Intrathecal synthesis of specific antibodies in neuroborreliosis comparison of different ELISA techniques and calculation methods

64

Journal of the Neurological Sctences, 118 (1993) 64-72 © 1993 Elsexaer Science Pubhshers B V All rights reserved 0022-510X/93/$06.00

JNS 04069

Intrathecal synthesis of specific antibodies in neuroborreliosis Comparison of different ELISA techniques and calculation methods R. Kaiser and C.H. Liicking Department of Neurology, Umverstty of Frelburg, Fretburg. Germany (Recewed 4 November, 1992) (Revised, received 1 March, 1993) (Accepted 3 March, 1993)

Key words Antibody Index; Borreha burgdorfen, CSF, ELISA, Immunofluorescence assay Summary The sensitivity of six different ELISA techniques and calculation methods for the determination of mtrathecal synthesis of IgG antibodies specific to Borreha burgdorfert was investigated m paired CSF and serum specimens from 33 patients with neuroborrehosts. The dlagnosttc value of the Antibody Index (AI) was compared with the meaningfulness of serum anttbodles to B burgdorferi (Bb), established by immunofluorescence assay (IFA) The AI, as a standard for mtrathecal anttbody synthesis was determined from specific antibody ratios (QBb) in the CSF and serum and the CSF/serum ratto of IgG ( Q I ~ ) or albumin (QAIb) Using Western blottmg with identical concentrations of IgG in the CSF and serum all patients displayed intrathecal synthesis of specific antibodies to at least two B. burgdorfen protems. The different ELISA methods and calculation procedures were almost equivalent in demonstrating mtrathecal synthesis of specific antibodies (32 and 33/33) Calculation of AI from IFA tlters was somewhat less sensitive (30/33) In 5 patients ttters of serum IgG- and IgM-antibodies to B burgdorfen determined by IFA were within the normal range or borderline, whtle elevated AIBb values indicated an autochthonous immune response to B burgdorfen in the CSF. In uncertam cases of neuroborrehosls calculation of AI from ELISA tlters will be useful in clarifying the diagnosis.

Introduction Since the identification of Borrelia burgdorferi as the organism responsible for Lyme disease an increasing spectrum of neurological disorders has been associated with this infection (Reik 1991). In endemic areas exposure to this spirochete - and consequently, specific immunoreactlvity - is widespread, making it difficult to determine if an individual patient's neurological problems are causally related to B. burgdorferi infection (Mautner et al. 1990). Isolation of B. burgdorfen from the CSF, however, is not suitable as a diagnostic routine test (Barbour 1984). Previous studies have indicated that the assessment of mtrathecal production of anti-B, burgdorfert antibodies provides a useful marker of CNS infection (Hansen et al. 1990; Steere et al. 1990; Stiernstedt et al. 1985). Intrathecally produced antibodies may be

Correspondence to Dr R Kaiser, Department of Neurology, University of Frelburg, HansastraBe 9, D-7800 Frelburg, Germany Fax 0761-2705310

detected by qualitative or quantitatwe methods. A qualitative method is to compare specific antibody bands in Western blotting with CSF and serum diluted to the same concentration of IgG (Weber et al. 1987; Kaiser 1990). The sole or more intense staming of certain bands in the CSF indicates intrathecal synthesis of these antibodies. Although S D S - P A G E along with Western blotting is a suitable technique to prove intrathecal synthesis of specific antibodies, this method is too costly for routine invesUgations in terms of time and labor. A simple and rapid method to analyze an autochthonous immune response of the CNS is to calculate the fracUon of locally produced specific antibodies in the CNS from antibody txters in CSF and serum by enzyme-linked immunosorbent assay. A higher proportion of specific antibodies per unit IgG in the CSF indicates intrathecal synthesis of these antibodies (Retber 1988; D6rnes et al. 1991). Up to now a considerable number of different E L I S A and calculation procedures have been published. First, specific antibody titer values were used to calculate the AI (Ukkonen et al. 1981) Other authors have applied the same amounts

65 of total IgG for CSF and serum samples in the paired analysis (Rehse-Kiipper and Ackermann 1986; Biniek et al. 1988). Ratios were calculated not only from titers but also from optical densities in both fluids (Stiernstedt et al. 1985). Specific antibody ratios were related to the IgG ratio (Ql~o = concentration of IgGcsF/se~m) or to the albumin ratio (QAlb = concentration of albumincsF/se~m) (Klapper et al. 1981; Ukkonen et al. 1981). The aim of this study therefore was to compare the different methods to investigate an autochthonous immune response in the CNS to prove the clinical diagnosis of suspected Lyme neuroborreliosis.

