Analysis of cerebrospinal fluid and serum of patients with multiple sclerosis by means of anti-idiotypic antisera

Twelve antisera were prepared, each specific for idiotype(s) of IgG from the cerebrospinal fluid (CSF) of an individual patient with multiple sclerosis (MS). Idiotype-positive (Id(+)) IgG represented 24 to 33% of the CSF-IgG. Radioautographic analysis showed that Id(+)IgG in CSF and in serum had a restricted electrophoretic mobility. From serum of three patients, at least two different noncross-reacting Id(+)IgG fractions were isolated. The anti-Id antisera, which detect IgG considered to be of importance in MS, were used to measure Id(+)IgG in CSF and in serum. Relative to the total IgG, Id(+)IgG was always enriched in CSF, which suggests intrathecal synthesis of such IgG. However, in all patients the absolute concentration of Id(+)IgG was higher in serum than in CSF. Studies attempting to demonstrate cross-reactivity between the IgG of different patients were negative.     

Analysis of a common inheritable idiotype in IgA-deficient sera using monoclonal antibodies

In these studies we describe the production of three mAb raised to an idiotype on an IgG anticasein antibody isolated from the serum of one IgA-deficient blood donor. These are IgM kappa and block the binding of casein Ag to anticasein antibody. Sera of unrelated IgA-deficient donors were tested for the presence of the idiotype; 15 of 56 IgA-deficient sera (25%) contain the anticasein idiotype, whereas 1 of 45 normal sera was positive. Anticasein antibodies as a whole were predominantly of the IgG1 and IgG3 subclass; idiotype-positive anticaseins are predominantly of the IgG1 subclass. For IgA-deficient donors, the relative amount of idiotype-positive anticasein antibody was correlated with the level of anticasein present in the serum. Studies were done to investigate the potential inheritance of the idiotype in families; in three of four families the idiotype was inherited in an apparent autosomal dominant pattern. Our data show that a common cross-reactive idiotype can be detected in the sera of IgA-deficient individuals and their family members. This suggests that V region markers may be conserved in this humoral immunodeficiency disease.     

Idiotypic analysis of anti-GAT antibodies. III. Determinant specificity and immunoglobulin class distribution of CGAT idiotype.

The humoral response to poly-(L-glutamic acid60, L-alanine 30, L-tyrosine10), GAT, in mice is further characterized by both idiotype and fine specificity analyses. The common idiotype on murine anti-GAT antibodies (CGAT) was identified in anti-GAT antisera from seven additional strains of mice. These data confirm that the CGAT idiotype can be induced in all inbred strains of mice. Using a partially inbred strain of mice selected for their ability to respond to poly-(L-glutamic acid50, L-tyrosine50), GT, we demonstrated that the GT copolymer is capable of inducing antibodies that express the CGAT idiotype. In contrast, antisera directed against poly-(L-glutamic acid60, L-alanine40), GA, bind GAT but lack CGAT idiotype. These results indicate that GAT molecules contain determinants either similar or identical to those on GT molecules, which are responsible for the induction of CGAT idiotypic antibodies. We also demonstrated that both GAT responder and nonresponder strains of mice can produce anti-GAT antibodies with similar fine specificity patterns and CGAT idiotype. In addition, we demonstrate that antibodies of the IgM, IgG1, and IgG2 immunoglobulin classes express the CGAT idiotype.     

Idiotypic analysis of a monoclonal anti-Sm antibody. II. Strain distribution of a common idiotypic determinant and its relationship to anti-Sm expression

The idiotypes borne by Y2, a monoclonal anti-Sm antibody of MRL-lpr/lpr mouse strain origin, were investigated to elucidate genetic mechanisms in this autoantibody response. An anti-Y2 anti-idiotypic antiserum was raised in a rabbit and was rendered specific for idiotype by extensive absorption with globulins of the B6 and BALB/c strains as well as the BALB/c myeloma UPC 10. By using a sensitive assay for idiotype by inhibition ELISA, the Y2 determinant was found to be commonly expressed in sera of MRL-lpr/lpr and MRL-+/+ mice. Moreover, sera of several normal strain mice also bore the idiotype and, in mice bearing the lpr gene, idiotype levels were increased 1.5 to fivefold, even in the absence of a serum anti-Sm response. The relationship of this idiotype to anti-Sm expression was further assessed by determining the idiotype content of affinity-purified anti-Sm antibodies from MRL-lpr/lpr mice. Anti-Sm from serum pools or individual animals showed no significant enrichment of the Y2 idiotype in comparison to unselected MRL-lpr/lpr IgG. These results suggest that the Y2 idiotype defines only a minor component of the anti-Sm autoantibody response, and that most antibodies with this determinant express other antigenic specificities.     

