Journal of Autoimmunity (199 1) 4,87-96

Detection of Anti-Ro( SSA) Antibodies by Gel Double Diffusion and a ‘Sandwich’ ELISA in Systemic and Subacute Cutaneous Lupus Erythematosus and Sjiigren’s Syndrome

Thomas T. Provost,* L. Stefan Levin,t$ Rosemarie M. Watson,* Marvin Mayo* and Harry Ratrie III* *Department of Dermatology, TheJohns Hopkins University iDivision of Dermatopathology, Department of Dermatology, University

School of Medicine;

Baltimore,

School of Medicine; TheJohns Hopkins

Maryland,

USA

A newly described Ro ‘sandwich’ ELISA was compared to the gel double diffusion technique to detect anti-Ro(SSA) antibodies in Sjiigren’s syndrome, systemic and subacute cutaneous lupus erythematosus patients. This study demonstrates that the ELISA assay increased the frequency of detection of anti-Ro(SSA) antibodies in these well defined connective tissue disease patients by approximately S-10% compared to the gel double diffusion anti-Ro(SSA) antibody assay. The study also confirms that some patients make anti-Ro(SSA) antibodies directed solely at unique human Ro(SSA) antigen epitopes. We also detected the existence of a significant Sjogren’s syndrome patient population failing to make significant antiRo(SSA) antibodies. We conclude from our study that the gel double-diffusion technique employing human spleen extract as a source of the Ro(SSA) antigen is, at present, the most cost-effective test to detect anti-Ro(SSA) antibodies

Introduction

Anti-Ro( SSA) antibody determinations are now routinely employed in the investigation of patients suspected of having connective tissue diseases. The anti-Ro( SSA) antibody is directed against a polypeptide complex composed of at least two (52 kD and 60 kD) and possibly more polypeptides [ 11. Binding to these polypeptides are a unique species of RNAs (hY 1-hY5) [2]. Previous work in our and other laboratories IL. Stefan Levin deceased. 87 0896-8411/91/010087+

10 $03.00/O

0 1991 Academic Press Limited

88

T. T. Provost et al.

using gel double diffusion assays has indicated that anti-Ro(SSA) antibodies are present in approximately 40% of patients with Sjogren’s syndrome, approximately 70% of patients with subacute cutaneous lupus erythematosus, approximately 70% of patients with antinuclear antibody (ANA)-negative systemic lupus erythematosus, approximately 75% of homozygous C2- and C4-deficient patients with a lupus-like syndrome, and approximately 95% of mothers giving birth to children with the neonatal lupus syndrome (reviewed in [3]). The present study was designed to compare the gel double diffusion technique with a Ro(SSA) ‘sandwich’ ELISA to detect anti-Ro(SSA) antibodies amongst classic antinuclear antibody-positive systemic lupus erythematosus, Sjogren’s syndrome and subacute cutaneous lupus erythematosus patients. These studies indicate that the gel double diffusion technique employing human spleen extract as a source of the Ro(SSA) antigen is the most cost-effective technique now available for anti-Ro(SSA) antibody detection. These studies confirm the superiority of human Ro( SSA) macromolecules compared to a bovine source of Ro( SSA) macromolecules in detecting anti-Ro( S SA) antibodies. Furthermore, these studies demonstrate that there is a significant number of Sjogren’s syndrome patients who fail to demonstrate abnormal quantities of anti-Ro(SSA) antibodies.

Materials and methods Patient population Sjbgren’s syndrome (SS) patient sera were obtained on 39 consecutive SS patients whose minor salivary gland biopsies demonstrated features consistent with the diagnosis of Sjogren’s syndrome (Class III and Class IV biopsies using Greenspan’s and Tarpley’s classification [4, 51). All but one of these 39 Sjdgren’s syndrome patients were seen and examined by one of the authors at the time of lip biopsy. All were symptomatic for the sicca complex. No attempt was made to further classify these patients. Systemic lupus erythematosus (SLE) patients’ sera were obtained from our serum bank on 33 consecutive sera, which upon analysis were found to possess anti-nDNA antibodies (Crithidia Zuciliaeassay), anti-Sm and/or anti-nRNP (U,RNI?) antibodies (gel double diffusion). The presence of anti-Ro(SSA) antibodies, as determined by gel double diffusion, did not influence the selection of these patients’ sera. The sera of subacute cutaneous lupus erythematosus (SCLE) patients were obtained from 31 consecutive SCLE patients selected solely on the basis of morphologic features of either annular polycyclic or papulosquamous lesions associated with SCLE [6]. Sixteen of these SCLE patients had the papulosquamous variety, and 15 had the annular polycyclic variety of subacute cutaneous lupus erythematosus lesions. Serologic testing as part of the serologic evaluation All sera employed in this study were tested for antinuclear antibodies employing HEp-2 cells using standard immunofluorescent techniques. Anti-native DNA antibodies were detected employing the Crithidia luciliae assay [7]. Antiribonuclear protein (nRNP; UiRNP), anti-Sm, anti-Ro(SSA), and anti-La(SSB) antibodies

