A critical evaluation of enzyme immunoassays for detection of antinuclear autoantibodies of defined specificities. I. Precision, sensitivity, and specificity.

OBJECTIVE: To determine the performance characteristics of enzyme-based immunoassay (EIA) kits for the detection of antinuclear and other autoantibodies of defined specificities. METHODS: Nine manufacturers of EIA kits to detect antibodies of defined specificities participated in a study in which they received coded sera from the Centers for Disease Control and Prevention. These coded sera contained different dilutions of antibody of one specificity mixed with sera containing antibodies of other specificities. The manufacturers were asked to use their standard technology to determine antibody content and send the data to a committee of the International Union of Immunological Societies for analysis. The data were analyzed for sensitivity and specificity in the detection of anti-double-stranded DNA (anti-dsDNA), anti-single-stranded DNA, antihistone, anti-Sm, anti-U1 RNP, anti-SSA/Ro, anti-SSB/La, anti-Scl-70 (DNA topoisomerase I), anticentromere, and anti-Jo-1 antibodies. In addition, replicate samples were included in the coded sera to evaluate the precision of each EIA method. RESULTS: Lack of sensitivity and specificity was most evident in the anti-dsDNA and anti-Sm kits, although 2 kits for anti-dsDNA achieved acceptable sensitivity and specificity. Generally, anti-SSA/Ro, anti-SSB/La, anti-Scl-70, anticentromere, and anti-Jo-1 kits performed well. Many false-positive results were obtained with a multiple myeloma serum containing cryoprecipitates, but multiple myeloma sera without cryoprecipitates presented no problem in the EIA system. Precision, based on evaluation of replicate samples, varied from very good to poor. CONCLUSION: No single manufacturer was clearly superior to others in terms of their products' overall sensitivity, specificity, and precision. Areas that needed improvement were in kits for the detection of antibodies to dsDNA and to Sm antigen. Some EIA kits achieved good sensitivity and specificity. Individual manufacturers were informed of the performance of their respective kits so they could take measures to correct perceived deficiencies and thus improve the reliability of a group of important diagnostic assays used in the evaluation of systemic rheumatic diseases.

Range of antinuclear antibodies in "healthy" individuals.

OBJECTIVE: To determine the range of antinuclear antibodies (ANA) in "healthy" individuals compared with that in patients with systemic lupus erythematosus (SLE), systemic sclerosis (SSc; scleroderma), Sjögren's syndrome (SS), rheumatoid arthritis (RA), or soft tissue rheumatism (STR). METHODS: Fifteen international laboratories experienced in performing tests for ANA by indirect immunofluorescence participated in analyzing coded sera from healthy individuals and from patients in the 5 different disease groups described above. Except for the stipulation that HEp-2 cells should be used as substrate, each laboratory used its own in-house methodology so that the data might be expected to reflect the output of a cross-section of worldwide ANA reference laboratories. The sera were analyzed at 4 dilutions: 1:40, 1:80, 1:160, and 1:320. RESULTS: In healthy individuals, the frequency of ANA did not differ significantly across the 4 age subgroups spanning 20-60 years of age. This putatively normal population was ANA positive in 31.7% of individuals at 1:40 serum dilution, 13.3% at 1:80, 5.0% at 1:160, and 3.3% at 1:320. In comparison with the findings among the disease groups, a low cutoff point at 1:40 serum dilution (high sensitivity, low specificity) could have diagnostic value, since it would classify virtually all patients with SLE, SSc, or SS as ANA positive. Conversely, a high positive cutoff at 1:160 serum dilution (high specificity, low sensitivity) would be useful to confirm the presence of disease in only a portion of cases, but would be likely to exclude 95% of normal individuals. CONCLUSION: It is recommended that laboratories performing immunofluorescent ANA tests should report results at both the 1:40 and 1:160 dilutions, and should supply information on the percentage of normal individuals who are positive at these dilutions. A low-titer ANA is not necessarily insignificant and might depend on at least 4 specific factors. ANA assays can be a useful discriminant in recognizing certain disease conditions, but can create misunderstanding when the limitations are not fully appreciated.

