Immunoblot analysis of JS-1–glutathione S-transferase (GST) recombinant fusion protein with representative serum samples from patients with primary Sjögren syndrome. Lanes 1 and 2, Coomassie brilliant blue staining of the purified GST protein alone (27 kd) and the JS-1–GST fusion protein (98 kd). Lanes 3 and 4, GST and JS-1–GST protein reacted with patient serum samples, without incubation, showing binding to the fusion protein. Lanes 5 and 6, JS-1–GST fusion protein reacted with the same serum samples, preincubated with JS-1–GST and GST protein, respectively.
Correlation between anti–60-kd SS-A (Ro) and anti–52-kd SS-A (Ro) antibodies. Relative antibody titers are indicated as optical density units (ODUs). Three serum samples, positive for anti–JS-1 from patients with lupus erythematosus (LE) alone, are indicated by thick arrows. Two serum samples, negative for anti–JS-1 from patients with primary Sjögren syndrome (SS), are indicated by thin arrows.
Preliminary comparison of clinical and laboratory findings in JS-1–positive vs JS-1–negative patients. Upper bars indicate patients with anti–JS-1 antibody. Asterisk indicates P<.01 by χ2 analysis with Yates correction; SS, Sjögren syndrome; LE, lupus erythematosus; and dagger, P<.01 by Fisher exact test.
Watanabe T, Tsuchida T, Kanda N, Mori K, Hayashi Y, Tamaki K. Anti–α-Fodrin Antibodies in Sjögren Syndrome and Lupus Erythematosus. Arch Dermatol. 1999;135(5):535-539. doi:10.1001/archderm.135.5.535
To investigate the prevalence of anti–α-fodrin antibody in patients with Sjögren syndrome (SS), lupus erythematosus (LE), or both and the association of this antibody with other clinical manifestations.
A study of screening and diagnostic tests. Mean follow-up was 152 months (range, 4-572 months).
A university hospital associated with a research laboratory in Tokyo, Japan.
Nine patients with primary SS, 15 patients with SS secondary to LE, and 44 patients with LE alone.
Main Outcome Measures
Frequencies of clinical and laboratory findings, including anti–α-fodrin antibody.
Anti–α-fodrin antibody was more commonly detected in patients with primary (7/9; P<.001) and secondary (9/15; P<.001) SS than in those with LE alone (3/44). When patients with primary and secondary SS were combined and compared with those with LE alone, the sensitivity of anti–α-fodrin antibody was 67%, specificity was 93%, and both positive and negative predictive values were 84%. The presence of anti–α-fodrin antibody was associated with pernio, hyperglobulinemia, rheumatoid factor positivity, and the presence of anti–SS-B (La) antibody (P<.01) but not with annular erythema, photosensitivity, vasculitis, or renal disorder.
Although anti–α-fodrin antibody was detected in patients with SS and in those with LE, it seemed to be more valuable for the diagnosis of SS than was anti–SS-A (Ro) because anti–α-fodrin was much less prevalent in patients with LE alone. It may be possible to consider this novel autoantibody as pathophysiologically associated with some extraglandular manifestations characteristically seen in patients with SS.
Sjögren Syndrome (SS) and lupus erythematosus (LE) are diseases of unknown cause and uncertain pathogenesis. No single clinical or laboratory finding has been suggested as the sine qua non diagnostic criterion. The American Rheumatism Association1 and European Community Study Group2 conducted large, multicenter studies and proposed preliminary diagnostic criteria for systemic LE (SLE) and SS, respectively. The differential diagnosis of these 2 diseases is sometimes problematic because a close serologic relationship has long been recognized. Autoantibodies to the small ribonucleoprotein particles SS-A (Ro) occur in 70% to 90% of patients with SS and in 40% to 60% of patients with LE,3- 6 whereas the incidence of anti–SS-A (Ro) in patients with rheumatoid arthritis (RA), scleroderma, and polymyositis is relatively low (10%-15%).7 Provost et al8 described 10 patients with SS and LE who were positive for anti–SS-A (Ro) and pointed out a close clinical and pathological relationship. In addition, several authors9- 11 reported that annular erythema develops in Asian patients with SS and LE, but attempts to identify autoantibodies specific for this skin lesion have been unsuccessful.12 There is clearly a need for a new laboratory test to discriminate SS from LE.