Materials and methods

Subjects 33 paired CSF and serum samples from 25 patients with early Lyme neuroborreliosis (ELN) and 8 patients with late Lyme neuroborreliosis (LLN) (Reik 1991)

were available (Table 1). Patients were treated with Ceftriaxon 1 × 2 g intravenously for 14 days. In all patients lumbar puncture was performed before treatment. A preceding erythema migrans was reported in 12 patients, 11 individuals remembered a tick bite. All patients had lymphocytic pleocytosis in CSF: ELN: 23-610 × 106 cells/l (median 242 × 106 cells/l), LLN: 10-120 × 106 cells/l (median, 60 × 106 cells/l), and a breakdown of the blood-CSF barrier (QAlb = albumin CSF/serum ratio > 7.5 × 10-3): ELN: 8-58 × 10 -3 (median 24 × 10-3), LLN: 9.4-70.6 × 10 -3 (median 39 × 10-3). Calculations by the Reiber formula (Reiber and Felgenhauer 1987) revealed intrathecal synthesis of IgG in 20 patients, of IgM in 25 and of IgA in 10 patients. Oligoclonal IgG bands in the CSF were found in all patients (Kaiser 1991). As controls, matched CSF sera from 15 individuals with serum IgG-antibodies to B. burgdorferi in immunofluorescence assay (IFA) but no history compatible with neuroborreliosis were analyzed. Investigatton of the CSF in these patients who were examined be-

TABLE 1 CLINICAL D A T A F R O M 33 PATIENTS W I T H N E U R O B O R R E L I O S I S Patient 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 EM" erythema mlgrans

Course acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute acute chronic chrome chronic chronic chronic chronic chromc chronic

Neurological mamfestations Meningoradlculitis, N VII palsy Meningoradicuhtis, N VII palsy Menmgoradicuhtis, N VII palsy Menmgoradlculias, N VII palsy Menmgoradlculitis, N VII palsy Menlngoradiculitis, N VII palsy Meningoradicuhtis, N VII palsy Menmgoradlculihs, N VII palsy Memngoradlculitls, N VII palsy Meningoradicuhtis, N VII palsy MeningoradicuhUs, N VII palsy MenlngoradlculiUs, N VII palsy Myelomeningoradlculitis Menmgoradicuhtis Menlngoradlculitls Menmgoradiculitls Meningoradicuhtis Memngoradlcuhtls Menlngoradlculitls MeningoradlcullUs Memngoradlcuhtls Meningoradtcuhtis Menmgoradlculltis Menmgoradiculitls Meningoradicuhtis, Papdlitls Menlngoradlculitls, N VII palsy Spastic paraparesis SpasUc paraparesis Spastic herniparesis Spastic paraparesls Spastic paraparesls Spastic paraparesls Motor Neuron Disease

Tick bate no no yes no no no yes yes yes yes no no no yes no no no no no yes no yes yes no yes no no no no no no yes no

EM yes yes yes yes no no yes yes yes yes no no no no yes no rio yes no no no yes no no no no no no no no no yes no

66 cause of headache (cerebral infarction, classical migraine, endogenous depression, sinusitis) and lumbar pain (spinal disk syndrome, neurinoma) was inconspicuous with respect to cell count, b l o o d - C S F barrier function and autochthonous synthesis of lmmunoglobulins. As negative controls, matched CSF sera from 20 patients with negative results in IFA and Western blotting and without history of a tick bite were examined.

Reagents Microtiter plates were supplied by Nunc, Wiesbaden, Germany. Reagents were obtained from the following companies: peroxidase-conjugated antibodies from Dianova, Hamburg, Germany; o-phenylenediamine (OPD), 4-chloro-l-naphthol, HzO 2, Tween 20, bovine serum albumin (BSA), N a z H P O 4, N a H 2 P O 4, NaCI and urea from Sigma Chemical Co., St Louis, MO; dry milk powder from commercial sources (Nestl6); molecular weight standards, SDS, Tris, glycine from Bio-Rad, Munchen, Germany Dilution buffer (DIL): 3% milk powder, 1% BSA, 0 1% Tween 20 in PBS

SDS-gel and lmmunoblottmg SDS-PAGE, transfer of electrophoretically separated proteins to nitrocellulose, and subsequent blotting was done according to previously described methods (Laemmh et al. 1973; Towbln et al. 1979) with the following modifications: B. burgdorfen lysates were electrophoresed in a discontinuous SDS gel, using a 4% acrylamide stacking gel and a 12% acrylamtde resolving gel. Gels were loaded with 100/~g antigen/7 cm width. Eiectrophoresls was performed at a constant voltage of 200 V for 1 h in a Bio-Rad Minicell. Proteins were electrophoretically transferred from the gel to a nitrocellulose (NC) membrane in phosphate buffer (pH 9.0) at 1 A for 1 h NC membranes were cut into 5-mm strips and used for testing After blocking for 1 h at room temperature with DIL the strips were incubated with 0.5 ml of CSF and sera, each adjusted to an IgG concentration of 10 mg/1. After 2 h the strips were washed 3 times with 0 1% Tween 2 0 / P B S and Incubated for 1 h at room temperature with peroxidaseconjugated goat anti-human IgG (1 1000). The NC stripes were developed by the addition of a freshly prepared solution of 0.05% 4-chloro-l-naphthol and 0.015% H 2 0 2 in PBS for 10 min. The reaction was stopped by washing the strips with distilled water. Intrathecal synthesis of antibodies to B. burgdorfen was assessed from the single or darker staining of antibody bands in CSF The appropriate dilutions of CSF and serum were determined with the purpose to identify different antibody reactions in both fluid and not to detect the total serum response to the agent.