Reference sera for antinuclear antibodies. II. Further definition of antibody specificities in international antinuclear antibody reference sera by immunofluorescence and western blotting.

OBJECTIVE: To define the fine specificity of the 10 reference sera used for determination of antinuclear antibodies (ANA) and ANA subsets which are available from the Arthritis Foundation (AF) and from the Centers for Disease Control and Prevention (CDC). METHODS: AF/CDC sera were assessed by experienced laboratory staff, using indirect immunofluorescence and Western blotting. RESULTS: The original assignment of fluorescence patterns to 4 reference sera was confirmed, and the fluorescence intensities were determined using reference fluorescent beads. On Western blots, sera AF/CDC2 (anti-SS-B/La) and AF/CDC7 (anti-SS-A/Ro) did not detect Ro antigens, sera AF/CDC9 and AF/CDC10 appeared to be monospecific anti-Scl-70 and anti-Jo-1 sera, respectively, serum AF/CDC4 (anti-U1 small nuclear RNP) recognized the 70-kd band, and serum AF/CDC5 recognized the Sm antigen with its multiple bands. Semiquantitative analyses revealed that AF/CDC5, AF/CDC2, and AF/CDC10 were strongly reactive sera, whereas AF/CDC4 and AF/CDC9 were much weaker and should be used at lower dilutions on Western blots. CONCLUSION: The AF/CDC ANA reference sera, originally described as reference reagents for indirect immunofluorescence and double immunodiffusion techniques, are also useful for Western blotting. The data presented herein further support the use of these sera for reference purposes.

Localization of histidyl-tRNA synthetase (Jo-1) in human laryngeal epithelial carcinoma cell line (HEp-2 cells).

The present study was designed to determine the subcellular localization of histidyl-tRNA synthetase (Jo-1) in human laryngeal epithelial carcinoma cell line (HEp-2 cells). Indirect immunofluorescence using commercial HEp-2 cells with human serum and human-affinity-purified anti-Jo-1 antibodies was performed using confocal microscopy. Anti-histidyl-tRNA-synthetase-positive sera showed distinct nuclear and cytoplasmic granular staining in HEp-2 cells. Affinity purified anti-Jo-1 produced an identical pattern to the whole serum, whereas the serum fraction that did not bind to the affinity column was negative by immunofluorescence on HEp-2 cells. Two commercial human anti-Jo-1-positive control sera and seven anti-Jo-1-positive sera from patients with myositis reproduced the nuclear and cytoplasmic granular pattern. We conclude that Jo-1 is present in cytoplasm and in intact nuclei from HEp-2 cells. The presence of tRNA synthetases in intact nuclei suggests that they have an unsuspected function in the nucleus.

Chiasmal herniation as a complication of bromocriptine therapy.

INTRODUCTION: Medical treatment of macroprolactinomas with dopamine agonists decreases tumor mass and improves visual defects. We report an unusual complication of a macroprolactinoma responding to bromocriptine: a visual field defect caused by downward herniation of the optic chiasm. MATERIALS AND METHODS: A 64-year-old woman was found to have a 4.5 cm macroprolactinoma with superior displacement of the optic chiasm, bitemporal hemianopia, and serum prolactin concentration (P) of 17,060 micrograms/L. Bromocriptine was initiated at 2.5 mg/day and increased to 7.5 mg/day over 2 months. RESULTS: After 2 months, visual fields improved significantly and tumor height decreased to 3 cm with resolution of the optic chiasm displacement. P decreased to 1,180 micrograms/L. After 5 months of therapy, visual fields were normal, and P was 8 micrograms/L. After 8 months of therapy, new bilateral visual defects were observed. Magnetic resonance imaging (MRI) revealed further decrease of the tumor height to 1.5 cm, and inferior and leftward traction of the optic chiasm as the probable mechanism for the new visual field deficit. P was < 1 microgram/L. Bromocriptine was decreased to 5 mg/day to allow reduced traction on the optic chiasm and its blood supply. Over the next 4 months, visual field abnormalities resolved. CONCLUSIONS: We report the development of a visual field abnormally that is explained by chiasmal herniation caused by a shrinking macroprolactinoma. This complication resolved with a decrease in the bromocriptine dose. We suggest that patients undergoing bromocriptine therapy for macroprolactinomas be followed for this potential complication.