In 1997, Haneji et al13 found that serum samples from an NFS/sld mouse model of human SS reacted with a 120-kd protein, α-fodrin, which was specifically expressed in the lesional salivary glands. There is some evidence that this protein plays a critical role in the development of exocrinopathy. First, it induces proliferative T-cell responses and their type I cytokine production. Second, neonatal immunization with this protein prevented development of the disease in mice. Third, serum samples from 41 (95%) of 43 patients with primary SS and 5 (63%) of 8 patients with secondary SS reacted with this antigen by immunoblotting, even though serum samples from all SLE, RA, and healthy controls were negative. Fodrin is a major component of the membrane cytoskeleton of most eukaryotic cells and is speculated to be involved in exocrinal secretion.14,15 It forms a heterodimer composed of α (240-kd) and β (235-kd) subunits. It was recently observed (Hinoko Inoue, oral communication, Department of Microbiology, Tokyo Medical School, October 14, 1998) that the 240-kd α subunit is degraded to 120-kd products, probably via calpain I (Ca++-dependent cysteine proteinase), in Epstein-Barr virus–infected cells immediately before the induction of apoptotic death . The overactivation of calpain in synovial fluid and subsequent cartilage destruction are also suggested to be part of the pathomechanisms of RA.16 Anticalpastatin antibodies that may break the calpain-calpastatin balance and enhance proteolytic activity of calpain are specifically detected in serum samples from more than 50% of patients with RA.17,18 Considered together, it is not surprising that α-fodrin, a substrate of calpain, is involved in the development of SS.
We investigated (1) whether anti–α-fodrin antibody is specific enough, even if validated with a large number of patients with SS and LE, and (2) whether there is any correlation of reactivity between anti–α-fodrin and anti–SS-A (Ro) antibody. We also examined the association between anti–α-fodrin and the clinical manifestations characteristically seen in patients with SS and LE.
The serum samples analyzed in this study were collected from 9 patients with SS, 15 patients with SS secondary to LE (SS+LE), and 44 patients with LE alone whose disease course had been followed up by members of the Department of Dermatology at Tokyo University Hospital between January 1995 and December 1997. All 24 patients with SS fulfilled the 1993 European Community Study Group criteria for SS.2 Diagnosis of LE was made by the presence of characteristic morphologic features and confirmed by skin biopsy results.19 Diagnosis of SLE was based on the 1982 American Rheumatism Association criteria.1 Patients with an associated connective tissue disorder, including RA, dermatomyositis and polymyositis, and scleroderma, were excluded.20,21
Antibody binding to the 60- and 52-kd components of SS-A (Ro) antigens was quantified using a commercially available solid-phase enzyme-linked immunosorbent assay (ELISA) kit22,23 according to the manufacturer's protocol (Medical & Biological Laboratories Co, Nagoya, Japan). The optical density (OD) of ELISA plates was read at 450 nm and expressed as OD units, which were calculated as follows: (OD of sample−OD of negative control)/(OD of positive control−OD of negative control)×100. We used the positive and negative control serum samples that were provided with the kit. In addition, we withdrew 10 negative control serum samples from the World Health Organization, Geneva, Switzerland, and the National Serum Reference Bank/Tokyo (National Institute of Infectious Diseases, Tokyo) and confirmed that the cutoff points indicated in the protocol were appropriate. We defined the findings of anti–SS-A (Ro) antibody as implying a serum sample that was reactive to either or both of the 60- and 52-kd components of SS-A(Ro) antigen. Precipitating autoantibodies against SS-B (La) were routinely examined by Ouchterlony double immunodiffusion.24 To prepare the α-fodrin protein as the antigen source for immunoblotting, JS-1 complementary DNA25—encoding the NH2 terminal portion (base pairs 1-1784) of human nonerythroid α-fodrin (7787 nucleotides)—was inserted into the EcoRI site of the expression vector pGEX-4T-2. Glutathione S-transferase (GST) and recombinant JS-1–GST fusion protein were expressed and purified as previously described.26 Electrophoretic blotting was performed using the method of Towbin et al27 with some modifications. Glutathione S-transferase was used as a negative control protein because it is a component of the JS-1–GST fusion protein. To assess the specificity of antibody binding to the JS-1–GST fusion protein, patient serum samples were preincubated with purified JS-1–GST fusion protein and GST protein.
Correlations between antibody profiles and clinical manifestations were made after completion of all laboratory studies. Possible differences in the frequency of anti–JS-1 and anti–SS-B (La) responses among primary SS, SS secondary to LE, and LE alone were assessed using the Fisher exact test. Other statistical significances were determined using the Fisher exact test or χ2 analysis with Yates correction.