Intrathecal synthesis of B. burgdorfen-specific antibodIes Specific antibody synthesis in the CSF was calculated by the ratio between the C S F / s e r u m quotients for specific antibodies (QAs) and the C S F / s e r u m quotients of total IgG (Qlg6) or albumin (QAIb)" The QAB was determined from IFA and ELISA titers and from specific antibody concentrations in ELISA. The upper reference values for the Antibody Index (AI = QAB/QtgG or QAb/QAIb) are dependent upon the respective assay. In case of local synthesis of polyclonal IgG in the CNS (IgGloc > 0) QIgG was replaced by Qhm with Qhm :0.8~/Q21b + 1 5 - 1 0

6 _ 1 8 ' 10 -3

In this procedure Qhm includes only immunoglobullns which statistically enter the CSF by diffusion at a certain b l o o d - C S F barrier permeability represented by the C S F / s e r u m ratios of albumin (QAlb) (Reiber and Lange 1991)

Indirect tmmunofluorescence assay (IFA) Screening for B. burgdorfen-specific antibodies in sera and CSF was done by indirect immunofluorescence assay according to standard techniques (Barbour 1984; Russel et al. 1984) Specimens were tested at the Department of Immunology in the Institute of Medical Microbiology of the University of Frelburg. Serum IgM tlters _> 1.48 (after adsorption of IgG) and serum IgG titers > 1 "32 (after adsorption with Treponema phagedents) were considered positive. Intrathecal synthesis of IgG antibodies to B. burgdorfen was calculated according to the mtrathecal Treponema palhdum antibody (ITPA) index for patients with neurosyphilis' lntrathecal synthesis of IgGantibodies to B. burgdorfen was assumed for values of AIIF g _> 2 (Prange et al. 1983).

A I [FA --

IFA CSF

IgGcsF

IFA . . . . .

IgG . . . . .

Preparanons for ELISA procedures Microtlter plates were coated overnight at 4°C with 50/~1 of B. burgdorfen (20 p~g/ml) in PBS. One isolate of B burgdorfert from ticks collected in the Freiburg (Germany) area (Z-136) was used in this study. Spirochetes were grown in modified Kelly's medium according to standard techniques (Barbour 1984), centrifuged at 10000 × g for 30 rain, and washed 3 times in PBS. For IFA spirochetes were fixed to glass slides by methanol fixation, for use in ELISA they were sonl-

67 cated for 5 mln and dissolved in 6 M urea to a final concentration of 2 m g / m l . The optimal antigen concentration for coating of ELISA plates had been determined in preliminary experiments by checkerboard titratlons of 2-fold dilutions of antigen and high-titered sera. Unspecific protein binding was blocked with 100 /.d DIL for 1 h. Specimens to be tested (50 ~zl) were incubated for 2 h. Specific antibody binding was detected by the addition of 50 /zl peroxidase-labelled goat anti-human IgG (1:20000 in DIL) for 1.5 h at room temperature. Samples were finally developed by adding 50 /.~l 1 m o l / l o-phenylenediamine (OPD) in 0.1 mol/1 citrate buffer (pH 50) with 0.4 /~l/ml H a O 2 / P B S to each well. The optical densities (OD) at 492 nm were measured with a Tltertek Multiscan reader (FlowLab, Meckenheim, Germany). OD ratios, which were 3 SD above the mean O D of control subjects, were defined as a positive antibody response to B. burgdorferi. Samples were tested in duplicate, and the mean value was calculated. If the two values differed more than 10% from the mean, the samples were retested Negative control samples were included on every plate

Concentration ELISA (EIA-1) This ELISA allows to quantify concentrations of antigen-specific antibodies in CSF and serum using a positive standard serum pool as a reference. The method of Relber and Lange (1991) was performed with slight modifications. As a reference antibody-positive sera from 15 patients were pooled, prediluted with DIL 1000-fold, and then diluted 2-fold with six serial dilution steps. Standard dilutions were chosen so that the corresponding absorbances were between 0.10 and 1.80 A. Samples and controls were diluted 1000- and 10 000-fold (serum) and 50- and 500-fold (CSF). Detection of specific antibodies was performed as described above. OD values >= 0.10 were regarded as positive. The standard values (reference sample) were evaluated in a log/log diagram ( O D / r e l a t i v e concentrations). The greatest measurable standard concentration (OD 1.8) was defined to be 1000 arbitrary concentration units. Final antibody concentrations of B. burgdorfen (Bb-conc) in CSF and serum were calculated by multiplication of arbitrary concentration units with the dilution factor (DF): EIA-I: Bb-conccsF • DFcs F Alc°nc = Bb-conc . . . . m ' DFserum :Qlgo Any AI >= 1.5 was considered indicative of a local antibody response to B. burgdorfen antigens (Reiber and Lange 1991)

