INFO (Version 6.04b, Centers for Disease Control radioactivity was measured in a gamma counter. Patient and normal control and Prevention, Atlanta, GA) and a public domain Fisher’s exact test pro- sera were assigned specimen numbers and analyzed blindly by one of us gram (http://www.nr.no/home/langsrud/fisher.htm) was used for these (C.B.) with no prior knowledge of clinical diagnosis. RIAs were routinely calculations. performed in paralleled triplicates, the background binding of BSA was subtracted, and Ab binding was expressed as cpm. Positive binding was defined as Ն2 SD above the mean of the appropriate ethnically matched Results control group. Specificity of anti-fibrillin-1 autoantibodies Competitive inhibition of anti-fibrillin-1 Abs. Experiments were con- ducted to determine the concentration of purified human anti-fibrillin-1 IgG anti-fibrillin-1 Abs from the sera of five SSc patients showing Abs for maximal binding to rFbn1 with microplates coated with 10 ␮g/ml high binding in the RIA were purified on a fibrillin-1 Sepharose 4B rFbn1. Abs were purified from patient’s sera on a fibrillin-1 protein-Sepha- column and used for competitive inhibition in RIA. A dose-de- rose 4B column. Abs at concentrations corresponding to 50% binding were incubated with various Ags at different dilutions and then transferred to pendent inhibition of binding of autoantibodies was observed with fibrillin-1 or BSA-coated plates. the rFbn1, but no significant inhibition was observed with topoisomerase I, RNA polymerase I, or centromere protein (Fig. 1). Western blot analysis of anti-fibrillin-1 Abs These results demonstrated the specificity of purified anti- fibrillin-1 Abs from SSc patients. It is well known that some au- One microgram per ml of rFbn1 was fractionated by SDS-PAGE (Bio-Rad, toantibodies such as anti-ssDNA or dsDNA autoantibodies display Hercules, CA). Proteins were transferred onto PVDF membranes by semi- dry electrophoresis at 100 mA current for 1 h using a Multiphor II appa- nonspecific binding properties characteristic of natural Abs. Thus, ratus (Pharmacia). Membranes were blocked overnight at 4°C with 5% additional experiments examined the binding of purified anti- fat-free milk (Carnation, San Francisco, CA), and then incubated for2hat fibrillin-1 autoantibodies to a panel of target autoantigens from room temperature with 1 ␮g/ml rabbit anti-fibrillin-1, or 20 ␮g/ml purified various autoimmune diseases, as well as to murine IgG to deter- anti fibrillin-1 Abs from a patient with SSc, or CREST, previously described, or with 20 ␮g/ml human IgG Abs. Bound Abs were visualized mine rheumatoid factor activity. The results shown in Table I dem- with 125I-labeled goat anti-rabbit IgG Abs or anti-human IgG Abs (3 ϫ 105 onstrate that the purified anti-fibrillin-1 Ab bound only to rFbn1 cpm/ml) for 90 min, washed, and exposed to x-ray film. and not to other autoantigens, BSA, collagen, or IgG. Because The Journal of Immunology 1069

FIGURE 3. Dose response relationship between serum dilution and binding to rFbn1 in SSc, CREST and MCTD sera. The binding of Abs was assessed as described in Materials and Methods. Titer of serum Ab was considered the highest dilution that gave 50% binding. Representative patients with SSc are shown in the top two panels; CREST patients (left) and MCTD Downloaded from patients (right) in the lower panels. http://www.jimmunol.org/

rFbn1 is a fusion protein, its fusion partner was tested for Ab 50% rFbn1 binding was observed with 1:300–1:1000 dilution of by guest on September 25, 2021 reactivity. In a group of 51 diffuse SSc patients having anti- three SSc and two CREST sera and with one of four MCTD sera fibrillin-1 Abs, no binding to the glutathione S-transferase fusion studied. Taken together, the results demonstrate that anti-fibrillin-1 partner higher than BSA background was observed (data not Abs showed only fibrillin-1-specific binding. shown). The binding of Abs to fibrillin-1 was also confirmed by Western Dose-effect binding was performed on a selected panel of pa- blotting (Fig. 4). The results show that purified anti-fibrillin-1 Abs tient sera with SSc and other connective tissue diseases. The data from a SSc patient (Fig. 4, lane 2) and a CREST patient (Fig. 4, presented in Fig. 2 demonstrate that in three Choctaw SSc patients, lane 3), as well as Abs from an immunized rabbit (Fig. 3, lane 1) no significant binding to BSA was observed with SSc sera (Fig. 2, bound to rFbn1, but control human IgG (Fig. 4, lane 4) showed no top), and 50% binding to rFbn1 was observed with 1:300–1:1000 binding. These Abs did not exhibit any nonspecific binding to dilution (Fig. 2, bottom) of SSc sera. In Fig. 3, the data show that markers such as phosphorylase B, OVA, or lysozyme.