Mean age and follow-up of the 68 patients was 35.6 years (range, 11.0-63.0 years) and 152 months (range, 4-572 months), respectively. All patients were Japanese (60 female and 8 male). Forty-six patients (68%) received glucocorticosteroid therapy; total dose of prednisolone averaged 34,368.3 mg (16.1 mg/d).
Antibody profiles of the 3 groups (primary SS, SS+LE, and LE alone) are outlined in Table 1. Analysis of the JS-1–GST fusion protein on polyacrylamide gels demonstrated an apparent protein band at 98 kd (Figure 1). Seven serum samples from 9 patients with primary SS (78%) and 9 serum samples from 15 patients with SS+LE (60%) bound to the 98-kd protein, whereas only 3 (7%) of 44 serum samples from patients with LE alone bound to this protein (P<.001 for both, by Fisher exact test). When patients with primary SS and SS+LE were combined and compared with patients with LE alone, the sensitivity of anti–JS-1 antibody was 67%, specificity was 93%, and both positive and negative predictive values were 84% (Table 1). Specificity of the interaction of JS-1 protein with patient serum was confirmed by competitive preincubation tests: when a positive serum sample was preincubated with JS-1–GST or GST, the JS-1–GST fusion protein inhibited the reactivity of the serum, whereas the GST protein alone did not (Figure 1). Analogous results were obtained using a commercially available mouse monoclonal antibody to α-fodrin (results not shown). It was theoretically possible to establish an ELISA for measuring the titer of anti–JS-1 antibody; however, we have yet to obtain reproducible data with an ELISA system.
Although there was no significant difference, the anti–SS-A (Ro) antibody was more frequently detected in patients with primary SS (6/9; P=.10) and SS+LE (8/15; P=.20) than in those with LE alone (16/44). This pattern was almost the same as in anti–JS-1, which excluded the higher percentage of anti–SS-A (Ro) in patients with LE alone. The prevalence of anti–SS-B (La) antibody in these 3 groups was similar to that of anti–JS-1 antibody: 6 (67%) of 9 serum samples from patients with primary SS contained anti–SS-B (La) antibody, whereas 5 (11%) of 44 serum samples from patients with LE bound to this protein (P=.001, Fisher exact test). However, anti–SS-B (La) antibody was present in 5 (33%) of 15 serum samples from patients with SS+LE, indicating a proportion was not significantly higher than that of patients with LE alone (5/44; P=.06).
Correlations between anti–60-kd and anti–52-kd antibodies are shown in Figure 2. Both antibodies occurred together in 24 (80%) of the 30 serum samples that were positive for anti–SS-A (Ro). Serum samples reacting with the 52-kd antigen but not with the 60-kd antigen were found only in patients with primary SS, whereas serum samples reacting with the 60-kd antigen but not with the 52-kd antigen were found only in patients with LE. Two serum samples that were negative for anti–JS-1 from patients with primary SS and 3 serum samples that were positive for anti–JS-1 from patients with LE alone reacted with 60- and 52-kd proteins (Figure 2). In this study, the titers of anti–60-kd and anti–52-kd antibodies failed to show any association with clinical manifestations (data not shown), as reported in previous articles.12,28,29
Because we found that anti–JS-1 antibody was more prevalent in patients with primary SS and SS+LE, we investigated whether any clinical features are associated with this type of autoantibody. Figure 3 presents the summary of a preliminary comparison of positive and negative anti–JS-1 groups. Pernio, hyperglobulinemia, rheumatoid factor positivity, and the presence of anti–SS-B (La) antibody were characteristic of JS-1–positive patients (P<.01). No significant differences could be demonstrated regarding annular erythema, photosensitivity, vasculitis, or renal disorder.
We found that anti–JS-1 was common in patients with primary SS and, to a lesser extent, in those with secondary SS—78% and 60%, respectively, which are similar to findings of a previous study.13 It has also been emphasized that the anti–α-fodrin antibody is specific to SS because it lacks reactivity with serum samples from patients with SLE and RA.13 In contrast, we detected anti–JS-1 antibodies in 3 (7%) of 44 serum samples from patients with LE alone and consequently concluded that anti–JS-1 might not be a single serologic abnormality that is in itself predictive of definite SS. However, we still think that the detection of anti–JS-1 antibody is of great diagnostic value for distinguishing SS from LE because anti–JS-1 antibody was as specifically present in serum samples from patients with SS as was anti–SS-B (La) antibody, and the JS-1–positive serum samples did not necessarily contain anti–SS-B (La), and vice versa (data not shown). All 3 patients who had anti–JS-1 manifested pernio and rheumatoid factor, and 2 of them revealed xerophthalmia, hyperglobulinemia, anti–SS-A (Ro), and anti–SS-B (La) (Figure 3). Although 3 is a small number on which to base a conclusion, these findings agreed with these manifestations being common among JS-1–positive patients. Further work may identify additional patients with LE with anti–α-fodrin antibodies to define a novel clinical entity of LE and to better determine whether there is any practical prognostic significance of the antibodies in patients with LE.
