APMIS 111: 1053–9, 2003 Copyright C APMIS 2003Printed in Denmark . All rights reserved

ISSN 0903-4641

Improved serodiagnosis of early Lyme borreliosis:Immunoblot with local Borrelia afzelii strain

VILMA Lj. JOVICIC,1 EDITA M. GREGO,2 BRANISLAV L. LAKO,3 BLAGOJE M. RISTOVIC,3

ZORICA A. LEPSANOVIC3 and NOVICA T. STAJKOVIC3

1Institute of Public Health of Serbia ‘‘Dr Milan Jovanovic Batut’’, Department for Microbiology, Belgrade,2Institute of Molecular Genetics and Genetic Engineering, Belgrade, and 3Institute for Microbiology, Military

Medical Academy, Belgrade, Serbia and Montenegro

Jovicic V, Grego E, Lako B, Ristovic B, Lepsanovic Z, Stajkovic N. Improved serodiagnosis of earlyLyme borreliosis: Immunoblot with local Borrelia afzelii strain. APMIS 2003;111:1053–9.

To improve the serodiagnosis of Lyme borreliosis (LB) the performances of four tests were evaluated.An indirect immunofluorescent assay based on Borrelia burgdorferi s.s. , enzyme-linked immunosorb-ent assay (ELISA) based on local isolates of Borrelia afzelii and B. burgdorferi s.s., and immunoblot(IB) of B. afzelii were prepared. The serum panels contained 214 serum samples: control group (nΩ120) and patients at different stages of LB (nΩ94). The specificity of IB was 96%, of in-house ELISA93%, and of IFA 89%. In early LB the sensitivity of IFA was 36%, ELISA 67%, and IB 93%. In late-stage LB the sensitivity was: 72% for IFA, 80% for ELISA, and 94% for IB. Comparison of in-houseand Behring ELISA showed that the sensitivity of the serological assay could be increased when thetest was based on local Borrelia strains. IgM and IgG antibodies from sera of patients with early andlate LB most frequently demonstrated reactivity to OspC. The other significant proteins in early LBwere: p39, p41 in IgM IB, and p83/100, p39, Osp17 in IgG IB; in late LB: p39, p41 in IgM IB, andp83/100, Osp17, p21 and p43 in IgG IB. Using IB based on local B. afzelii isolates improves theserodiagnosis of early LB in our geographical region.

Key words: Lyme borreliosis; serodiagnosis; indirect immunofluorescent assay; enzyme-linked immu-nosorbent assay; immunoblot.

Jovicic Vilma, Institute of Public Health of Serbia ‘‘Dr Milan Jovanovic Batut’’, Department for Micro-biology, 5, Dr Subotica St, 11000 Belgrade, Serbia and Montenegro. e-mail: vilma/batut.org.yu

Lyme borreliosis (LB) is a multisystemic dis-order caused by the tick-borne spirochete Bor-relia burgdorferi (1). The disease may involvemany organs, most commonly the skin, nervoussystem, heart and joints. The basic clinicalcharacteristics of LB are similar worldwide, butthere are regional variations, primarily betweenthe illness found in America and that found inEurope and Asia (2). The diagnosis is based onclinical signs, epidemiological information anddetection of specific antibodies in blood andcerebrospinal fluid. Direct detection of the

Received March 3, 2003.Accepted August 28, 2003.

1053

pathogen by culture or polymerase chain reac-tion (PCR) usually plays a minor role in theroutine diagnosis of LB.

Commonly used serological assays include in-direct immunofluorescence assay (IFA), en-zyme-linked immunosorbent assay (ELISA) andimmunoblot (IB) or Western blot (WB). ELISAis considered more sensitive and less subjectivethan IFA (3). Mitchell et al. (4) compared theresults of four immunoserological assays andfound that immunoglobulin M (IgM) IFA hadthe highest sensitivity. IB is used as a confirma-tory test (5).