Structural studies on induced antibodies with defined idiotypic specificities. II. The light chains of anti-p-azophenylarsonate antibodies from A/J mice bearing a cross-reactive idiotype.

N-terminal amino acid sequence analyses have been performed on three preparations of light chains of A/J mice. Light chains derived from the IgG of unimmunized animals were compared to light chains of anti-p-azo-phenylarsonate (anti-Ar) antibodies possessing a cross-reacting idiotype (CRI); the latter were derived from the ascites fluid of a single A/J mouse, or from the pooled ascites fluids of 18 A/J mice. The heavy chains of these same two antibody preparations had previously been shown to comprise a single, homogeneous sequence to position 40. With few exceptions, the first 26 positions of light chains derived from unimmunized animals were extremely heterogeneous; the heterogeneity is comparable to that observed in a composite of sequence data on light chains of BALB/c myeloma proteins. Although the light chains obtained from anti-Ar antibodies possessing the CRI (whether from the pool of 18 A/J mice or from a single mouse) were more restricted in their sequence, at several positions as many as four alternative amino acids were detected. These studies indicate that an antibody population with defined idiotypic specificity, and very possibly identical heavy chain sequences, may contain at least four distinct light chains. The feasibility of structural studies on antibodies induced in individual mice is further demonstrated.     

Characterization of feline serum ferritin-binding proteins: the presence of a novel ferritin-binding protein as an inhibitory factor in feline ferritin immunoassay

Ferritin-binding proteins (FBPs) such as anti-ferritin antibody, α-2-macroglobulin, apolipoprotein B are expected to interact with circulating ferritin to eliminate it from circulation. However, we found that feline serum more strongly inhibits the detection of canine liver ferritin by immunoassay than its apoferritin; putative FBPs probably conceal ferritin epitopes detected by anti-ferritin antibodies. After complex formation between affinity-purified FBPs and canine liver ferritin, co-immunoprecipitates of the complex by anti-bovine spleen ferritin antibody were found to contain autoantibodies (IgG, IgM, and IgA) to ferritin by immunoblot analysis with antibodies specific for feline IgG, IgM, and IgA. On the other hand, affinity-purified samples did not show any inhibitory effect in the ferritin immunoassay. This result shows that feline serum has another FBP, which inhibits ferritin immunoassays, but not anti-ferritin autoantibody. A feline FBP was partially purified from feline serum by (NH₄)₂SO₄ fractionation (33–50%), gel filtration chromatography, and anion exchange chromatography. After binding of the partially purified sample with canine liver ferritin coupled-Sepharose gel, the FBP was separated and purified from complexes formed in a native-PAGE gel. SDS–PAGE analysis showed that the purified FBP is a homomultimer composed of 31 kDa monomeric subunits connected by intermolecular disulfide bonds. Detection of feline liver ferritin by immunoassay was inhibited by FBP in a dose-dependent manner. The purified protein molecules appeared to be conglomerate of pentraxin-like molecules by its electron micrographic appearance. These results demonstrate that feline serum contains a novel FBP as inhibitory factor of ferritin immunoassay with different molecular properties from those of other mammalian FBPs, in addition to auto-antibodies (IgG, IgM, and IgA) to ferritin.     

Separation of kallikrein 6 glycoprotein subpopulations in biological fluids by anion-exchange chromatography coupled to ELISA and identification by mass spectrometry

Kallikrein 6 (KLK6) has been shown to be aberrantly glycosylated in ovarian cancer. Here, we report a novel HPLC anion exchange method, coupled to a KLK6-specific ELISA, capable of differentiating KLK6 glycoform subgroups in biological fluids. Biological fluids were fractionated using anion exchange and resulting fractions were analyzed for KLK6 content by ELISA producing a four-peak elution profile. Using this assay, the KLK6 elution profile and distribution across peaks of a set (n = 7) of ovarian cancer patient matched serum and ascites fluid samples was found to be different than the profile of serum and cerebrospinal fluid (CSF) of normal individuals (n = 7). Glycosylation patterns of recombinant KLK6 (rKLK6) were characterized using tandem mass spectrometry (MS/MS), and found to consist of a highly heterogeneous KLK6 population. This protein was found to contain all of the four diagnostic KLK6 peaks present in the previously assayed biological fluids. The rKLK6 glycoform composition of each peak was assessed by lectin affinity and MS/MS based glycopeptide quantification by product ion monitoring. The combined results showed an increase in terminal alpha 2-6 linked sialic acid in the N-glycans found on KLK6 from ovarian cancer serum and ascites, as opposed to CSF and serum of normal individuals.     