Standard ELISA (EIA-2, EIA-3, EIA-4) (a) Serum and CSF were prediluted 1 : 500 and 1 : 50, respectwely. Specimens were then diluted in two steps and incubated for 2 h at room temperature. Detection of specific antibodies was performed as described above The endpoint titers of the patient's sera and CSF were fLxed at OD values that were 3 SD above the mean of ten negative controls (cut-off: OD > 0.15). Intrathecal synthesis of B. burgdorfen-specific IgG antibodies was calculated according to the following methods: EIA-2:

Alsto~-

Titercs F Titers,rum : Q I g G

Intrathecal antibody synthesis was considered for AI >= 2 (Ukkonen et al. 1981). (b) The ratio QIgG w a s replaced by the ratio album i n - C S F / s e r u m = QAII~: EIA-3: TitercsF Alsto,,ab = Tlterserum "QAIb Intrathecal antibody synthesis was considered for AI 1.9 (Klapper et al. 1981) (c) Ratios of optical densities at serum dilution 1:1000 and CSF dilution 1:100 were related to the ratio of albumin C S F / s e r u m (DF = dilution factor). EIA-4: ODcsF " DFcsF AlstoD = ODserum . DFseru m ' Q ~ b Intrathecal antibody synthesis was considered for AI >__ 2 (Stlernstedt et al. 1985).

Equivalence ELISA (EIA-5, EIA-6) CSF and sera were prediluted to an IgG concentration of 5 m g / I with dilution buffer and then serially diluted 1:2. The end-point dilution was determined graphically by interpolation between the two dilutions, giving an optical density reading above and below the cut-off level (OD __>0.1) of the ELISA. (a) The proportion of B. burgdorferi-specific antibodies per weight unit IgG in CSF and serum can be compared by a ratio expressed as: EIA-5: AI Eq..... =

End-point dilutioncs F End-point dilution . . . . .

tltercs F titerserum

Intrathecal synthesis of B. burgdorferi-specific antibodies was assumed if AIiden t was => 2 (Rehse-KiJpper et al. 1986; Biniek et al. 1988).

68

(b) Since the serum and CSF were diluted to the same concentration of IgG, the antibody ratio can also be calculated from the optical densmes. EIA-6'

ODcsF

A I Eqoo -- O D . . . . . .

Former studies (Hansen et al. 1990) have proposed any ratio > 1 to indicate intrathecal synthesis of antibodies. In this study the mean plus 3 s of AIBboD of 20 negative controls was 1 3, any ratio > 1.5 was considered positwe.

Results The findings from the immunofluorescence studies are shown in Table 2. Serum IgG-antibodles to B burgdorfert (before absorption with T. phagedems) were detected m all patients with neuroborreliosis and in all controls. Elevated titers of serum IgG-antibodles to B burgdorfert (> 32, after absorption of cross-reactive antibodies) were found m 2 9 / 3 3 patients with neuroborrehosis and m 3 / 1 5 controls (1" 128) with spinal disk syndrome but without history of a recent B burgdorfert infection In the latter 3 patients, who recovered after surgery because of a slipped disk, the CSF revealed no abnormalities. Low titers of serum

TABLE 2 INDIRECT

IMMUNOFLUORESCENCE

Patient

ASSAY

normalvalue

AND CSF ANTIBODY

T I T E R S T O BORRELIA BURGDORFERI

IgG

IgM

Serum

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

SERUM

CSF

Serum

0 a

+a

256 128 512 256 128 32 128 256 256 64 256 512 256 128 64 256 64 64 256 64 1024 4 096 2048 512 1 024 4096 1024 256 512 2048 256 1 024 128

64 64 256 128 64 16 64 128 256 32 64 128 256 128 32 128 16 16 64 16 256 256 256 64 256 512 1 024 256 512 1 024 64 1 024 64

128 32 64 4 32 16 8 256 32 2 256 256 16 512 16 16 256 32 16 8 128 64 64 8 128 64 128 32 128 1024 128 512 32

< 2 16 < 2 < 2 32 8 2 256 32 < 2 256 128 8 64 8 4 256 32 16 8 64 64 64 8 128 16 128 32 128 1024 32 512 16

< 32

< 2

< 48

< 2

" 0 = no adsorption with

CSF

+

0

T phagedents + = after adsorption with T phagedents

48 48 192 48 24 24 12 48 192 192 < 2 96 < 2 < 2 < 2 < 2 48 24 < 2 < 2 48 192 ~ 2 < 2 < 2 < 2 < 2 < 2 48 < 2 24 ~ 2 12