Anti-fibrillin-1 Abs in scleroderma patients from various ethnic groups With the exception of African-Americans, SSc patients from all other ethnic groups studied had significantly higher levels of anti- fibrillin-1 Ab compared with their respective controls (Table II). However, the mean levels of anti-fibrillin-1 IgM were not significantly elevated in the Caucasian SSc compared with Caucasian controls (Table II). Sera from 13 of 16 (81%) Choctaw Native Americans with SSc had IgG Abs to fibrillin-1 compared with only 1 of 20 (5%) Choctaw controls ( p ϭ 2 ϫ 10Ϫ6) (Table II). Sim- ilarly, a high frequency of anti-fibrillin-1 IgG Abs was observed in Japanese SSc patients (78%) compared with Japanese controls (12%, p ϭ 2 ϫ 10Ϫ6) (Table II). In both of these ethnic groups, the FIGURE 4. Western blot analysis of purified human anti-fibrillin-1 Abs. Lane 1, polyclonal rabbit anti-fibrillin-1; lane 2, chromatographically pu- frequency of anti-fibrillin-1 IgM was also significantly increased rified anti-fibrillin-1 Ab from a SSc patient; lane 3, chromatographically compared with controls. Only 31% of Caucasian SSc patients were purified anti-fibrillin-1 Ab from a CREST patient; lane 4, human IgG my- positive for IgG anti-fibrillin-1, although the frequency was still eloma protein. significantly elevated compared with Caucasian controls (Table 1070 ANTI-FIBRILLIN-1 Abs IN SYSTEMIC SCLEROSIS

Table II. Mean levels and frequencies of anti-ﬁbrillin-1 Abs in diffuse scleroderma cases and controls from various ethnic groups

Mean level Ϯ SD Mean Levels of Anti-Fibrillin-1 Frequency of Ab Frequency N (cpm) (SSc vs control)a Anti-Fibrillin-1b (SSc vs control)c

Choctaw controls 20 IgM 361 Ϯ 347 0 IgG 629 Ϯ 1445 1 (5%) Choctaw SSc 16 IgM 3325 Ϯ 3723 0.0001 9 (56%) 1 ϫ 10Ϫ4 IgG10902 Ϯ 9921 Ͻ0.0001 13 (81%) 2 ϫ 10Ϫ6 Japanese controls 17 IgM 762 Ϯ 382 0 IgG 1284 Ϯ 1365 2 (12%) Japanese SSc 54 IgM 7866 Ϯ 8237 0.0006 32 (59%) 7 ϫ 10Ϫ6 IgG15194 Ϯ 11977 Ͻ0.0001 42 (78%) 2 ϫ 10Ϫ6 Caucasian controls 51 IgM 1450 Ϯ 2301 1 (2%) IgG 965 Ϯ 777 3 (6%) Caucasian SSc 128 IgM 1639 Ϯ 2301 NS 7 (6%) NS IgG 3088 Ϯ 6569 0.0003 40 (31%) 2 ϫ 10Ϫ4 African-American controls 20 IgM 1092 Ϯ 876 1 (5%) IgG 1999 Ϯ 5255 1 (5%) African-American SSc 27 IgM 1473 Ϯ 2632 NS 2 (7%) NS IgG 3491 Ϯ 6783 NS 1 (4%) NS

Mann-Whitney U test. Downloaded from b Positive anti-ﬁbrillin-1 deﬁned as Ն2 SD above the mean of the appropriate control group. c Comparison within each ethnic group by the two-tailed Fisher’s exact test.