Detection of antibodies to Angiostrongylus cantonensis in serum and cerebrospinal fluid of patients with eosinophilic meningitis

Adult and young adult antigens of Angiostrongylus cantonensis were purified by immuno-affinity chromatography and used to detect antibody in serum and cerebrospinal fluid (CSF), by enzyme-linked immunosorbent assay (ELISA), in cases of human eosinophilic meningitis or meningoencephalitis. The levels of IgG, IgA, IgM and IgE antibodies to A. cantonensis in these patients were higher than levels in control subjects. Antibodies in patients detected against adult and young adult worm antigens of A. cantonensis did not differ significantly. Significantly higher IgM and IgE antibody levels were observed in serum compared with CSF from infected patients (Student's t-test, P less than 0.05). Both adult and young adult A. cantonensis antigens proved to be highly sensitive in ELISA for serum antibodies; however, the sensitivity was significantly lower in tests on CSF.     

Ultrastructural Localization of Monoclonal IgG Antibodies for Antigenic Sites of Eimeria tenella Oocysts,Sporocysts, and Sporozoites1

Monoclonal IgG antibodies against sporozoites of Eimeria tenella were obtained from the ascites fluid of BALB/c mice. Oocysts, sporocysts, and sporozoites were exposed to medium 199, normal ascites fluid, or monoclonal antibodies 1A, 9D, 3D3II, or 2G8f. Specimens were then incubated with ferritin-conjugated goat anti-mouse IgG antibody. Ferritin was uniformly distributed over the surface of sporozoites exposed to 1A, 9D, or 3D3II; ferritin was localized in patches on sporozoites exposed to 2G8f. A uniform layer of ferritin was present on the inner layer of oocyst walls and on the Stieda body and outer surface of sporocysts exposed to 1A, 9D or 3D3II. In specimens treated with 2G8f, ferritin was present on the inner layer of the oocyst wall and the Stieda body, but not on the sporocyst wall. No ferritin was found on specimens exposed to medium 199 or normal ascites fluid. Monoclonal antibodies 1A, 9D, and 3D3II, but not 2G8f, caused complement-mediated lysis of sporozoites. These findings indicate that oocysts, sporocysts, and sporozoites of E. tenella contain common antigens specific for each monoclonal antibody tested.     

Mapping the idiotopes of a monoclonal anti-myoglobin antibody with syngeneic monoclonal anti-idiotypic antibodies: detection of a common idiotope

A panel of syngeneic monoclonal anti-idiotypic antibodies was prepared by immunizing A.SW mice with keyhole limpet hemocyanin-coupled A.SW monoclonal anti-myoglobin (HAL 19, IgG1) and screening the cloned hybridomas for production of IgG2 binding to idiotype but not to certain other anti-myoglobin antibodies of the same subclass in an ELISA. With these antibodies, we identified three nonoverlapping idiotopes, based on three clusters of monoclonal anti-idiotopes that mutually inhibit within each cluster, but not between clusters (Cluster I: S2, S6, S8; Cluster II: S5, S7; Cluster III: S9). Only Cluster II antibodies block the binding of myoglobin to HAL 19 and so identify a binding site-related idiotope(s). Binding of both Cluster II monoclonals (S5 and S7) to Hal 19 is inhibited by a rabbit anti-idiotype that we previously reported detects a common cross-reactive anti-myoglobin idiotope in immune sera. However, only one of these, S7, and not S5, identifies an idiotope that is present on 20 to 30% of A.SW anti-myoglobin antibodies in immune sera and ascites. The panel of syngeneic monoclonal anti-idiotype antibodies also detects new idiotopes not detected by the rabbit anti-idiotype. The development of a panel of syngeneic monoclonal anti-idiotypic antibodies to different clusters of idiotopes on the same antibody molecule, including one that identifies a major common idiotope in immune sera, should allow the analysis of possible idiotype network regulation in vivo and in vitro in a completely syngeneic system.     