< < < <

2 2 2 2 4 32 4 64 8 64 < 2 64 <. 2 < 2 < 2 < 2 < 2 16 < 2 < 2 < 2 4 < 2 < 2 < 2 < 2 -~ 2 < 2 < 2 < 2 4 < 2 ~ 2

69 TABLE 3 D A Y - T O - D A Y P R E C I S I O N (CV%) OF THE B BURGDORFERI-SPECIFIC A N T I B O D Y I N D E X D E T E R M I N A T I O N IN SIX D I F F E R ENT ELISA M E T H O D S

CV%

EIA-1

EIA-2

EIA-3

EIA-4

EIA-5

EIA-6

6.2

16_6

16 6

5_5

16 6

75

IgG-antibodles ( > 32 after absorption of cross-reactive antibodies), however, did not rule out the possibility of a CNS infection as demonstrated m 6 patients with neuroborreliosis (nos. 6, 10, 15, 17, 18, 20). One patient (no. 10) had elevated serum IgM titers to B. burgdorferi, a second patient (no. 17) had serum IgM-antibodies just at the cut-off level (1 : 48) None of the controls had IgM-antibodies to B burgdorfen in CSF and serum. Only 12 of 33 patients with neuroborreliosis had elevated B. burgdorfen-speciflc IgM-antibodies m serum ( > 48). With regard to elevated serum antibodies (IgG a n d / o r IgM) the diagnosis of neuroborreliosls could be

supported in 30/33 patients, 2 / 3 0 patients had tlters at the cut-off level. Western blotting performed with B. burgdorferi lysate, using CSF and serum diluted to the same IgG concentration, was the reference method for measuring the mtrathecal antibody reaction to mdividual B. burgdorferi antigens. An autochthonous synthesis of specific IgG-antlbodies to individual B. burgdorfen antigens, as evidenced by a more intense staining of single antigen bands in the CSF compared to serum was demonstrated in all patients. Fig. 1 shows a typical finding from a patient with chronic neuroborreliosls

TABLE 4 I N T R A T H E C A L SYNTHESIS OF B BURGDORFERI-SPECIFIC IgG-ANTIBODIES Results from redirect lmmunofluorescence assay (IFA) and different ELISA methods. Patzent 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 normal value

IFA

EIA-1

EIA-2

EIA-3

EIA-4

EIA-5

125 2 83 1 112 47 13_4 41 6 31 1 40.6 86 16 62 4 16 5 9_1 26 6 10 5 52 20 8 33 0 27 1 73 57 18 47 10 31 2 39 10 4 19.5 83 11 1 41_6 10 0 81 2

27.6 88 1 85 4_4 43 0 47 2 10 8 14 6 32 25 60 2 11_0 84 64 56 45 19 0 36 2 21_7 11 8 65 28 36 25 10 2 33 41 17 5 47 10_4 35 8 60 24 6

25 1 16 0 90 38 21 5 33 3 40 0 8.0 52 1.3 98 7 26 6 14 7 53 0 10 4 4_1 18 0 57 0 43 8 23 4 90 31 29 35 25 0 31 8.3 39 66 35 5 33 0 16_0 33 4

16 6 11 1 70 20 17 0 23 9 20 2 56 5.2 09 72 0 18 9 88 41 0 45 30 92 39 4 24 4 14 3 64 21 19 25 15 8 29 59 2_6 49 27 5 22 4 11 3 21.2

10_7 66 67 54 10 1 13 6 95 64 42 1.3 88 0 109 99 54 55 37 32.0 28 0 17 6 100 70 72 8.8 30 15 0 40 71 5.2 50 2_8 60 37 27 0

2 16 8 4 4 16 32 2 2 2 64 32 32 32 8 4 8 32 16 16 8 4 4 2 16 8 4 4 4 8 16 16 16

< 2

< 15

<2

< 19

<2

<2

EIA-6 3_3 60 45 38 4.4 129 15 0 2.3 21 17 34 89 9.1 22 8.9 20 5.0 15 8 87 13 0 28 1_5 22 19 28 2_6 20 24 25 19 2.8 27 7_2 < 1.5

70 kDo

B burgdorfcrJ

%or

Ser

%or LSF

her L%F

97 67

Iif

45

b

m

31

mk~

i

21

•

AI < 15

. AI > 1 5

Fig 1 Western blotting of B burgdorfen protein CSF and serum were adjusted so thai concentrations of lgG were equal (10 m g / L ) A negative ( e ) and posltwe ( ~ ) serum was included m every assay Control patient (AI < 1 5 ) with an anamnest~c response to B hurgdorfen diffusion of serum antibodies into the CSF Patient with neuroborrehosLs (A1 > 1 5) The more intense staining of individual bands m the CSF indicates mtrathecal synthesis of B burgdor]enspecific antlbod~e,~