II). In contrast, African-American SSc patients did not have sig- of low titer, also were found in Յ10% of patients with SLE, pri- nificantly elevated levels of anti-fibrillin Ab compared with Afri- mary Sjo¨gren’s syndrome, and normal controls. can-American controls. The frequency of anti-fibrillin-1 IgG was significantly higher in http://www.jimmunol.org/ Table III compares the frequencies of anti-fibrillin-1 Abs in SSc patients with diffuse skin disease compared with those with Caucasian, African-American, Japanese, and Choctaw Native limited skin involvement ( p ϭ 9 ϫ 10Ϫ6) (Table IV). Only seven American SSc patients. Choctaw Native Americans and Japanese patients with morphea were studied; none was positive for anti- SSc patients had significantly increased frequencies of anti- fibrillin-1 (Table IV). In general, the highest levels of anti- fibrillin-1 Abs compared with Caucasians and African-Americans. fibrillin-1 occurred among patients with diffuse SSc, although a On the other hand, anti-fibrillin-1 Abs were significantly less fre- few patients with limited SSc, CREST syndrome, and MCTD and quent in African-Americans with SSc than in Caucasians. MCTD-SSc expressed similar levels. by guest on September 25, 2021 Anti-fibrillin-1 Abs in scleroderma subsets and other connective tisssue diseases Discussion The high frequency of anti-fibrillin-1 autoantibodies in the SSc Fibrillin-1 is the major component of microfibrils in the ECM patients prompted an investigation of scleroderma subsets and found in skin and other connective tisssue (9). Comparisons of other connective tissue diseases. Mean levels of IgM and IgG an- mouse and human fibrillin-1 transcripts reveal Ͼ95% DNA and tifibrillin-1 among Caucasian, African-American, and Japanese amino acid sequence homology, implying strict structural require- controls showed no statistically significant differences (Kruskal- ments for the assembly and integrity of extracellular microfibrils Wallis test, IgM, p ϭ 0.37; IgG, p ϭ 0.89); and these three groups (24). As a result of a gene duplication, the tsk1 mouse, an animal were combined (Table IV). Compared with pooled controls, the model of fibrosis and perhaps SSc, produces an abnormally large frequencies of anti-fibrillin-1 Abs (especially IgG) were signifi- fibrillin-1 protein (11). Recently, we demonstrated that the tsk1 cantly increased in patients (combined ethnic groups except fibrillin-1 bound larger than normal amounts of TGF-␤ when com- Choctaw) with diffuse SSc, with CREST syndrome, and in MCTD, pared with wild-type fibrillin-1 by virtue of a duplication of latent MCTD-SSc and poly- and dermatomyositis (Table IV). In con- TGF binding protein-like domains (30). It also has been demon- trast, anti-fibrillin-1 was found infrequently in sera of patients with strated that these mice spontaneously produce anti-fibrillin-1 Abs SLE, Sjo¨gren’s syndrome, or RA. IgG antifibrillin-1 Abs, usually (16), as well as other SSc-specific autoantibodies (3).

Table III. Comparison of the frequency of anti-ﬁbrillin-1 Abs in scleroderma patients from various ethnic groupsa

Positive Anti-Fibrillin-1 N (IgG or IgM) pb

Choctaws 16 15 (94%) vs Caucasians, 4 ϫ 10Ϫ6 vs African-Americans, 1 ϫ 10Ϫ8 Caucasians 128 43 (34%) vs African-Americans, 0.005 African-American 27 2 (7%) Japanese 54 47 (87%) vs Caucasians, 1 ϫ 10Ϫ11 vs African-Americans, 2 ϫ 10Ϫ12 vs Choctaws, NS

a Patients of unknown ethnicity excluded. b Two-tailed Fisher’s exact test. The Journal of Immunology 1071

Table IV. Frequencies of anti-ﬁbrillin-1 Abs among non-Choctaw patients with scleroderna and other autoimmune diseases

Frequency of Ab Frequency Disease N Isotype Anti-Fibrillin-1a (disease vs controls)b