Hypergammaglobulinemic purpura in systemic autoimmune rheumatic diseases: predictive value of anti-Ro(SSA) and anti-La(SSB) antibodies and treatment with indomethacin and hydroxychloroquine.

OBJECTIVE: To define the clinical manifestations, autoantibody associations, optimal treatment, and prognosis of hypergammaglobulinemic purpura associated with systemic autoimmune rheumatic diseases. METHODS: Of 303 consecutive patients with systemic autoimmune rheumatic diseases evaluated over 5 years, 17 French Canadian patients with hypergammaglobulinemic purpura with systemic lupus erythematosus (SLE) (n = 12) or another systemic autoimmune rheumatic disease (n = 5) were identified and followed prospectively. Mild secondary Sjörgren's syndrome developed in 9 (53%) patients. RESULTS: Sixteen (94.1%) patients were women. Attacks of hypergammaglobulinemic purpura occurred in the pretibial (76.5%) or perimalleolar (70.5%) areas or the dorsal aspect of the feet (52.9%). Triggering factors included walking, prolonged standing, and alcohol intake. The mean duration of attacks was 6.1 days. Systemic manifestations consistent with a flare of the underlying systemic autoimmune rheumatic diseases accompanied hypergammaglobulinemic purpura attacks in 15 (88%) patients. Arthralgias (n = 13, 86.6%), arthritis (n = 9, 69.2%), and periarthritis were characterstically localized adjacent to the purpura. Anti-Ro antibodies were expressed in all (100%) patients with hypergammaglobulinemic purpura with SLE, but in only 11 (28.9%) of 38 consecutive patients with SLE without hypergammaglobulinemic purpura (P < 0.000001, odds ratio 84, 95% confidence interval 4.6, 1525). The positive predictive values for hypergammaglobulinemic purpura in SLE were: anti-Ro plus anti-La 73%, anti-La 57%, and anti-Ro 52%. The negative predictive value of anti-Ro was 100%. Although 11 (92%) patients with SLE with anti-Ro expressed anti-52 kDa Ro [4 (36.3%) of whom also expressed anti-60 kDa Ro], this frequency was not greater than in anti-Ro positive patients with SLE without hypergammaglobulinemic purpura. The effects of indomethacin or hydroxychloroquine were assessed over 6 months in 8 patients with recurrent incapacitating hypergammaglobulinemic purpura. Complete (n = 4) or partial (n = 4) remission of hypergammaglobulinemic purpura occurred. In 5 additional patients with severe hypergammaglobulinemic purpura, attacks stopped with prednisone 25 to 60 mg daily. The mean duration of hypergammaglobulinemic purpura followup was 5.4 years (range 1-6 years). At last followup, hypergammaglobulinemic purpura had resolved in 11 (64.7%) patients despite persistently abnormal serology. CONCLUSION: In the absence of anti-Ro antibodies, a presumptive diagnosis of hypergammaglobulinemic purpura secondary to SLE should be questioned. Prednisone should be used only in severe hypergammaglobulinemic purpura. Indomethacin and hydroxychloroquine are of value in the treatment of milder hypergammaglobulinemic purpura.

Autoantibodies in systemic sclerosis.

There are 3 major autoantibodies in sera from patients with scleroderma: 1) anticentromere antibodies (ACA), 2) anti-topoisomerase I (anti-topo I), and 3) anti-RNA polymerases. Each is present in about 25% of patients and are mutually exclusive. ACA are found in patients with primary and secondary Raynaud's disease and in patients with primary biliary cirrhosis. Anti-topo I and anti-RNA polymerases are found exclusively in scleroderma. Each autoantibody is present in specific subsets of scleroderma patients. ACA and anti-topo I have been well studied and their presence and titer are stable over time. The anti-topo I and ACA are of all three isotypes, recognize multiple epitopes on the antigens and have stable cross reactive or private idiotypes. The antigen, topoisomerase I, has domains which have homology to viral proteins. Other autoantibodies predominantly recognize nucleolar antigens, are found in less than 15% of patients, and are not specific for scleroderma.

