Lyme Neuroborreliosis: A New Sensitive Diagnostic Assay for Intrathecd Synthesis of Immunoglobdn G, A, and M Klaus Hansen, MD, and Anne-Mette Lebech, MD ~~

An antibody capture enzyme-linked immunosorbent assay was developed to measure directly intrathecal immunoglobulin (Ig)G , A, and M synthesis specific for Bowelia burgdoderi. Purified, biotin-avidin-peroxidase-labeled B. burgdorferi flagella was used as test antigen. Paired cerebrospinal fluid and serum specimens from 100 patients with clinically definite neuroborreliosis and 35 control subjects with neurological diseases were examined. Significant B . burgdogerispecific intrathecal IgG, A, and M production was found in 89%, 65%, and 67% of patients with neuroborreliosis. Local synthesis of specific IgA was only seen in patients with significant local I g G synthesis. Antibody production in cerebrospinal fluid began by 2 weeks after onset of neurological symptoms. At the end of the second week specific IgM, IgG, or both, was detected in 88% of the patients. Specific IgG synthesis was present in all patients by 6 weeks after onset. Specific local IgM synthesis usually disappeared by 3 to 6 months after therapy, whereas specific IgG synthesis persisted after recovery. Even in patients with a severely altered blood-brain barrier, the assay discriminated between intrathecal antibody synthesis and antibody leakage from serum. The assay makes diagnostic measurement of B . burgdogeri-specific intrathecal antibody synthesis reliable, rapid, and accessible as a routine serological test. Hansen K, Lebech A-M. Lyme neuroborreliosis: a new sensitive diagnostic assay for intrathecal synthesis of Borreha burgdorferi-specific immunoglobulin G, A, and M. Ann Neurol 1991;30:197-205 Neuroborreliosis is a frequent and serious manifestation of Lyme borreliosisf caused by the tick-borne spirochete B o d i a bzlrgdoderi [I, 2). The best current indicators of neuroborreliosis are lymphocytic pleocytosis and B. bargdor$eri-specific antibody synthesis within cerebrospinal fluid (CSF). Patients with neuroborreliosis develop a specific intrathecal immune response that is demonstrated by the preferential compartmentalization of B. burgdoderi-specific antibody-secreting B cells [ 3 } , specific T cells 141, and the finding of ohgoclonal immunoglobujin (Ig) G specific for B. burgdovferi [5}. These techniques, however, are not feasible for routine investigation of CSF. In this study, our aim was to develop a simple diagnostic assay for B. burgdor-ri-specific intrathecal antibody production based on a capture enzyme-linked immunosorbent assay The idea Of this assay was originally introduced in 1982 by Burke and ColleWes L61, who measured specific antibodies in CSF in Patients with Japanese encephalitis; it has since been applied to a few viral neuroinfections 17-91. Normally, immunoglobulin concentration in CSF is low; there-

fore, specific antibody production in CSF will, in contrast to serum, represent a large proportion of the total immunoglobulin. This proportion can be measured directly by a capture ELISA. Based on the investigation of paired CSF and serum samples from 100 patients with neuroborreliosis, this technique proved to be a sensitive and reliable assay for demonstration of intrathecal B . bwgdorferi-specific IgG, IgA, and IgM synthesis. The assay avoids complicated corrections for impairment of the blood-brain barrier (BBB).

From rhe Borrelia Laboratory, Department of Infection-Immunology, Statens Seruminstitut, Copenhagen, Denmark.

Address correspondence to Dr Hansen, Borrelia Laboratory, Department of Infection-Immunology, Starens Seruminstitut, Artillerivei 51 DK-2300 'Openhagen " Denmark.

Received Jul 25, 1990, and in revised form Oct 16, and Jan 28 and Feb 18,1991. Accepted for publication Feb 23, 1991.

Materials and Methods

SEC bjectJ Paired CSF and serum samples were obtained from 100 patients with neuroborreliosis who were hospitalized in the period 1984 to 1989. In 91 of the patients, the diagnosis was second-st;ige lymphocytic meningoradiculitis (LMR)C 10121, and in 9 patients, third-stage chronic progressive encephalomyelitis [13}. A preceding erythema migrans was reported by 56 patients. All patients had lymphocytic pleocytosis in CSF, 13-830 x lo6 cells/L (median, 211 x lo6

Copyright 0 1991 by the American Neurological Association 197

Table I . Intru-Assay Variation of the IgG. I g A , and ISM Capture ELISA" IgA


Sample 1 Sample 2 Sample 3


Mean OD


Mean OD


Mean O D


0.224 1.352 1.815

0.014 0.07 1 0.074

0.101 0.721 2.017

0.01 1 0.052 0.075

0.409 0.778 1.931

0.045 0.055 0.100

"Based on 24 measurements on the same plate of each of 3 samples with a low, medium, and high OD signal in the indicated assay. Ig = immunoglobulin; ELISA = enzyme-linked immunosorbent assay; OD = optical density.

cellsiL), and an elevated protein concentration (>0.5 gm/L) in all but 5 patients (0.2-12 gmiL; median, 1.2 gmiL). All measurements were done o n pretreatment samples obtained 4 days 106 years (median, 26 days) after onset of neurological symptoms. Consecutive CSFiserum samples were available from 10 patients with LMR,who were followed up with 2 to 5 control lumbar punctures 3 to 12 months after treatment. Paired CSFiserum samples served as controls from patients with clinically definite multiple sclerosis (MS; n = 17), Guillain-Bar& syndrome (GRS; n = 8), neurosyphilis (n = 4), and acrodermatitis chronica atrophicans (ACA; n = 6) [141.