Detection of Anti-Ro(SSA)

Antibodies

89

were detected utilizing 0.6% agarose gel double-diffusion techniques employing human spleen extract as a source of the Ro antigen and rabbit thymic extract as a source of La(SSB) nRNl? (U,RNP) and SM antigen, as previously described [8]. In addition, selected sera were tested by gel double-diffusion techniques for antiRo(SSA) antibodies utilizing bovine spleen extracts as a source of the Ro(SSA) antigen. Sera were tested for anti-Ro(SSA) antibodies utilizing a modification of a ‘sandwich’ ELISA technique recently described by Reichlin et al. [9]. Pooled IgG (void-volume of a DE-52 column) containing IgG anti-Ro(SSA) antibodies from five lupus erythematosus patients (reactive by immunoblotting using human spleen extract against the 52 kD and the 60 kD Ro(SSA) associated peptides) was subjected to pepsin digestion. (Four hundred microlitres of 1 M acetate buffer (pH 4.15) was added to 200 mg of protein (2 ml pool of IgG) + 4 mg of pepsin and incubated at 37°C for 18 h). The digestion was stopped by adjusting the pH to 7.5 with solid trizma base. The F(ab’), fragments were separated from undigested IgG by Sephacryl-200 filtration. The F(ab’), fragments of anti-Ro(SSA) antibodies were diluted in 0.05 M sodium carbonate-bicarbonate buffer pH 9.2 and and pipetted into 96 well micro assay plates (Linbro/Titertek) at 0.25 pg/50 ~1. The plates were incubated overnight at 4°C. Unbound sites were blocked with PBS-l y0 bovine serum albumin (1 y0 BSA) pH 7.2 for 30 min at room temperature. A working dilution of human spleen extract in PBS pH 7.2 containing the Ro(SSA) antigen was added to duplicate wells. To check for non-specific binding of the test sera to the F(ab)‘, fragment, PBS pH 7.2 was also added to duplicate wells. A 1:2,000 dilution of test serum in PBS-1 y0 Tween 20-l “/o BSA pH 7.2 was added to the appropriate wells. The secondary antibody was horseradish peroxidase-labelled IgG fraction of goat antihuman IgG (FC specific) (Cappel, Westchester, Pa) diluted in PBS-1 % BSA pH 7.2. Fifty pi/well of 2,2’azino-di[3’-ethyl-benzothiazoline sulfonate) [6] (ABTS) was employed as the colour developer and the OD was read at 405 nm using an automated plate reader. Washings between steps were performed with PBS-0.5% Tween-20, pH 7.2 in an automated plate washer. Except where noted, incubations were for one hour at room temperature. Analysis

The determination of antibody status of test serum in the ‘sandwich’ ELISA for anti-Ro(SSA) antibodies was calculated as follows: A panel of 10 normal human sera (NHS) was assayed for each test run. The mean OD value and standard deviation was calculated for the 10 NHS. To determine the number of standard deviations of the test sera OD from the mean OD of NHS, the mean OD of the NHS was subtracted from the test serum OD value. This value was then divided by one standard deviation of the OD of the NHS. Those sera greater than two standard deviations above the mean of the NHS were considered positive, and those less than two standard deviations were considered negative. In addition, the same calculations were made for the non-specific binding wells. Those test sera exhibiting OD levels greater than the NHS mean plus two standard deviations were corrected for non-specific binding.