Controlsc 88 IgM 2 (2%) IgG 4 (5%) Diffuse SSc 257 IgM 51 (20%) 4 ϫ 10Ϫ5 IgG 94 (37%) 6 ϫ 10Ϫ8 Limited SScd 65 IgM 7 (11%) 0.04 IgG 6 (9%) NS CREST 37 IgM 15 (41%) 9 ϫ 10Ϫ8 IgG 19 (51%) 8 ϫ 10Ϫ8 Morphea 7 IgM 0 NS IgG 0 NS MCTD ϩ SSc 10 IgM 1 (10%) NS IgG 3 (30%) 0.009 MCTD 59 IgM 10 (17%) 0.004 IgG 20 (34%) 2 ϫ 10Ϫ5 SLE 79 IgM 0 NS IgG 1 (1%) NS Sjo¨gren’s syndrome 33 IgM 0 NS Downloaded from IgG 2 (6%) NS RA 35 IgM 0 NS IgG 0 NS Polymyositis-dermatomyositis 10 IgM 0 NS IgG 4 (40%) 0.009

a Positive anti-fibrillin-1 Abs defined as Ն2 SD above the mean of the pooled control group. b Two-tailed Fisher’s exact test. http://www.jimmunol.org/ c Normal Caucasian, African-American, and Japanese controls (Table II) were pooled. No significant difference in anti- fibrillin levels was found among the three ethnic control groups by the Kruskal-Wallis test (IgM, p ϭ 0.37; IgG, p ϭ 0.89). d Compared with diffuse SSc, frequency of anti-fibrillin IgM p ϭ 0.10, IgG p ϭ 9 ϫ 10Ϫ6.

The data presented here demonstrate for the first time sponta- The frequencies of anti-fibrillin-1 Abs in SSc patients of differ- neously occurring anti-fibrillin-1 Abs in human SSc. These Abs ent ethnic backgrounds also showed striking differences (see Table exhibited specificity for a recombinant fibrillin-1 protein. Autoan- III). Choctaw Native Americans and Japanese patients were pos- tibody binding in a RIA was inhibited by recombinant fibrillin-1 itive for anti-fibrillin-1 Abs (94 and 87%, respectively) signifi- protein but not by three other scleroderma autoantigens (DNA to- cantly more often than Caucasians (34%) and African-Americans by guest on September 25, 2021 poisomerase I, RNA polymerase I, and centromeric proteins). Fur- (4%). The reasons for these ethnic differences are unknown but thermore, anti-fibrillin-1 Abs failed to bind to a panel of tissue and could be due to genetic variation in the immune response to fibril- nuclear/cytoplasmic autoantigens. In addition, the presence of anti- lin-1. Both Native Americans and the Japanese are relatively ge- fibrillin-1 Abs was confirmed by Western blotting. netically homogeneous groups who may share more common an- Anti-fibrillin-1 Abs occurred most commonly and at highest lev- cestral backgrounds than the other two ethnic groups studied. In els in patients with diffuse SSc, but also in some patients with other fact, recent comparative analyses of clinical and serological fea- fibrosing scleroderma-like syndromes, including the CREST vari- tures in SSc patients of different ethnic backgrounds have shown ant and MCTD, and in those who had evolved into diffuse SSc. that the diffuse form of SSc and anti-topoisomerase I Abs are more Two patients with CREST and two with MCTD show a dose-effect typical clinical features in the Choctaw and Japanese populations binding pattern for anti-fibrillin-1 Abs (Fig. 3). Patients defined as than in Caucasians (26). Moreover, sera from Japanese SSc having limited SSc had a significantly lower frequency and gen- patients preferentially recognize the amino-terminal portion of erally lower levels of anti-fibrillin-1 Abs than those with diffuse topoisomerase I, whereas sera from Caucasians and African-Amer- SSc, suggesting that there might be a correlation between the ex- icans more frequently recognize the carboxy terminus; Choctaw tent of cutaneous fibrosis and the autoantibody response. This in- SSc sera tend to recognize both (26). There are also ethnically terpretation would be premature, however, in view of the findings related genetic similarities and differences at the MHC level. Anti- in some patients with CREST syndrome and morphea, in whom topoisomerase I-positive Choctaw and Japanese SSc patients show ,1502ءand DRB1 1602ءcutaneous involvement is limited. Moreover, it must be empha- associations with HLA-DR2 haplotypes (DRB1 sized that this was a retrospective study and the most accurate and respectively), whereas their Caucasian and African-American –1101ءquantitative measures of cutaneous involvement, i.e., skin scores counterparts possess HLA-DR5 (DR11) haplotypes (DRB1 Thus, it is possible that SSc patients in this .(28 ,27 ,12) (1104ء could not be obtained. Thus, a prospective study correlating ,(25) skin scores with both the presence and levels of anti-fibrillin-1 Abs study, particularly Caucasians and African-Americans, will subse- will be necessary to answer this question. quently be found to recognize other portions of the fibrillin-1 pro- IgM Abs to fibrillin-1 occurred infrequently in other connective tein. Only one small fragment of the large fibrillin-1 protein was tissue diseases compared with SSc patients (see Table IV). The examined for Ab reactivity in this study. In addition, possibly dif- clinical and biological significance of these findings remains un- ferent HLA associations may be found between SSc patients who clear, but the high IgG autoantibody levels suggest a T-cell-me- do or do not produce autoantibodies to the recombinant fibrillin-1 diated Ag-driven response. The Ab response against the large gly- protein studied here. Such investigations are in progress. coprotein fibrillin-1 likely requires CD4ϩ T cell and B cell Another possible reason for the ethnic differences in anti- cooperation. Further studies aimed at determining the frequency of fibrillin-1 Ab frequencies may be genetic heterogeneity in the etio- fibrillin-1-specific T cells are warranted. pathogenesis of SSc. Perhaps the Choctaw show the highest anti- 1072 ANTI-FIBRILLIN-1 Abs IN SYSTEMIC SCLEROSIS