In the United States, the diagnosis is usuallybased on recognition of characteristic clinical

JOVICIC et al.

findings, a history of exposure in an area wherethe disease is endemic, and except in patientswith erythema migrans, an antibody responseto B. burgdorferi by ELISA and WB is inter-preted according to the criteria of the Centersfor Disease Control and Prevention (2, 6). Theantigenic heterogeneity among European iso-lates complicates the comparability and stand-ardization of assay systems. Differences in theregional distributions of borrelial species mayfurther influence the preferential reactivities ofsera from patients with LB (7). The definitionand use of optimal antigens are critical for theperformance and clinical utility of B. burgdor-feri serological assays. A few different antigensor antigenic fractions have been evaluated sofar. Hansen et al. showed that ELISA based onpurified flagella antigen was superior to soni-cated antigen (8). In a study from Finland, Oksiet al. concluded that ELISA based on sonicateantigen was more sensitive than flagella-basedELISA (9). A report from Sweden demon-strated the potential significance of strain selec-tion in assay design (10).

Several third-generation tests that use recom-binant spirochetal proteins or synthetic peptideshave shown promising results (11–13).

To date, no single sensitive and highly specificlaboratory test is available.

The aim of the present study was to investi-gate whether using tests (IFA, ELISA, IB)based on local Borrelia isolates improved theserological diagnosis of LB in our geographicalregion.

MATERIALS AND METHODS

StrainsIn the present investigation four isolates of B. burg-

dorferi were used: strain B31(ATCC 35210) and ourisolates named K1, K2, Z1 from Apodemus flavicollis,Belgrade. Borrelia was grown in BSK II medium at35 æC (14). Whole-cell lysates of our isolates wereanalysed by sodium dodecyl sulfate polyacrylamidegel electrophoresis (SDS-PAGE) and their antigeniccharacteristics were compared with referent strainB31, prepared in the same manner. According to theprotein profile of B. burgdorferi our isolates K1 andZ1 belong to Borrelia afzelii (15). Isolate K2 belongsto B. burgdorferi s.s. The immunodominant proteinsof our isolates were not identified by monoclonalantibodies because a panel of specific monoclonalantibodies was not available.

1054

IFAAntigen for in-house IFA was prepared from strain

B31 (previous investigation showed that IFA basedon K1 had the same sensitivity and specificity). Spiro-chetes were grown for 3 to 5 days and then centri-fuged at 10000 g for 30 min, washed three times in0.15 M phosphate-buffered saline (PBS) ;0.005... MMgCl2, pH 7.4 and fixed on glass slides by cold ace-tone fixation.The presence of IgM and IgG anti-bodies was determined in sera diluted 1:40 and 1:80,respectively. Fluorescein isothiocyanate-labelled anti-human IgM and IgG (Inep-Zemun, Belgrade, Serbiaand Montenegro) were used as secondary antibodies.Positive and negative control serum samples were in-cluded in each test. The endpoint was defined as thehighest titre at which all spirochetes still showed abright fluorescence.

ELISAAntigen for in-house ELISA was prepared from a

mixture of our isolates K1, K2 and Z1 by the stan-dard method of Craft et al. (16) with modifications.The spirochetes were pelleted by centrifugation at10000 g for 30 min, washed twice with 0.15 M PBS;0.005... M MgCl2, pH 7.4, and then washed twicewith 0.15 M PBS. The final pellet was resuspended in0.15 M PBS with 1 M NaCl, and then sonicated onice with a cell sonicator (MSE 150 watt, MSE Scien-tific Instruments, Sussex, England). After a final cen-trifugation at 21000 g for 30 min the supernatant wasused as antigen. Protein concentration was deter-mined by the method of Lowry et al. (17). The opti-mal coating concentration was found to be 10 mg/ml for both IgM and IgG antibodies. Linbro 96-wellmicrotitration plates were coated with 0.1 ml solutionof the spirochetal proteins in 0.05 M sodium carbon-ate, pH 9.6 overnight at room temperature. The wellswere blocked with 0.15 M PBS, pH 7.4; 5% nonfantdry milk (NFM); 0.05% Tween 20, and incubatedwith patient sera serially diluted (1:100, 1:200 forIgM; and 1:200, 1:400 for IgG). Bound antibodieswere detected with anti-human IgM and IgG conju-gated to peroxidase (Bio-Rad). The substrate wasOPD (o-phenylenediamine). The reaction wasstopped with 2 M H2SO4, and the absorbances weremeasured at 492 nm (Microwell system-reader 510‘‘Organon Teknika’’). The sera were tested in dupli-cate. Positive, negative and blank controls were in-cluded in each test.