90

T. T. Provost et al. Tgble 1. SLE wa

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

Precipitin antibodies

HSE’ Ro(SSA) ‘sandwich’ ELISA (standard deviations)

nRNP very weak Ro, uni, La Ro, nRNP nRNP Ro, La Ro, nRNP uni, nRNP, Sm Ro nRNP nRNP Ro, Sm, nRNP uni, Sm, nRNP uni, Sm, nRNP Ro, nRNP uni, Sm, nRNP Ro Ro, nRNP uni, Sm, nRNP nRNP uni, Sm, nRNP uni, Sm, nRNP uni Ro, uni, Sm, nRNP Ro, uni, Sm, nRNP Ro, uni, Sm, nRNP Ro, Sm, nRNP

2.3 3.2 6.3 0.6 13.6 2.5 0.3 12.2 -0.5 -0.1 -0.3 0.5 15.1 1.1 0.5 -0.4 0.0 8.8 0.2 3.4 8.0 -0.3 0.4 -0.6 -0.7 -0.1 -0.8 -0.4 -0.2 0.8 7.9 6.7 8.8

‘HSE Humanspleenextract. uni-unidentified precipitinantibody.

Results Comparison between the frequency of anti-Ro(SSA) antibody positively between gel double diffusion and the ELISA in systemic lupus erythematosus, SjGgren’s syndrome and subacute cutaneous lupus erythematosus patients is made in Tables 1,2, and 3. In this cohort of systemic lupus erythematosus patient population, 13 of 33 (39.4%) demonstrated anti-Ro(SSA) antibody positively by both gel double diffusion and ELISA. However, in two instances, patients #l and $30, there was a discrepancy between the Ro gel double diffusion and Ro sandwich ELISA.

Detection of Anti-Ro(SSA)

Antibodies

91

Table 2. Sjiigren’s syndrome

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39.

Precipitin antibodies

HSE’ Ro(SSA) ‘sandwich’ ELISA (standard deviations)

Ro Ro Ro Ro, La Ro Ro, La Ro, La Ro La

19.0 28.8 0.5 2.0 30.3 42.3 -0.5 -1.7 -0.1 -0.6 -1.1 0.5 -1.3 0.1 15.5 -0.3 0.1 0.2 0.8 -0.6 1.4 0.3 1.0 8.2 -0.1 -0.1 19.7 16.8 1.0 2.1 1.1 0.9 -0.8 8.3 -0.5 -0.2 7.7 0.3 15.4

‘HSE Human spleen extract.

In Table 2 the results of the comparison between gel double diffusion and ELISA for anti-Ro(SSA) antibody detection in Sjiigren’s syndrome patients are reported. In this group of 39 patients who were symptomatic for the sicca complex and who demonstrated a lip biopsy consistent with, if not diagnostic of, Sjogren’s syndrome, eight were anti-Ro( S SA) antibody positive by gel double diffusion (20.5 %), whereas

92

T. T. Provost et al. Table 3. Subacute cutaneous lupus erythematosus

HSE’ Ro(SSA) HSE’ precipitin

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

antibodies

BSE2 precipitin antibodies

‘sandwich ELISA (standard deviations)

Ro -

Ro -

Ro -

Ro -

24.9 0.4 2.0 -0.3 21.8 9.7 8.3 8.8 0.3 1.5 6.7 19.5 34.6 46.2 20.1 1.3 0.5 34.3 0.5 1.7 1.8 20.1 23.9 11.7 53.4 4.4 22.2 63.0 11.3 3.0 71.5

Ro Ro -

Ro -

-

Ro Ro Ro Ro -

Ro Ro Ro Ro -

Ro -

Ro -

-

-

Ro Ro Ro Ro Ro Ro Ro Ro Uni3 Ro

Ro Ro Ro Ro Ro Ro Ro Ro Uni Ro

‘HSE, human spleenextract. 2BSE, bovinespleenextract. Wni, unidentifiedprecipitin antibody.