CENP-C, an autoantigen in scleroderma, is a component of the human inner kinetochore plate.

We have isolated and characterized a set of overlapping cDNA clones that encode the human centromere autoantigen centromere protein C (CENP-C). The identity of these clones has been established using several criteria. First, they were shown to encode a polypeptide that migrates at the expected position for CENP-C on SDS-polyacrylamide gel electrophoresis. Second, we have demonstrated that this polypeptide shares at least two epitopes with human CENP-C. Polyclonal antibodies were raised to fusion proteins encoded by nonoverlapping regions of the cDNA clones. These antibodies were shown to recognize a protein at a position appropriate for CENP-C on immunoblots of human chromosomal proteins. In addition, we used indirect immunofluorescence to demonstrate that these antibodies recognize centromeres of HeLa chromosomes in the expected pattern for CENP-C. Localization of CENP-C by immunoelectron microscopy reveals that this protein is a component of the inner kinetochore plate.

Injection of anticentromere antibodies in interphase disrupts events required for chromosome movement at mitosis.

We have used autoantibodies to probe the function of three human centromere proteins in mitosis. These antibodies recognize three human polypeptides in immunoblots: CENP-A (17 kD), CENP-B (80 kD), and CENP-C (140 kD). Purified anticentromere antibodies (ACA-IgG) disrupt mitosis when introduced into tissue culture cells during interphase. We have identified two execution points for antibody inhibition. Antibodies injected into the nucleus greater than or equal to 3 h before mitosis prevent the chromosomes from undergoing normal prometaphase movements in the subsequent mitosis. Antibodies injected in the nucleus during late G2 cause cells to arrest in metaphase. Surprisingly, antibodies introduced subsequent to the beginning of prophase do not block mitosis. These results suggest that the CENP antigens are involved in two essential interphase events that are required for centromere action in mitosis. These may include centromere assembly coordinate with the replication of alpha-satellite DNA at the end of S phase and the structural maturation of the kinetochore that begins at prophase.

Extensive calcinosis cutis with systemic lupus erythematosus.

Calcinosis cutis is a common clinical feature of dermatomyositis and scleroderma but is only rarely reported in association with systemic lupus erythematosus (SLE). We describe three patients with long-standing systemic lupus erythematosus in whom extensive calcinosis cutis developed. We identify characteristics our patients share in common with 23 previously described patients.

Clinical associations of anticentromere antibodies and antibodies to topoisomerase I. A study of 355 patients.

Anticentromere antibodies (ACA) and anti-topoisomerase I (anti-topo I) were assayed in serum samples from 355 patients: 89 with proximal scleroderma; 54 with CREST syndrome (calcinosis, Raynaud's phenomenon, esophageal dysmotility, sclerodactyly, telangiectasias), without proximal scleroderma; 154 with primary and secondary Raynaud's disease; and 58 with other rheumatic diseases, without Raynaud's disease. Sera from healthy control subjects were also assayed. Using immunoblotting techniques, anti-topo I was detected in 28% of the patients with proximal scleroderma; using immunodiffusion techniques, this antibody was found in only 20% of the same group of patients. Anti-topo I and ACA were found primarily in patients with scleroderma, CREST syndrome, and Raynaud's phenomenon. ACA identified patients with less severe disease, whereas anti-topo I identified patients with skin and cardiac involvement and patients with malignancies.

Molecular cloning of cDNA for CENP-B, the major human centromere autoantigen.