(AI .... > 1 5) and an lntrathecal synthesis of IgG to various B. burgdorfen proteins. In 3 / 1 5 controls with elevated IgG-antlbodies m serum (IFA), Western blotting revealed IgG-antlbodies to various B burgdorfert proteins in serum and CSF. As antibody bands were of similar staining intensity in both fluids (AI .... < 1 5) no lntrathecal IgG response to B burgdorfen could be assessed from this finding (Fig. 1). The day-to-day variation (CV%) of the AI for the SEXdifferent E L I S A methods was analyzed in 10 different series (posmve control) The detailed data are listed in Table 3 The CV% ranged from 5.5% (EIA-4) to 16 6% (EIA-2; EIA-3, EIA-5) Calculation based on B burgdorfen-speclfic antibody titers in IFA demonstrated lntrathecal synthesis in 30/33 samples The AIIFA ranged from 3 9 to 125.2 in patients with positive results. In 3 patients (nos. 10, 22, 24) the AIIFA was negative, although calculations from E L I S A values indicated lntrathecal synthesis of speofic antibodies. Locally produced antibodies in these three latter patients were mainly reactive with p39 and p41, as shown by Western blotting Patients nos 10 and 22 had elevated IgM-antlbodies to B burgdorfen in serum in addition The occurrence of IgG-antibodles to B_ burgdorfert m the CSF and sera was demonstrated by different ELISA procedures in 3 3 / 3 3 samples In most of the patients CSF antibodies were calculated to be produced lntrathecally Using the "Concentration E L I S A " (EIA-1) elevated antibody indices were determined in 3 3 / 3 3 samples

The AI ranged from 2 5 to 88.1 The same sensitivity was obtained with the "Equivalence ELISA", either by calculating the antibody index from tlters (EIA-5, AI 2.0-64) or from optical densities (EIA-6, A1 1 5-15 8) The "Standard EL1SA", with positive results in 32/33 samples, was almost as sensitive when the AI was calculated from antibody titers and QI~(; (EIA-2, AI 2.9-98.7) or from antibody tlters and QAJD (EIA-3, AI 20-72.0). Similar results (32/33 patients) were obtained with this ELISA procedure if the AI was calculated from optical densities at serum dilutions of 1 1000 and CSF dilutions of 1_ 100 and the QAJb instead of QIgG (EIA-4, AI 2 8-88.0). The data are hsted m detail in Table 4. In the controls IgG-antibodles to B. burgdorfi, rt were detected by various E L I S A procedures (E1A 1-6) m 13 serum and three CSF samples An lntrathecal synthesis of B. burgdorfen-speclfic antibodies could not be derived either by calculation of the AI from antibody tlters (IFA or ELISA), or from antibody concentrations or from optical densities (ELISA)

Discussion Determination of a specific intrathecal antibody reaction represents a useful diagnostic tool to confirm the diagnosis in suspected cases of CNS infection The aim of our study was to compare various technical procedures and calculations methods for the investigation of the lntrathecal immune response to an infectious agent Due to the expenditure of time and labor in performing the Western blot, this assay is not suitable for the examination of a large number of samples m routine diagnostic work. The autochthonous antibody response in the CSF however, might be directed against other antigens of the infectious agent (antigen processing and presentation by astrocytes) (Fierz et al 1985) than the serum response In this case the Western blot would detect an autochthonous immune reaction, even when the relative amounts of specific IgG antibodies in the CSF and serum are very similar and calculation methods fail to indicate specific intrathecal synthesis Therefore, the Western blot, employing identical concentrations of IgG in CSF and serum, was chosen as a reference method to study the antibody response to individual B burgdorferi antigens. An autochthonous synthesis of specific antibodies was assumed if the CSF revealed a more intense staining than the serum with at least two proteins of B. burgdorfen. Of the two proteins, one had to be a typical or probably specific antigen of B burgdorfert (p22, p31, p34, p41, pl00). Referring to this criterion no lntrathecal immune reaction to B burgdorfert could be demonstrated in 15 controls, and in 20 samples from patients with MS who