12 of 39 (30.7%) were positive by ELISA. Note that patient #39 demonstrated antiLa( SSB) antibody activity in the apparent absence of anti-Ro(SSA) antibody activity by gel double diffusion but demonstrated a very significant anti-Ro( SSA) antibody response by ELISA (15.4 standard deviations). Table 3 shows that 18 of 31 subacute cutaneous lupus erythematosus patients (59%) were anti-Ro(SSA) antibody positive by gel double diffusion employing human spleen extract as the antigen source, whereas 22 of 31 were anti-Ro(SSA) antibody positive by the ELISA method (70.9%). In addition, this table demonstrates a comparison between positivity of these two assays using human versus

Detection of Anti-Ro(SSA)

Antibodies

93

60

BSE

HSE

Figure 1. Ro ‘sandwich’ ELISA data on 31 subacute cutaneous lupus erythematosus (SCLE) patients. The results are reported in standard deviations (see text). Bovine spleen extract (BSE) and human spleen extract (HSE) were employed to charge the microtiter plates and the system optimized to produce maximum sensitivity and the least non-specific background. Note that in all instances the human Ro ‘sandwich’ ELISA demonstrated higher titers than the bovine Ro ‘sandwich’ ELISA. Furthermore, at least four SCLE patients failing to demonstrate abnormal quantities of anti-Ro(SSA) antibodies in the bovine Ro ‘sandwich’ ELISA demonstrated abnormal quantities in the human Ro ‘sandwich’ ELISA.

bovine spleen sources of the Ro antigen. One patient (#7), detected as being antiRo( SSA) positive by gel double diffusion employing human spleen extract was not detected as being positive when bovine Ro was employed in the gel double diffusion assay. (These studies were repeated several times.) In addition, in Figure 1, one can see a comparison of the anti-Ro(SSA) antibody titers in the ‘sandwich’ ELISA using human spleen extract versus bovine spleen extract to ‘charge’ the microtiter wells. You will note that several sera positive in the human Ro(SSA) ELISA failed to demonstrate significant anti-Ro(SSA) antibody titers in the bovine Ro(SSA) ‘sandwich’ ELISA. Discussion

This study demonstrates that 39 of 103 (38%) ofwell-characterized connective tissue disease patients (systemic lupus erythematosus, Sjogren’s syndrome and subacute cutaneous lupus erythematosus) are anti-Ro(SSA) antibody positive employing gel double-diffusion techniques. By comparison, 47 of these 103 connective tissue disease patients were positive for anti-Ro(SSA) antibody (45.6%) by ELISA. This study compares well with another recently published study comparing anti-Ro( SSA)

94

T. T. Provost et al.

antibody positivity by gel double diffusion and ELISA in systemic lupus erythematosus [lo]. These studies collectively indicate that the ELI SA assay probably detects an additional 5-1Oo/oof anti-Ro(SSA) positivity compared to the gel double diffusion technique. Obviously, the most important reason for this discrepancy between the two assays is the increased sensitivity of the ELISA assay. Indeed, patient $1 in Table 1, patients #4 and #30 in Table 2, and patients #3 and #30 in Table 3 support the contention that this ELISA assay will detect very low levels of abnormal quantities of anti-Ro(SSA) antibody activity. However, the data reported in these tables also indicate that the ELISA assay detects non-precipitating antibodies (i.e., conceptually antibodies directed against a restricted number of epitopes on the antigen, resulting in poor lattice formation and thus making poor precipitating antibodies). In Table 2, patients #24, #37 and #39, and in Table 3 patients $6 and $11 make relatively significant quantities of autoantibodies as determined by the ELI SA assay, and yet the anti-Ro(SSA) gel double-diffusion technique on repeated occasions failed to demonstrate the presence of anti-Ro(SSA) antibodies. There is also preliminary data that on unusual occasions the gel double diffusion assay may be more sensitive in detecting anti-Ro(SSA) antibodies than the ELISA (Patient #30, Table 1). Thus far, we have detected one other patient, a La( SSB) positive mother of a neonatal lupus infant who demonstrated anti-Ro(SSA) antibodies by gel double diffusion, but failed to demonstrate anti-Ro(SSA) antibodies in our ELISA. It is conceivable this is a technical problem, but it could also indicate that the patient may possess an anti-Ro(SSA) antibody detected against an additional polypeptide not represented in our sandwich ELISA. [Our Ro ‘sandwich’ ELISA contains antibodies reactive by immunoblotting against 52 kD and 60 kD polypeptides.] Our studies indicate the ELISA does have utility in detecting low-titer antiRo(SSA) antibody activity as well as detecting those individuals who make significant quantities of non-precipitating antibodies. The ELISA studies employing a human source of Ro( SSA) antigen have played a very valuable role in demonstrating a link between anti-La(SSB) antibody activity and the presence of anti-Ro(SSA) antibody activity. It has been a controversial point whether or not anti-La(SSB) antibody activity can occur in the absence of anti-Ro(SSA) antibody activity. In many previous anti-Ro(SSA) antibody studies, bovine spleen extract has been the source of the Ro(SSA) antigen. Reichlin et al. has demonstrated that in contradistinction to most other autoantibodies described in connective tissue disease patients, the anti-Ro( S SA) antibody activity is not directed against highly-conserved epitopes but rather against epitopes unique to the human species [9-l 11. We believe this fact explains the previous studies demonstrating anti-La( S SB) antibodies in the apparent absence of anti-Ro(SSA) antibodies. In fact, neither Reichlin and Harley, nor our group, have ever detected an anti-La(SSB) serum failing to demonstrate antiRo(SSA) antibodies when either human spleen has been used as a source of the Ro antigen or the serum was examined by the Ro ‘sandwich’ ELISA [ 121. Based upon our experience with the ELISA and gel double diffusion assays, we believe that gel double diffusion techniques for the determination of anti-Ro( SSA) antibody activity which utilize human tissue as a source of the antigen, despite their inherent deficiency in sensitivity, will remain, for the present time, the standard technique for the detection of anti-Ro( SSA) antibody activity in the clinical setting of