We have isolated a series of overlapping cDNA clones for approximately 95% of the mRNA that encodes CENP-B, the 80-kD human centromere autoantigen recognized by patients with anticentromere antibodies. The cloned sequences encode a polypeptide with an apparent molecular mass appropriate for CENP-B. This polypeptide and CENP-B share three non-overlapping epitopes. The first two are defined by monoclonal antibodies elicited by injection of cloned fusion protein. Epitope 1 corresponds to a major antigenic site recognized by the anticentromere autoantibody used to obtain the original clone. Epitope 2 is a novel one not recognized by the autoantibody. These epitopes were shown to be distinct both by competitive binding experiments and by their presence or absence on different subcloned portions of the fusion protein. The third independent epitope, recognized by a subset of anticentromere-positive patient sera, maps to a region substantially closer to the amino terminus of the fusion protein. DNA and RNA blot analyses indicate that CENP-B is unrelated to CENP-C, a 140-kD centromere antigen also recognized by these antisera. CENP-B is the product of a 2.9-kb mRNA that is encoded by a single genetic locus. This mRNA is far too short to encode a polypeptide the size of CENP-C. The carboxy terminus of CENP-B contains two long domains comprised almost entirely of glutamic and aspartic acid residues. These domains may be responsible for anomalous migration of CENP-B on SDS-polyacrylamide gels, since the true molecular mass of CENP-B is approximately 65 kD, 15 kD less than the apparent molecular mass deduced from gel electrophoresis. Quite unexpectedly, immunofluorescence analysis using antibodies specific for CENP-B reveals that the levels of antigen vary widely between chromosomes.

The clinical utility of the lupus band test.

In order to determine the clinical utility of the lupus band test, the presence of dermal-epidermal junction (DEJ) deposits of IgG, IgM, IgA, C3, C4, Clq, and properdin were studied in biopsies of clinically normal deltoid area skin from 102 patients with systemic lupus erythematosus (SLE) and 151 patients with other rheumatic diseases. One or more proteins were detected at the DEJ in 72.6% (74 of 102) of patients with SLE and in 36.4% (55 of 151) of the other patients, yielding a specificity of 64% and a predictive value of 57%. The predictive value for the diagnosis of SLE was greatest with C4 (100%), properdin (91.3%), and IgA (86.2%) and lowest with IgM (59%). Specificity and predictive value increased with the number of proteins detected at the DEJ. The results suggest that more rigid criteria are required before diagnostic significance is attached to a positive result on the lupus band test.

A long-term longitudinal study of anticentromere antibodies.

A longitudinal retrospective study of 37 patients previously tested for anticentromere antibodies (ACA) in 1982 was carried out. No ACA were found in stored sera from 22 ACA-negative patients including 1 patient with CREST (calcinosis, Raynaud's phenomenon, esophageal dysmotility, sclerodactyly, telangiectasia). All ACA-positive patients, with 1 exception, were found to have ACA in their sera over prolonged periods of time. One patient developed ACA for the first time when the third feature of the CREST syndrome (Raynaud's phenomenon) was noted. Anticentromere antibodies were IgG, and no consistent change in titer over time was found.

Familial partial deficiency of the third component of complement (C3) and the hypocomplementemic cutaneous vasculitis syndrome.

Familial hypocomplementemia of the third component of complement (C3) was found in four members of a family. The prospositus had cutaneous vasculitis, hypocomplementemia, arthralgia, proteinuria and thrombocytopenia. The combination of clinical, laboratory and pathologic findings resembled the "hypocomplementemic cutaneous vasculitis syndrome" (HCVS) or the "SLE-like syndrome" but serum C3 concentration was 35 to 57 per cent of normal in the propositus and in three relatives. Results of Clq precipitins, cryoglobulins and serologic tests for systemic lupus erythematosus were negative. Proteinuria (815 mg/day) but no hematuria was present. Analysis of the C3 phenotypes in this family showed that three hypocomplementemic members were apparent homozygous C3 slow but one was heterozygous C3 fast-slow. Metabolic studies with 125-Iodinated C3 in the clinically normal mother showed a 50 per cent reduction in C3 synthesis which was consistent with hypocomplementemia documented by serum protein assay. The occurrence of an immune complex-like disease (with characteristics of the HCVS) in a patient with a familial deficiency of C3 suggests that the preexisting C3 deficiency may predispose such persons to certain diseases.