71 revealed an autochthonous synthes~s of total IgG. Therefore, this procedure may be recommended for patients with suspected neuroborreliosls but borderhne results In AI calculations. If antibody bands were of similar staining intensity in both fluids, antibodies in the CSF were interpreted to result from diffusion from the serum into the CSF. This was shown in three controls with a spinal disk syndrome and an anamnestic response to B. burgdorfert. The finding of B. burgdorfen-specific IgG-antlbodies in the CSF by Western blotting but not by IFA indicates the higher sensitivity of the Western blot Results from Western blotting are hard to compare with findings from quantitative examinations in IFA and ELISA. In Western blotting, the extend of antibody reactions is not only expressed by the number but also by the staining intensity of individual bands. Quantification of the latter criterion requires a scanner which was not available in this study. Assessment of results in Western blotting by three investigators revealed that the occurrence of four or more intense CSF IgG bands in Western blotting was frequently assooated with AI values > 5 in the "Concentration ELISA" IFA and ELISA are the most practical techniques for investigating the antibody response to an infectious agent. As results in IFA may vary between different laboratories and absolute measurements cannot be easily obtained, calculation of AI from IFA titers is not usually recommended. In our study determination of AI from IFA titers had a lower sensitivity (90.9%) in detecting a specific antibody synthesis in the CNS than calculation of this criterion from ELISA values. In the 3 patients (nos. 10, 22, 24) without intrathecal synthesis of B. burgdorfert-specific IgG-antibodies, as calculated from IFA-tlters, two had significantly elevated titers of IgM-antibodles to B. burgdorferi in serum ( 1 192). The various ELISA techniques and calculation methods did not differ s~gnificantly in respect of the sensitivity in detecting a specific lntrathecal immune response to B. burgdorfen. However, these assays differed in their day-to-day variation (CV). A high CV was obtained if the AI was calculated from end-point dilutions (antibody titers) of samples (EIA-2, EIA-3, and EIA-5) In each of these assays only one of ten determinations differed from the average. According to the twofold dilution steps the resulting AI value was doubled or divided by two. If the AI was calculated from OD (EIA-1, EIA-4, and EIA-6), the CV was reduced to 5.5-7.5%. In these assays antibody concentrations were determined always from the same ddution step. Probably the most commonly used ELISA procedure is the "Standard ELISA". Neither the use of OD (EIA-4) instead of titers (EIA-2, EIA-3) nor the replacement of the QIgG with the QA~b (EIA-3, EIA-4) significantly influenced the sensitivity in determining a

specific autochthonous antibody reaction. However, in patients with very high antibody levels, determination of the AI from a single dilution step (EIA-4, EIA-6) in CSF and serum might not be accurate. As the OD reflecting the antibody concentration of a sample corresponds to a sigmoid curve in ELISA, a linear rise of optical densities occurs only In a limited range of antibody concentrations Very high or low antibody concentrations will not differ significantly with respect to optical densities and therefore gwe rise to false calculations. EIA-4 is recommended only for samples with optical densities within the linear range of the sigmold a n t i b o d y / O D curve. Calculation of the AI from "Standard ELISA" titers and Qlgc will provide satisfactory results in patients with high levels of locally produced specific antibodies. For the chnician who is sent specific antibody tlters from a microbiological laboratory and protein data from another chnical laboratory the determination of the AI from "Standard ELISA" t~ters still represents a suitable method in most cases The most sensitive tests however were the "Concentration ELISA" and the "Equivalence ELISA", both specially performed to investigate the lntrathecal immune response to an infectious agent (Hansen et al. 1990; Reiber and Lange 1991). Calculanon of the AI from OD in the Equivalence ELISA requires only one dilution of CSF and serum, which mimmizes consumption of anUgen. Adjustment of CSF and serum to the same IgG-concentratlon might, however, provide false negative results in calculating the AI if there is a sigmficant mtrathecal synthesis of antibodies with specificity for different infectious agents (i.e. HIV). Due to the additional synthesis of lmmunoglobulins with other speclficities the CSF will be diluted disproportionately. This disadvantage is avoided by the "Concentration ELISA", determining specific antibody concentrations in CSF and serum from the OD of a reference-serum and using the Ql~m to calculate the AI. In most cases simultaneous analysis of two dilutions of speomens is sufficient to obtain OD in the range of the reference serum. In our opinion this procedure represents a very accurate method to demonstrate intrathecal synthesis of specific antibodies, even in patients with a low autochthonous synthesis of speofic antibodies. Values of the antibody index determined from different calculation methods or ELISA procedures might vary greatly in the same patient Therefore, assessment of a specific antibody synthesis over the course of an infection should always be done by the same method As individual data from different assays and calculation methods are hard to compare, the main purpose of this study was to investigate the sensitivity of different methods in determining the AI to confirm the chnical diagnosis m suspected cases of neuroborrello-

72 sis. In a considerable proportion of patients the diagnosis of neuroborreliosis will be established by a preceding tick bite a n d / o r erythema mlgrans and, the typical clinical finding of a painful menmgoradicuhtis, the presence of elevated IgM a n d / o r I g G serum antibodies to B burgdorfen and the demonstration of a lymphocytic pleocytosis in the CSF in accordance with a moderate impairment of the b l o o d - C S F barrier and an autochthonous synthesis of IgM, IgG and sometimes IgA This was the case in 28/33 patients of our study. In five patients however serum tlters to B. burgdorfert were borderline or inconspicuous with respect to a recent tnfect~on with this agent. In these patients diagnos~s of neuroborreliosls was finally made by the calculation of an elevated B burgdorfen-spec~fic antibody index. Acknowledgments This study was supported by grant no 01K1 9001/0 from the German Federal Ministry for Research and Technology (Bundesmmlstermm fur Forschung und Technologie) We thank Prof Vogt, Institute of Medical Microbiology, University of Frelburg, for donating somcates of B burgdorferi, and Glsela Flelg for expert and skillful technical assistance Prof_ Batsford is gratefully acknowledged for help m editing the manuscript