Detection

of Anti-Ro(SSA)

Antibodies

95

patient evaluation. Although there is increased sensitivity with the ELISA assay, the cost of instrumentation, as well as technician time, makes the gel double diffusion assay at the present time more cost effective. This study also is important because it demonstrates the existence of a significant Sjiigren’s syndrome patient population who, despite the employment of a very sensitive Ro ‘sandwich’ ELISA capable of detecting antibodies directed against the 52 kD, as well as the 60 kD Ro(SSA) polypeptides, failed to demonstrate abnormal quantities of anti-Ro(SSA) antibodies. This result is contrary to a previous study we have performed in which virtually 100% of SS patients demonstrated anti-Ro(SSA) antibodies [13]. We believe there are two major reasons for the apparent discrepancies between our initial study and this study. First many of the Sj(igren’s syndrome patients evaluated in the current study did not have extraglandular manifestations of Sjiigren’s syndrome. They were patients who had predominantly a sicca complex. The frequency of vasculitis, the coexistence of subacute cutaneous lupus erythematosus, as well as neuropsychiatric manifestations appeared to be markedly less in this group of patients compared to our original study. In contrast to the original study, which was composed of patients followed over a prolonged period of time by a rheumatology service, the patients in the current study were predominantly newly diagnosed Sjijgren’s syndrome patients referred to our clinic by physicians representing many different specialties including allergy (for the evaluation of non-allergic sinusitis), dermatologists (for the evaluation of xerosis and sicca symptoms), from gynecologists (for evaluation of vaginal dryness), etc. The second reason for the apparent discrepancy is a statistical one dealing with the data. In the original study, the emphasis was upon detecting low-grade antiRo(SSA) antibody activity in these patients’ sera. The results were presented in absolute terms, expressed as the log of OD reading at a lob7 dilution of the test serum. Using this statistical analysis, four of the 40 normal control sera demonstrated low titers of anti-Ro(SSA) antibodies. Approximately 60% of the SS patients in the original study also demonstrated low levels of anti-Ro(SSA) antibody activity comparable to the four normal control sera. We suspect that all the low-grade anti-Ro(SSA) antibody activity detected in the original study would have fallen within two standard deviations of the normal range using the methods employed in this study and thus would not be considered by us as being abnormal values. [Unfortunately, none of the original sera are available for a comparison study.] The presence of anti-Ro(SSA) antibodies in Sjijgren’s syndrome (SS) is an important issue. We have detected a group of anti-Ro(SSA) antibody-negative SS patients who have features of multiple sclerosis (14). They have a central nervous system disease process characterized by involvement of the entire neural axis, exacerbations and remissions over a prolonged period of time. Sixty-five percent of these S S patients with features of multiple sclerosis in our original study were anti-Ro(SSA) antibody negative [ 141. Regrettably, one recent study of multiple sclerosis patients in which approximately 20% of their patient population had features of the sicca complex (dry eyes and/or dry mouth) failed to perform routine diagnostic evaluation to rule out a concomitant Sjdgren’s syndrome, based on the erroneous assumption that a negative anti-Ro( SSA) antibody determination eliminated the possibility of Sjijgren’s syndrome [ 151. The point to be emphasized and demonstrated in this study is that there