Determination of cut-off values and interassayvariation

The cut-off value was calculated by adding threestandard deviations (SD) to the mean of the absorb-ance for the group of healthy blood donors with nohistory of disease . The cut-off value was ODΩ0.360for IgM ELISA and ODΩ0.730 for IgG.The interas-says coefficients of variation (CV) were 12% for IgMand 10% for IgG. All samples with absorbance valuescut-off∫CV were retested.

IMPROVED SERODIAGNOSIS OF EARLY LYME BORRELIOSIS

Behring ELISA is a commercially available test forLB based on B. afzelii, strain PKo.

SDS-PAGE and immunoblottingElectrophoresis in 12.5% polyacrylamide gels was

performed following standard procedures (18). Pro-tein standards (Pharmacia Biotech) were included ineach gel. Proteins were transferred to a polyvinyl di-fluoride membrane (PVDF). The membranes wereblocked overnight in 0.05 M Tris-buffered saline(TBS, pH 7.6), containing 5% NFM; 0.5% Tween 20and 0.25% gelatin and incubated with patient sera ata dilution of 1:100. The secondary antibodies wereanti-human IgM and IgG conjugated to alkalinephosphatase (DAKO, Copenhagen, Denmark). Thereactivity was visualized by substrate p-nitro bluetetrazolium/5-bromo-4-chloro-3-indolyl phosphate(NBT/BCIP). The reaction was stopped by 0.5 MEDTA in PBS.

Interpretation of IBThe following interpretation criteria were used for

IB positivity: a positive IgM IB was defined by serumreactivity with at least one band corresponding to thefollowing: p39, OspC, p17 or a strong p41 band; apositive IgG IB was defined by serum reactivity withat least two bands of the following: p83/100, p58,p43, p39, p30, OspC, p21, p17, p14 (19). The positivecontrols were selected from patients with early andlate disease, and were reactive with all diagnosticallyrelevant bands. The use of internal controls in eachassay allowed semiquantitative evaluation.

Study samplesSix groups of serum samples were tested: 50 sera

from healthy blood donors 20 to 60 years of age withno history of disease residing in an area where LB isendemic; 30 sera from healthy blood donors with nohistory of LB or tick bites (age 18 to 27 years), per-sons from the other area of Serbia without data ofendemicity; 40 sera from patients with syphilis, rheu-matoid arthritis (RA) and systematic lupus ery-thematosus (SLE) with no known exposure to ticksand positive by Treponema pallidum hemagglutin-ation (TPH) test , rheumatoid factor (RF) and anti-nuclear antibodies (ANA), respectively; 40 sera frompatients with erythema migrans (EM): 31 collected 2to 6 weeks and 9 collected 2 to 6 months after the tickbite; 4 sera from patients with Lyme carditis (LC); 50sera from patients with late LB clinically defined asneuroborreliosis (NB) or as Lyme arthritis (LA). Theclinical diagnosis of LB was based on the EUCALBclinical case definitions LB (20).

All sera were tested for the presence of IgM andIgG antibodies by IFA, ELISA and IB.

Sensitivity of in-house ELISA was compared withthe sensitivity of Behring ELISA using sera from thepatients with EM and late manifestations of LB.

1055

Statistical analysisSpecificity (Sp) was estimated as the relative fre-

quency of negative test results in non-infected indi-viduals: SpΩtrue negative/(true negativeπfalse posi-tive).

Sensitivity (Se) was estimated as relative frequencyof positive results in infected individuals: SeΩtruepositive/true positiveπfalse negative.

RESULTS

Specificity and sensitivity of testsTo determine test specificity we tested sera

from persons with no previous history of LB:healthy blood donors, and potentially cross-re-active samples from patients with syphilis, RAand SLE. IB was the most specific assay with aspecificity of 96%. The specificity of ELISA was93%, and of IFA 89%. False-positive reactionsdetected by IFA and ELISA were defined by IB.

Sensitivity was determined by testing serumpanels composed of the sera from patients withearly and late manifestations of LB (Table 1).Sensitivity of IFA, ELISA and IB in the serodi-agnosis of early Lyme disease was 36%, 67%,and 93%, respectively, and in the serodiagnosisof late Lyme disease 72%, 80% and 94%, respec-tively. IgM and IgG antibodies were assayed to-gether.