fibrillin-1 Ab frequencies because SSc in this population is caused DQB1 first domain with the anticentromere autoantibody response in systemic by a genetically mediated abnormality of fibrillin-1, whereas dif- sclerosis (scleroderma). J. Clin. Invest. 89:1208. 6. Kuwana, M., T. A. Medsger, Jr., and T. M. Wright. 1995. T cell proliferative ferent genetic mechanisms may play roles in other ethnic groups, response induced by DNA topoisomerase I in patients with systemic sclerosis and especially Caucasians and African-Americans. In previous studies, healthy donors. J. Clin. Invest. 96:586. 7. Green, M. C., H. O. Sweet, and L. E. Bunker. 1976. Tight-skin, a new mutation we have found a striking association of SSc in the Choctaw with of the mouse causing excessive growth of connective tissue and skeleton. a haplotype on chromosome 15q containing the fibrillin-1 gene Am. J. Pathol. 82:493. (13). It is not yet known whether the fibrillin-1 gene is abnormal 8. Siracusa, L. D., R. McGrath, Q. Ma, J. J. Moskow, J. Mane, P. J. Christner, A. M. Buchberg, and S. A. Jimenez. 1996. A tandem duplication within the in these SSc cases; however, examination of fibrillin-1 metabolism fibrillin 1 gene is associated with the mouse tight skin mutation. Genome Res. in explanted fibroblasts from these SSc cases have shown abnor- 6:300. mal incorporation into the ECM, and electron microscopic studies 9. Reinhardt, D. P., D. R. Keene, G. M. Corson, E. Poschl, H. P. Bachinger, J. E. Gambee, and L. Y. Sakai. 1996. Fibrillin-1: organization in microfibrils and have revealed markedly diminished microfibril content and mor- structural properties. J. Mol. Biol. 258:104. phologically abnormal microfibrils (15). 10. Dietz, H., and R. Pyeritz. 1995. Mutations on the human gene for fibrillin-1 in the Marfan syndrome and related disorders. Hum. Mol. Genet. 4:1799. Although the studies described here clearly demonstrate that a 11. Kielty, C., M. Raghunath, L. D. Siracusa, M. J. Sherratt, R. Peters, large proportion of patients with SSc spontaneously produce spe- C. A. Shuttleworth, and S. A. Jimenez. 1998. The tight skin mouse: demonstra- cific autoantibodies to fibrillin-1, it is unclear whether this immune tion of mutant, fibrillin-1 production and assembly into abnormal microfibrils. J. Cell Biol. 140:1159. response plays a primary role in disease pathogenesis or is a sec- 12. Arnett, F. C., R. F. Howard, F. Tan, J. M. Moulds, W. B. Bias, E. Durban, ondary phenomenon. Fibrillin-1, being a major component of skin, H. D. Cameron, G. Paxton, T. W. Hodge, P. E. Weathers, and J. D. Reveille. is a likely target of the SSc pathological process. As discussed 1996. Increased prevalence of systemic sclerosis in a native American tribe in Oklahoma. Arthritis Rheum. 39:1362. above, the fibrillin-1 produced by SSc fibroblasts appears to be 13. Tan, F. K., D. N. Stivers, M. W. Foster, R. Chakraborty, R. F. Howard, Downloaded from unstable or unusually susceptible to degradation, perhaps by ex- D. M. Milewicz, and F. C. Arnett. 1998. Microsatellite markers near the fibrillin-1 gene on human chromosome 15q are associated with scleroderma in a Na- tracellular proteases. Fragmented fibrillin-1 could reveal cryptic tive American population. Arthritis Rheum. 