Several investigators have studied the types of specified diseases and clinical manifestations that are associated with the presence of precipitating autoantibodies. There is general agreement that anti–SS-A (Ro) antibodies are highly associated with vasculitis,30,31 hyperglobulinemia,32 hypocomplementemia,31,32 leukopenia,32 and rheumatoid factor positivity32,33 in patients with primary SS and strongly correlated with photosensitive skin damage8,19,30,32- 35 (including annular skin lesion), interstitial pneumonitis,36 and rheumatoid factor positivity33,35 in patients with LE. With a few exceptions, similar tendencies are observed in patients with anti–SS-B (La).4,8 We could not perform a comparative analysis between patients who were positive and those who were negative for anti–JS-1 because their numbers were too small to evaluate. The preliminary results shown in Figure 3 are consistent with the finding that antibodies to JS-1 are more closely related to SS than to LE. We found that pernio, hyperglobulinemia, and rheumatoid factor positivity were unequally distributed when patients were divided according to serum reactivity with the JS-1 antigen. These statistical differences were significant because they were frequently present in most patients with SS who were positive for anti–JS-1. The frequency of pernio is higher in patients with SS, as are the 2 serum abnormalities that might cause hyperviscosity and induce impairment of peripheral microcirculation. Rheumatoid factor was once a candidate when criteria for SS were being validated.2 All of these results support the idea that occurrence of the anti–JS-1 antibody is probably not a mere epiphenomenon but rather, essential to the pathogenesis of these clinical and laboratory findings.
Results of a few large studies32,37,38 show that serum samples from patients with SLE and SS reacted with the 60- and 52-kd components of SS-A (Ro) proteins. Antibodies to the 52-kd antigen without concomitant antibodies to the 60-kd antigen were seen only in patients with SS, whereas antibodies to the 60-kd antigen without concomitant antibodies to the 52-kd antigen were detected only in patients with SLE. These patterns were again confirmed in our study. Two serum samples from patients with SS did not react with JS-1 antigen; in both, the level of antibody binding to the 52-kd protein was relatively low (Figure 2). However, 2 of 3 serum samples from patients with LE and anti–JS-1 antibodies showed higher levels of anti–52-kd antibodies. We speculate that there is a close serologic linkage between antibody production to JS-1 and to 52-kd SS-A (Ro).
In 1979, Sontheimer et al19 divided LE-specific skin lesions into 3 major categories (chronic, subacute, and acute) and proposed a new concept, "subacute cutaneous LE," from the dermatologic perspective. Annular-polycyclic and papulosquamous skin lesions with LE-specific histopathologic features are the only required findings for the diagnosis of subacute cutaneous LE. Later, annular erythema resembling the subacute lesion was frequently observed in Asian patients with SS, and anti–SS-A (Ro) antibodies have been detected in 50% to 60% of serum samples from these patients, the percentage of which is similar to that in subacute cutaneous LE.9- 11,33- 35 There has been a close linkage between annular erythema and anti–SS-A (Ro) antibodies, but we did not find any association of annular lesions with JS-1 antibody production (Figure 3). Therefore, we speculated that a disease-specific expression of α-fodrin could be observed at the eccrine sweat duct and may play a crucial role during the development of SS-relevant skin lesions because it has been reported13 that 120-kd α-fodrin is intensely expressed on the epithelial duct of inflamed salivary glands. We performed immunohistochemical staining of several skin lesions with 2 different monoclonal anti–α-fodrin antibodies, but we observed no significant immunoreactivity in skin lesions with annular erythema or in those with pernio (results not shown). Taking these findings together, α-fodrin seems to not be directly involved in the pathogenesis of annular erythema. As previously suggested,39 SS-A (Ro) antigen expression on the surface of keratinocytes via ultraviolet irradiation and subsequent antigen-autoantibody reaction may be important.
Accepted for publication November 5, 1998.
We thank Randall T. Moon, PhD, for providing the JS-1 complementary DNA, and Yuriko Itoh, MD, and Yoshitsugu Ueda, MD, for their helpful discussion.
Reprints: Takahiro Watanabe, MD, PhD, Department of Dermatology, Faculty of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan (e-mail: firstname.lastname@example.org).