Helper T lymphocytes from xid and normal mice support anti-phosphocholine antibody responses with equivalent T15, 511, and 603 idiotypic composition

We have examined the idiotypic composition of secondary adoptive transfer antibody responses to phosphocholine (PC) supported by KLH-primed helper T cells derived from normal mice or xid mice. CBA/N x BALB/c F1 male xid mice have diminished anti-PC responses and virtually undetectable levels of the T15 idiotype; xid mice do express the 511 and 603 idiotypes. Nonetheless, we find helper T cells derived from such mice are indistinguishable from T cells primed in a normal environment in their ability to cooperate with B cells producing anti-PC antibody bearing the T15, 511, or 603 idiotype markers. This result is in contrast to a previously published report from this laboratory. T cells from xid mice did support more IgG PFC than normal T cells, but serum IgG anti-PC antibody levels were similar in both groups. The IgM anti-PC response was predominantly of the T15 idiotype, whereas the 511 idiotype was associated with a minor fraction of IgG1 antibodies. The majority of the secondary IgG "anti-PC" antibody response bore none of the idiotypic markers associated with PC-binding myeloma or hybridoma antibodies, and was directed against phenyl-PC rather than PC. The phenomenon of T15 clonal dominance in the anti-PC response therefore is largely confined to the IgM response. We would conclude that the idiotype levels in the T cell priming environment do not influence the subsequent ability of such primed T cells to support anti-PC antibody responses.     

Importance of antibody isotype in monoclonal anti-idiotype therapy of a murine B cell lymphoma. A study of hybridoma class switch variants

An initial panel of four syngeneic monoclonal antibodies directed against the idiotype of a murine B cell lymphoma was used to treat this tumor in vivo. The antibody in the panel of the IgG2a isotype was more effective in treatment than the other antibodies, which were of the IgG1 and IgG2b isotypes. To independently assess the role of antibody isotype in mediating antitumor effects, switch variant hybridoma families were isolated from the hybridomas secreting the less effective IgG1 and IgG2b antibodies. A family isolated from an IgG1-secreting parent consisted of IgG1-, IgG2b-, and IgG2a-secreting members, and an IgG2a variant was isolated from an IgG2b-secreting parent for another family. Antibody members of each family differed only in heavy chain composition and were the same with respect to their light chains and their affinity and specificity for idiotype. The IgG2a members of both families were superior to the other members in inhibiting tumor growth with an order of effectiveness of IgG2a greater than IgG1 greater than IgG2b. These in vivo results paralleled the abilities of these different isotype antibodies to mediate antibody-dependent cellular cytolysis in vitro. For the IgG2b----IgG2a family, in vivo treatment with the IgG2a member given i.p. after i.p. tumor challenge at one-tenth the dose of the IgG2b member was still superior to the latter. At one-hundredth the dose of the IgG2b, the IgG2a was still superior to the latter when the antibodies were given i.p. and tumors subcutaneously. These data and those showing that the clearance of these antibodies from the serum differed in only a relatively minor way indicate that the IgG2a antibodies in this system had greater antitumor effects primarily by virtue of their greater capacity for host effector interaction.     

Measles-virus-specific IgG in optic neuritis and in multiple sclerosis after optic neuritis.

Measles-virus-specific IgG was measured in the serum of 100 patients who had presented with optic neuritis (ON) during 1960-74. When reviewed 41 of them were found to have developed definite symptoms and signs of multiple sclerosis (MS), their serum containing significantly higher titres of the antibody than sera from either the rest of the patients or a group of normal healthy controls. In a few patients from whom cerebrospinal fluid (CSF) was obtained in the acute phase of ON, titres of measles IgG in the serum was higher in those in whom the antibody was detected in the CSF than the serum of patients without CSF antibody.     