41:1729. epitopes, which could become the targets of an immune response. 14. Tan, F. K., J. T. Le, D. Milewicz, and F. C. Arnett. 1998. Sequencing of the Thus, the presence of antifibrillin-1 Abs may be a secondary pro- fibrillin-1 gene in native Americans with scleroderma (systemic sclerosis or SSc). Arthritis Rheum. 41:S103(Abstr.). cess resulting from presentation of proteolytic fragments in a host 15. Wallis, D. D., F. K. Tan, F. C. Arnett, and D. Milewicz. 1998. Abnormal fibril- with altered immune tolerance. Alternatively, an autoimmune re- lin-1 incorporation into the extracellular matrix of scleroderma dermal fibro- http://www.jimmunol.org/ sponse to ECM components may be a potential mechanism for blasts. Arthritis Rheum. 41:S103(Abstr.). 16. Murai, C., S. Saito, K. N. Kasturi, and C. A. Bona. 1998. Spontaneous occurrence fibrosing disease, because binding of autoantibodies to the target of anti-fibrillin-1 autoantibodies in tight skin mice. Autoimmunity 28:151. autoantigen could cause immunologic injury. Some data from an- 17. Masi, A. T., G. P. Rodnan, T. A. Medsger, Jr., R. D. Altman, W. A. D’Angelo, imal models are more consistent with the former hypothesis. De- J. F. Fries, E. C. LeRoy, A. B. Kirsner, A. H. MacKenzie, D. J. McShane, A. R. Myers, and G. C. Sharp. 1980. Preliminary criteria for the classification of spite anti-fibrillin-1 autoantibody production, the tsk1 mouse does systemic sclerosis (scleroderma). Arthritis Rheum. 23:581. not develop the inflammatory component seen in human early SSc 18. Sharp, G. C., W. S. Irvin, E. M. Tan, R. G. Gould, and H. R. Holman. 1972. Mixed connective tissue disease: an apparently distinct rheumatic disease syn- skin lesions which is characterized by perivascular and interstitial drome associated with a specific antibody to an extractable nuclear antigen lymphocytic infiltrates (7). Furthermore, Kasturi et al. (29) have (ENA). Am. J. Med. 52:148. demonstrated that B-lymphocyte deficient tsk1 mice (tsk/ϩ, 19. Kasturi, K. N., A. Hatakeyama, H. Spiera, and C. A. Bona. 1995. Antifibrillarin by guest on September 25, 2021 Ϫ Ϫ autoantibodies present in systemic sclerosis and other connective tissue diseases JHD / ), still develop cutaneous fibrosis, thus suggesting a dis- interact with similar epitopes. J. Exp. Med. 181:1027. sociation of autoimmunity and cutaneous hyperplasia. 20. Tan, E. M., A. S. Cohen, J. F. Fries, A. T. Masi, D. J. McShane, N. F. Rothfield, Finally, are autoantibodies to fibrillin-1 potentially useful in as- J. G. Schaller, N. Talal, and R. J. Winchester. 1982. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum. 25:1271. sessing the clinical diagnosis or prognosis of patients with con- 21. Arnett, F. C., S. M. Edworthy, D. A. Bloch, D. J. McShane, J. F. Fries, nective tissue diseases? Could future fibrosing disease be predicted N. S. Cooper, L. A. Healey, S. R. Kaplan, M. H. Liang, H. S. Luthra, T. A. Medsger, Jr., D. M. Mitchell, et al. 1988. The American Rheumatism in individual patients by measuring these autoantibodies? Al- Association 1987 revised criteria for the classification of rheumatoid arthritis. though the data presented here demonstrate that antifibrillin-1 Abs Arthritis Rheum. 