ELISA Test Antigens In indirect ELISAs, we used the 41-kd native B. burgdorjeri flagellar protein purified from strain DK1 [l5]. For capmre ELISAs, purified flagella were biotin labeled 1161.

IgG, A, und M Antibody Capture ELISA Separate microtiter plates were coated with y-chain-specific rabbit anti-human IgG, a-chain-spechc rabbit anti-human IgA, and pchain-specific rabbit anti-human IgM (code A423, A262, and A425; Dako, Copenhagen, Denmark) diluted 1:1000 in phosphate-buffered saline (PBS; p H 7.2). The wells were then sequentially incubated with (1) serum (1:200) or CSF (1:lO) for 2 hours and (2) biotin-labeled B . burgdorfeeri flagella complexed with avidin peroxidase for 3 hours 1161. Corresponding CSFiserum samples were tested in duplicate and on the same plate. A sample, positive for the respective antibody isotype, and a negative control were included on every plate. The intra-assay variation of the IgG, A, and M capture ELISA was determined at three optical density (OD) levels (Table 1). Definite intrathecal antibody synthesis was assumed if the net OD difference of CSF to serum was positive and exceeded 3 SDs of the intra-assay variation in the corresponding OD interval. A specific capture antibody index, considering both the ratio and the net OD difference





was defined and used as a numerical expression of the test result. An antibody index 20.3 corresponds to net OD differences exceeding the + 3 SD level of the intra-assay variation for low, medium, and high O D values and is therefore a reliable indicator of intrathecal antibody synthesis.

198 Annals of Neurology Vol 30 No 2 August 1991

Routine Lyme BoweLiosis SwoLogy for I& and IgM in Serum and CSF Anti-B. burgdo~eriIgG in CSF and serum and anti-B. burgdorjeri IgM in CSF were quantitated using an indirect ELISA 1151. B . burgdorfri-specific IgM in serum was measured by a pcapture ELISA {16]. For detection of specific IgG and IgM in serum, the 98% specific diagnostic cutoff was OD = 0.160 and 0.300. Using an indirect IgG and IgM ELISA and a standard CSF dilution (1:25), an arbitrary high cutoff at OD = 0.160 was chosen fl5].

Results Routine Lyme Borveliosts Sepology Figure 1 shows the anti-B. burgdorferi flagella IgG and IgM reactivity in sera and CSF samples from 100 patients with neuroborreliosis; 84% were IgG and 57% were IgM sero-positive, whereas 79% and 55% had elevated I g G and IgM concentrations, respectively, in CSF. All samples from the 25 patients with MS and GBS were negative. O n e serum and two CSF samples from 2 patients with neurosyphilis were slightly IgG positive (OD range, 0.160-0.270).

Dose-response Relationship of the B. burgdorferispec$c IgG and IgM Capture EUSA Figure 2 A and B show titration curves for B. burgdorferi-specific IgM and IgG antibodies in paired CSFI serum samples from 2 patients with LMR. T h e significantly higher OD signal obtained by CSF compared with serum indicates an increased proportion of specific IgG and IgM versus total IgG and IgM due to intrathecal antibody synthesis. If specific antibodies in CSF are derived from serum, the proportion of specific antibodies to total IgG and IgM will be identical or lower in CSF compared with serum. The net OD difference of CSF to serum will then be 5 0 as shown in Figure 2C, where a paired CSF/serum sample from a patient with ACA was examined. Saturation of the solid-phase bound anti-c~and anti-y is only achieved at low CSF sample dilutions, e x p k n i n g the plateau of the CSF titration curves (see Fig 2A, B). Therefore, CSF was diluted 1:lO. The dilution of serum is less critical as saturation will be achieved until dilutions exceed lo3 to lo4 (see Fig 2A, B, C) [16].

lntrathecal B. burgdorferi-specific IgG, A, and M Synthe.fis in Neuroborreliosis

OD IttJt

2 1.8


. . ..




.. 0


. .. .









. .. . . ..




. . . . H




.* 6


. .ace




i: d3


. t


.. I



I 2



i 0.



..+. *E


*#r 0::





Fig 1. Anti-B. burgdorferi immunogLobulin (Ig) G and IgM levels in paired serum and cerebrospinalfluid (CSP) samples fmm 100 patients with neuroboweliosis. Using a standard dilution of 1 :200 for serum and 1:25 for CSF, specifc IgG in serum and CSF and specific IgM in CSF were measured by indirect enzyme-linked immunosorbent assay (ELlSAI. Specific IgM in serum was measured by p-capture ELISA. The horizontal lines mark the diagnostic cutoff limits.