IgG autoantibody activity in normal mouse serum is controlled by IgM

In the serum of normal BALB/c mice, IgG antibody reactivity to mouse actin and tubulin, DNA, and TNP groups was very low compared to that of the IgM. This activity was considerably increased when IgG was separated, by affinity chromatography on protein A-Sepharose, whereas no difference in the IgM activity was observed. Addition of IgM to IgG isolated from the same serum resulted in the inhibition of IgG binding to these Ag. Isolation of IgG antibodies on actin, TNP, and tubulin immunoadsorbents has indicated that at least part of the IgG antibodies is polyreactive. In order to understand this inhibition better, experiments with F(ab')2 fragments of IgG were performed. IgM inhibited the binding of F(ab')2 to the antigens in a dose-dependent manner and reacted with immobilized F(ab')2. IgM isolated on F(ab')2 immunoadsorbent, as compared to the initial IgM preparation, were less active toward the Ag but more inhibitory for IgG binding to the Ag. In some pathologic situations, IgM failed to inhibit some IgG antibody activities. The anti-DNA IgG activity from (NZB x NZW)F1 mice was not affected by autologous IgM. Similarly the anti-tubulin IgG from mice infected with Trypanosoma cruzi were less inhibited by IgM from autologous serum than antitubulin IgG from normal mice. These results are compatible with the existence in normal mice of an idiotypic-like network, regulating via an IgM population in the serum, the binding of IgG autoantibodies to self Ag. Modifications of this idiotype-anti-idiotype system might lead to the expression and/or expansion of autoreactive IgG-producing clones.     

Usefulness of IgG4 subclass antibodies for diagnosis of human clonorchiasis

The present study analyzed serum IgG subclass antibody reaction to major antigenic bands of Clonorchis sinensis to investigate improvement of its serodiagnosis. Of the four subclass antibodies, IgG1 and IgG2 antibodies were produced but not specific, IgG3 antibody was least produced, and IgG4 antibody was prominent and specific. The serum IgG antibody reaction to any of 43-50, 34-37, 26-28, and 8 kDa bands was found in 65.5% of 168 egg positive cases while IgG4 antibody reaction was found in 22.0% of them. The positive rates of IgG and IgG4 antibodies were directly correlated with the intensity of infection. All of the sera from heavily infected cases over EPG 5,000 showed positive reaction for specific IgG and IgG4 antibodies. The specific serum IgG4 antibody disappeared within 6 months after treatment. The bands of 35 kDa and 67 kDa cross-reacted with IgG antibodies but not with IgG4 antibodies in sera of other trematode infections. The present findings suggest that serum IgG4 antibody reaction to 8 kDa band is specific but not sensitive. Any method to increase its sensitivity is required for improved serodiagnosis.     

Evaluation of the usefulness of the ELISA method for determining the level of enterovirus antibodies in serum and cerebrospinal fluid

The usefulness of the methods was compared: complement fixation test (CFT), neutralization test (NT) and ELISA IgG and IgM against enteroviruses for the evaluation of specific immune reaction in sera and cerebrospinal fluid (CSF) samples of patients with confirmed enterovirus infections. The criteria were established for the assessment of ELISA results in rapid diagnosis of enterovirus neuroinfections. The criteria accepted by the producer lowered the sensitivity of the method and the possibility of recognition of local synthesis of antibodies in the CNS. The use of serum negative in CFT and negative CSF as reference for the determination made possible using of that kit for rapid diagnosis of neuroinfections. The modified ELISA IgG test makes possible determination of antibodies in CSF and serum, and accepting the generally recognized criteria for local production of antibodies in the CNS the ELISA test makes possible rapid diagnosis of neuroinfections which is not possible by other methods.     

Structural and antigenic studies of an idiotype-bearing murine antibody to the arsonate hapten.

Mice of strain A/J responded to repeated intraperitoneal injection of Limulus hemocyanin derivatized with arsanilic acid by producing large quantities (approximately 5 mg/mL of ascites fluid) of IgG antibodies specific for this hapten. The antibodies possessed a characteristic idiotypic determinant and exhibited restricted heterogeneity as demonstrated by isoelectric focusing and primary N-terminal amino acid sequence analysis of isolated light and heavy polypeptide chains. Both light- and heavy-chain sequences were comparable to those of myeloma proteins in lack of heterogeneity. The N terminus of the light chain exhibited V KI sequence and only one position in the first 30 residues showed more than one amino acid. No variability was observed in the first 10 N-terminal residues of the heavy chain. Rabbit antiserum to the idiotype blocked binding of hapten by the purified antibody. The presence of both light- and heavy-chain antigenic determinants was required for optimal formation of the idiotypic determinant.     