31:315. are associated with scleroderma syndromes, additional prospective 22. Bohan, A., and J. B. Peter. 1975. Polymyositis and dermatomyositis (parts 1 and 2). N. Engl. J. Med. 292:344,403. studies in connective tissue diseases are essential to answering 23. Zhang, H., S. D. Apfelroth, W. Hu, E. C. Davis, C. Sanguineti, J. Bonadio, those questions. Nonetheless, the finding of an autoantibody to a R. P. Mecham, and F. Ramirez. 1998. Structure and expression of fibrillin-2, a component of the ECM, the probable primary focus of the patho- novel microfibrillar component preferentially located in elastic matrices. J. Cell Biol. 124:855. logical process in SSc, is provocative and opens new potential 24. Yin, W., E. Smiley, J. Germiller, C. Sanguineti, T. Lawton, L. Pereira, avenues of research into both pathogenesis and clinical manage- F. Ramirez, and J. Bonadio. 1995. Primary structure and developmental expres- ment of scleroderma syndromes. sion of Fbn-1, the mouse fibrillin gene. J. Biol. Chem. 270:1798. 25. Steen, V. D., T. A. Medsger, and G. P. Rodnan. 1982. D-Penicillamine therapy in progressive systemic sclerosis (scleroderma): a retrospective analysis. Ann. In- References tern. Med. 97:652. 26. Kuwana, M., J. Kaburaki, F. C. Arnett, R. F. Howard, T. A. Medsger, Jr., and 1. Medsger, T. A., Jr. 1993. Systemic sclerosis (scleroderma), localized forms of T. M. Wright. 1999. Influence of ethnic background on clinical and serologic scleroderma, and calcinosis. In Arthritis and Allied Conditions. features in patients with systemic sclerosis and anti-DNA topoisomerase I anti- D. J. McCarty and W. J. Koopman, eds. Lea and Febiger, Philadelphia, pp. body. Arthritis Rheum. 42:465. 1253–1299. 27. Kuwana, M., J. Kaburaki, Y. Okano, H. Inoko, and K. Tsuji. 1993. The HLA-DR 2. Arnett, F. C. 1995. HLA and autoimmunity in scleroderma (systemic sclerosis). and DQ genes control the autoimmune response to DNA topoisomerase I in Int. Rev. Immunol. 12:107. systemic sclerosis (scleroderma). J. Clin. Invest. 92:1296. 3. Bona, C., and N. Rothfield. 1994. Autoantibodies in scleroderma and tight skin 28. Reveille, J. D. 1993. Molecular genetics of systemic sclerosis. Curr. Opin. Rheu- mice. Curr. Opin. Rheumatol. 6:931. matol. 5:753. 4. Reveille, J. D., E. Durban, M. J. MacLeod, R. Goldstein, R. Moreda, 29. Kasturi, K. N., A. Hatakeyama, C. Murai, R. Gordon, R. G. Phelps, and R. D. Altman, and F. C. Arnett. 1992. Association of amino acid sequences in the C. A. Bona. 1997. B cell deficiency does not abrogate development of cutaneous HLA-DQB1 first domain with the anti-topoisomerase I autoantibody response in hyperplasia in mice inheriting defective fibrillin-1 gene. J. Autoimmun. 10:505. scleroderma (progressive systemic sclerosis). J. Clin. Invest. 90:973. 30. Saito, S., H. Nishimura, T. D. Brumeaunu, S. Casares, A. C. Stien, T. Honjo, and 5. Reveille, J. D., D. Owerbach, R. Goldstein, R. Moreda, R. A. Isern, and C. A. Bona. 1999. Characterization of mutated protein encoded by a partially F. C. Arnett. 1992. Association of polar amino acids at position 26 of the HLA- duplicated fibrillin-1 gene in tight skin (tsk) mice. Mol. Immunol. In press.