Figure 3 shows the intra-BBB production of specific IgG (panel A), IgA (panel B), and IgM (panel C) in 100 patients with neuroborreliosis. Due to the relatively high proportion of specific IgM versus total IgM in serum from these patients, the net OD difference of CSF to serum was often lower for IgM than for IgG measurements (see Fig 3A, C). Regarding specific 1 6 , only few patients with high specific serum IgG titers showed a significantly increased ratio of antigenspecific IgG versus total IgG in serum, resulting in a high serum y-capture signal (see Fig 3C). The specific capture antibody index was defined to obtain a single numeric expression of the test result and to define an upper reference limit of the assay. A specific antibody index 20.3 was found for IgG in 89%, for IgA in 65%, and for IgM in 67% of the patients with neuroborreliosis. Fifteen patients showed only intrathecal IgG synthesis, whereas only 2 patients had IgM without concomitant IgG synthesis. No patients had only specific IgA synthesis. Although only 2 patients showed an exclusively local IgM production, combined evaluation for intra-BBB IgG and IgM synthesis was advantageous because many patients with a low specific I g G synthesis showed a pronounced local IgM production and vice versa, thus strengthening the diagnostic evidence. Dividing the 100 patients with neuroborreliosis into 6 groups according to disease duration (Table 2) illustrates the rapid onset of anti-B. bargdo$eri-specific intrathecal antibody synthesis by the second week after onset. By the sixth week, all patients with neuroborreliosis showed intrathecal specific IgG production. IgM synthesis in CSF was common even in patients with a disease duration between 6 weeks and 6 months. Only in 1 patient with chronic neuroborreliosis was a low intrathecal specific IgM synthesis found (index = 0.36). The autochthonous character of the intrathecal immune response is further demonstrated in patients with a pronounced antibody response only in CSF. All 5 patients without detectable specific IgG or IgM in serum showed a specific IgG index >0.3 and, in 4 patients, a specific IgM index >0.3. Of 16 patients without detectable peripheral specific IgG, 13 (81%) had a specific IgG index B0.3; and of 43 patients without significant specific IgM levels in scrum, 23 (54%) had an IgM index >0.3. Even the antibody isotype of the peripheral and the intrathecal immune response could be different. Two patients, who in serum were only IgM positive, showed an exclusively intrathecal IgG immune response. B. burgdorferiSpecificity of the Capture ELISA specif;c Intrathecal Antibody Synthesis Figure 4 illustrates the specificity of the capture ELISA for B. bargdorfei-specific intrathecal antibody synthe-

Hansen and Lebech: Intrathecd B. burgdorferi-specific IgG, IgA, and IgM in Neuroborreliosis


OD 2’o





lo3 id

















B Fig 2 (A-Cj. Dose-response relationship of the p-capture and the y-capture enzyme-linked immuno.torbent assay using a tenfold dilution series of paired cerebrospinalfEuidlserum samples from 2 patients with neuroborreliosis (A, BJ and from a patient with acrodematitzs chronica atrophican (C), 0-0, CSF, @-@, serum. The plateau of the cerebrospinal&id titration carves (A,B) indicates saturation of the solid-phase bound anti-p and anti- y.







sis testing paired CSF/serum samples from 17 patients with MS, 8 patients with GBS, and 4 patients with neurosyphilis. Only one sample from 1 patient with neurosyphiiis showed an IgA index >0.3. The ability of the capture ELISA to discriminate between intrathecal antibody synthesis and an increased concentration of specific antibodies in CSF due to transudation from serum was evaluated. For this purpose, we tested 6 patients with ACA, all with very high specific serum IgG levels (OD, 1.380-2.000), but a normal albumin ratio and a normal CSF cell count. Although 2 of them showed a significantly increased specific IgG concentration in CSF by indirect ELISA, the capture technique excluded intrathecal antibody synthesis in all ACA patients. The example in Table 3 demonstrates that the test principle holds up in an extreme situation. To simulate blood contamination of CSF, serum from a patient with ACA and a high titer was added to autologous CSF specimen. None of the CSF samples showed specific intrathecal IgG or IgM production although the CSF concentration of anti-B. bzrgdorjiii IgG and IgM was considerably increased (see Table 3). Comparison of B. burgdorferi-specz$c CSF Antibody Detection by Indirect ELISA and by Capture ELISA By indirect ELISA with a standard dilution of CSF and an arbitrary high cutoff level, 21% of patients with 200

Annals of Neurology Vol 30

No 2

August 1991

neuroborreliosis were negative for IgG and 45% were negative for IgM antibodies in CSF. Using the specific capture antibody index, 12 of the 21 (5796) CSFIgG-negative patients had an IgG antibody index 20.3 and 15 of 45 (33%) CSF-IgM-negative patients showed an IgM antibody index 20.3. Specific IgG and IgM antibody indices

Lyme neuroborreliosis: a new sensitive diagnostic assay for intrathecal synthesis of Borrelia burgdorferi--specific immunoglobulin G, A, and M.

An antibody capture enzyme-linked immunosorbent assay was developed to measure directly intrathecal immunoglobulin (Ig) G, A, and M synthesis specific...
703KB Sizes 0 Downloads 0 Views