96

T. T. Provost et al.

are a significant number of patients with bonafide Sjiigren’s syndrome who are truly seronegative with regard to the presence of anti-Ro(SSA) antibodies. Acknowledgements This study was supported by grants #l ROl DE08570 and #5 ROl AR25650 the National Institutes of Health and a gift from the Noxell Corporation.

from

References 1. Ben-Chetrit, E., E. K. L. Chan, K. Sullivan, and E. M. Tan. 1988. A 52 kD protein is a novel component of the SSA/Ro antigenic particle.3. Exp. Med. 167: 1560-1571 2. Lerner, M. R. and J. A. Steitz. 1979. Antibodies to small nuclear RNAs complexed with proteins are produced by patients with systemic lupus erythematosus. Proc. Natl. Acad. Sci. (USA) 76: 5495-5497 3. Provost, T. T. 1988. The clinical significance of Ro(SSA) and La(SSB) antibodies in lupus erythematosus. Seminars in Dermatology 7: 130-139 4. Tarpley, T. M., L. G. Anderson, and C. L. White. 1974. Minor salivary gland involvement in Sjiigren’s syndrome. Oral Surg. 37: 64-74 5. Greenspan, J. S., T. E. Daniels, N. Talal, and R. A. Sylvester. 1974. The histopathology of SjBgren’s syndrome in labial salivary gland biopsies. Oral Pathol. 37: 217-219 6. Gilliam, J. N., and R. D. Sontheimer. 1981. Distinctive cutaneous subsets in the spectrum of lupus erythematosus. 3. Am. Acad. Dermatol. 4: 471-475 7. Aarden, L. A., E. R. deGroot, and T. E. Feltkamp. 1974. Immunology of DNA III Crithidia luciliae, a simple substrate for the determination of anti-dsDNA with the immunofluorescence technique. Ann N. Y. Acad. Sci. 254: 505-515 8. Madison, P. J., T. T. Provost, and M. Reichlin. 1981. Serological findings in patients with ‘ANA-negative’ SLE. Medicine 60: 87-94 9. Reichlin, M., M. Rader, and J. B. Harley. 1989. Autoimmune response to Ro/SSA particle is directed to the human antigen. Clin. Exp. Immunol. 76: 373-377 10. Hamilton, R. G., J. B. Harley, W. B. Bias, M. Roebber, M. Reichlin, M. C. Hochberg, and F. C. Arnett. 1988. Two Ro(SSA) autoantibody responses in systemic lupus erythematosus: Correlation of HLA-DR/DQ specificities with quantitative expression of Ro(SSA) autoantibody. Arthritis Rheum. 31: 496-505 11. Reichlin, M., and M. W. Reichlin. 1989. Autoantibodies to the Ro/SS-A particle react preferentially with the human antigen. SjGgren’s syndrome: A model for understanding autoimmunity. N. Talal, ed. Academic Press. pp. 51-57 12. Harley, J. B., A. L. Sestak, L. G. Willis, S. H. Fu, J. H. Hansen, and M. Reichlin. 1989. A model for disease heterogeneity in systemic lupus erythematosus: Relationships between histocompatibility antigens, autoantibodies and lymphopenia or renal disease. Arthritis Rheum. 32: 826-836 13. Harley, J. B., E. L. Alexander, W. B. Bias, 0. F. Fox, T. T. Provost, M. Reichlin, H. Yamagata, and F. C. Arnett. 1986. Anti-Ro(SSA) and anti-La(SSB) in Sjiigren’s syndrome. Arthritis Rheum. 29: 196-206 14. Alexander, E. L., K. Malinow, J. E. Lejewski, T. T. Provost, and G. E. Alexander. 1986. Primary Sjogren’s syndrome with central nervous system dysfunction mimicking multiple sclerosis. Ann. Int. Med. 104: 323-330 15. Noseworthy, J. H., B. H. Bass, M. K. Vanderwoort, G. C. Ebers, G. I’. A. Rice, B. G. Winsteulcer, C. J. L. McLay, and D. A. Bell. 1989. The Prevalence of primary Sjiigren’s syndrome in a multiple sclerosis population. Ann. Neural. 25: 95-98