References Barbour, A G (1984) Isolation and cultivation of Lyme disease spirochetes Yale J. BIol Med, 57 521-525 Bimek, R , Bartholome, M , Schulz, M , Lehmann, H J , Gesemann, H , Schelermann, N , Brockmeyer, N H , Derks, M_ and K Balzer (1988) Intrathecal production of HIV antibodies m suspected AIDS encephalopathy J Neurol, 235 131-135 Dornes, R , Kaiser, R , Schwender, S, Imrlch, H , Pohl-Koppe, A and K Dorries (1991) Recent aspects on the dmgnosis of wral refections m the central nervous system Lab Med, 15 99 Flerz, W , Endler, B, Weckerle, H. and A_ Fontana (1985) Astrocytes as antigen-presenting cells 1 Induction of la-antigen expression on astrocytes by T-cells via immune interferon and its effect on antigen presentation J. Immunol, 134 3785-3793 Hansen, K, Cruz, M and H Link (1990) Ohgoclonal Borreha burgdorfen-speciflc IgG antibodies in cerebrospmal fluid m Lyme neuroborreliosis J Infect Dis_, 161 1194-1202 Kaiser, R. (1991) Affinity lmmunoblottmg. Rapid and sensltwe detection of ohgoclonal IgG, IgA and IgM in unconcentrated CSF by agarose isoelectnc focusing J_ Neurol Sci., 101:216-225 Kaiser, R , Domes, R , ter Meulen, J , Kadenya, J.J, Pollath, M, Flelscher, K and V ter Meulen (1990) Serological evidence of

HIV refection of the central nervous system in African patients with acqmred lmmunodeflclency syndrome Eur Neurol, 30 27-31 Klapper, P E , Lamg I and M Longson (1981) Rapid non-invast'~e diagnosis of herpes encephahtls Lancet, u 607-609 Laemmh, U K and M Favre (1973) Maturation of the head ol bacteriophage T4_ 1 DNA packaging events J Mol Blol, 80 575-99 Mautner, V F , Glttermann, M, Freltag, V and E Schneider (1991)) Zur Epldemlologie der Borreha burgdorfer~-Infektlon Abhanglgkelt der Durchseuchungsrate von den serolog~schen Bestlmmungsverfahren Nervenarzt, 61 94-97 Prange, H W , Moskophldls, M , Schlpper, H I and F Muller (1983) Relationship between neurological features and mtrathecal synthesis of IgG anhbodles to Treponema palhdum in untreated and treated human neurosyphths J_ Neurol, 230 241-252 Rehse-Kupper, B and R Ackermann (1986) Demonstration of locally synthesized Borreha antibodies m cerebrospmal fluid Zbl Bakt Hyg A , 263' 407-411 Reiber, H O (1988) Aktuelle Methoden der Llquoranalytlk Lab Med, 12 101-109 Relber, H O and K Felgenhauer (1987) Protein transfer at the blood cerebrosplnal fluid barrier and the quant~tatlon of the humoral ~mmune response w~thm the central nervous system Chn Chlm Acta, 163 319-28 Relber, H O and P Lange (1991) Quantification of virus-speclhc antibodies m cerebrosplnal fluid and serum Sensitive and specific detection of antibody synthesis In brain Chn Chem, 37 1153-1160 Relk, L, Jr (1991) Lyme Disease and the Nervous System. Thieme Med_ Publishers, New York/Stuttgart, pp 52-110 Russel, H , Sampson, J_S and G P Schmld (1984) Enzyme-linked lmmunosorbent assay and indirect immunofluorescence assay for Lyme disease J Infect. Dis_, 149 465-470 Steere, A C, Berardl, V P, Weeks, K_E, Logigian, E L and R Ackermann (1990) Evaluation of the lntrathecal antibody response to Borreha burgdorfen as a dmgnostlc test for Lyme neuroborrehosls J Infect Dis 161, 1203-1209 St~ernstedt, G T, Granstrom, M, Hederstedt, B and B_ Skoldenberg (1985) Dmgnosis of spirochetal meningitis by enzyme-hnked immunosorbent assay and redirect ~mmunofluorescence assay m serum and cerebrosplnal fluid J Chn Mlcrobiol, 21 819-25 Towbm, H , Staehlehn, T and J Gordon (1979) Electrophoret~c transfer of proteins from polyacrylamlde gels to mtrocellulose sheets Precedures and some apphcatlons Proc_ Natl Acad_ Set USA, 76 4350-4 Ukkonen, P , Ganstrom, M L, Rasanen, J_, Salonen, E M and K Penttmen (1981) Local production of mumps lgG and IgM antibodies in the CSF of meningitis patients J Med Vlrol, 8 257-265 Weber, T, Buttner, W and K Felgenhauer (1987) Ewdence lot different immune response to HIV In CSF and serum Kiln Wochenschr, 65 259-263