Protection mediated by antibodies to iron-regulated outer-membrane proteins of S. typhi in a mouse peritonitis model

Iron limitation induces the expression of iron-regulated outer-membrane proteins, which are not expressed under iron sufficient growth conditions. In the present study, these proteins were purified in order to evaluate their protective potential in the experimental model. Anti IROMPs antiserum was raised in rabbits. In mice, passively transferred anti-IROMPs antibodies provided 60% protection against the serovar Typhi challenge dose (9.6 LD₅₀). The hyperimmune serum containing anti-IROMPs antibodies were also found to be bactericidal in the presence of complement whereas no bacterial killing was observed with pre-immunized serum. Bactericidal titre of anti-IROMPs serum was fond to be 2000 as more than 50% killing was observed with serum diluted to 1:2000. The role of IROMPs was assessed in actively-immunized mice followed by challenge with serovar Typhi. These proteins provided protection in 90% mice against challenge (480 LD₅₀) with the pathogen. The levels of isotypes of antibodies (IgG, IgM & IgA) in the sera and secretory antibodies (sIgA) in the gut fluid of immunized mice correlated with the protection. This study, thus indicates that anti IROMPs antibodies may play an important role in providing protection at systemic as well as at mucosal level.     

Recognition of two Dermatophagoides pteronyssinus-specific epitopes on antigen P1 by using monoclonal antibodies: binding to each epitope can be inhibited by serum from dust mite-allergic patients

Monoclonal antibodies were raised against Antigen P1, the major allergen of the house dust mite (Dermatophagoides pteronyssinus). The majority were Antigen P1 specific, isotype IgG1, and did not react with a comparable D. farinae allergen. These antibodies bound 38 to 50% of 125I Antigen P1 in antigen-binding assays (titer greater than or equal to 1/1,000,000), and the quantities of IgG antibody in ascites were 2 to 4 logs greater than those in polyclonal mouse antiserum or in serum from a mite-allergic patient. Two IgM antibodies showed weak binding to Antigen P1 but reacted strongly with D. pteronyssinus in enzyme immunoassay (titer greater than or equal to 1/100,000). Assessments of the specificity of the IgG antibodies by using two inhibition radioimmunoassays suggested that they were directed against two different epitopes. Antibodies 10B9 F6 and 5H8 C12 were purified by preparative isoelectric focusing (isoelectric points of pI 6.25 and 7.4, respectively) and radiolabeled with 125I. Cross-inhibition experiments, using ascites dilutions to inhibit binding of each radiolabeled antibody to Antigen P1, confirmed that these antibodies recognized two distinct epitopes. Analysis of antibodies from 39 clones/hybrids showed that the majority were directed against the same epitopes as either 10B9 F6 or 5H8 C12 (3 out of 39 [8%] and 29 out of 39 [74%], respectively). None of the monoclonal antibodies significantly inhibited (greater than 10%) human IgE binding to Antigen P1 in the radioallergosorbent test. However, 12 of 14 sera from mite allergic patients inhibited binding by the monoclonal antibodies. One serum from a mite-allergic patient inhibited binding of both 10B9 F6 and 5H8 C12 by greater than 85% and showed parallel inhibition curves. The results suggest that these monoclonal antibodies could be used to assay Antigen P1 in both D. pteronyssinus and house dust extracts. It should also be possible to use monoclonal antibodies in inhibition assays to define the antigenic/allergenic determinants recognized by human IgG and IgE antibodies on this mite allergen.     

Detection of antibodies in serum and cerebrospinal fluid to Toxoplasma gondii by indirect latex agglutination test and enzyme-linked immunosorbent assay.

Sensitivity of anti-Toxoplasma antibody (IgG) test by enzyme-linked immunosorbent assay (ELISA) was evaluated in comparison with indirect latex agglutination (ILA) using 2,016 paired human samples of serum and cerebrospinal fluid (CSF). The samples were collected from neurologic patients in Korea with mass lesions in central nervous system (CNS) as revealed by imaging diagnosis (CT/MRI). When the sera were screened for anti-Toxoplasma antibody by ILA, 76 cases(3.8%) were positive (1:32 or higher titers). In the paired samples of CSF, no positive reactions were observed. When ELISA was performed using PBS extract of Percoll purified tachyzoites as antigen, cut-off absorbance was determined as 0.40 for serum and 0.27 for CSF tests. The antibody positive rates by ELISA were 7.0% in serum and 5.6% in CSF. Of them, 40 cases (2.0%) showed positive reactions in both serum and CSF. The antibody positive rates were higher in groups older than 40 years. The rates were higher in male (4.7% by ILA, 8.3% by ELISA) than in female (2.2% by ILA, 5.0% by ELISA). The rates in CSF showed no such sex difference. ELISA showed twice higher positive rates when serum was tested, and was sensitive enough to detect specific antibodies in CSF. Etiologic relations between positive antibody tests and CNS lesions remained unknown.