Testing the same sera from patients with LBthe sensitivity of in-house ELISA was 86% (37/43), and of Behring ELISA 77% (33/43).

IgM and IgG antibody reactivity detected by IBIB reactivity of sera from healthy blood do-

nors.The cross-reactive antibodies to strain K1antigens were analysed by IB. In sera fromhealthy blood donors residing in an area whereLB is endemic, IB detected 14 positive (28%)and 36 negative samples. IgG antibodies reactedmost commonly with heat shock proteins (7sera); flagellar protein 41 kDa (4 sera); 13 kDaprotein (3 sera); and 32 kDa, 28 kDa,18 kDa,16 kDa, 15 kDa proteins (2 sera). In contrast,IgM antibodies from these samples did not rec-ognize polypeptides of strain K1 by IB.

IB reactivity of clinically defined LB seraThe patterns of IgM and IgG antibodies

binding to the polypeptides of strain K1 weredetermined in sera from patients with erythema

JOVICIC et al.

TABLE 1. Results of in-house tests for serodiagnosis of LB

Stage of No. of sera IFAa ELISAa Immunoblota

disease tested positive negative positive negative positive negativeEMb 40 13 27 26 14 37 3LCb 4 3 1 3 1 4 0LBc 46 33 13 37 9 43 3total 90 49 41 66 24 84 6a Either IgM or IgG antibodies scored positive.b Sera from patients with EM and sera from patients with LC.c Only 46 sera from patients with late manifestations of LB were used to determine sensitivity of tests.

migrans (EM) and sera from patients with latemanifestations of Lyme disease (Table 2).

IB detected 37 positive sera from patientswith EM (Fig. 1), and 3 negative. IgM and IgGantibodies specific for 23 kDa protein – ‘‘OspC’’were detected in 21 (53%) and 22 (55%) sera,respectively. The high reactivity to p39 wasevident in IgM (25%), and IgG IB (35%). Onenegative sample had reactivity of IgM against35 kDa protein – ‘‘OspB’’.

Testing of sera from patients with late mani-festations of Lyme disease by IB detected 47positive and 3 negative samples. The antibodiesfrom positive sera demonstrated the strongestreactivity to ‘‘OspC’’ in both IgM (34%) andIgG IB (50%).

TABLE 2. IB reactivity of IgM and IgG antibodies from sera of patients with EM and late LB

Number (%) of sera with IgM and IgG reactivity to protein bandSpecific proteins Erythema migrans (40)a Late Lyme borreliosis (50)a

IgM IgG IgM IgG83/100 3 (8%) 17 (43%) 3 (6%) 24 (48%)p58 – 3 (8%) 1 (2%) 2 (4%)p43 4 (10%) 2 (5%) 4 (8%) 14 (28%)p41 5 (13%) 7 (14%)p39 10 (25%) 14 (35%) 8 (16%) 9 (18%)OspBb 5 (13%) 10 (25%) 2 (4%) 8 (16%)OspAb 3 (8%) 5 (13%) 2 (4%) 7 (14%)p30 1 4 (10%) 4 (8%) 5 (10%)OspDb 3 (8%) 10 (25%) 3 (6%) 10 (20%)OspFb – 1 – 3 (6%)OspC 21 (53%) 22 (55%) 17 (34%) 25 (50%)p21 2 (5%) 6 (15%) 6 (12%) 15 (30%)OspEb 2 (5%) 9 (23%) 3 (6%) 7 (14%)Osp17 4 (10%) 11 (28%) 3 (6%) 19 (38%)p14 3 (8%) 4 (10%) 2 (4%) 7 (14%)a Total number of sera tested from patients with EM and late disease.b Specific proteins not included in the criteria for IB positivity: 32 kDa protein – ‘‘OspA’’, 35 kDa – ‘‘OspB’’,

28 kDa – ‘‘OspD’’, 26 kDa – ‘‘OspE’’, and 19 kDa – ‘‘OspF’’.

1056

DISCUSSION

We prepared the ELISA and IB based on localBorrelia s.s. and B. afzelii strains. Borrelia gari-nii occurs in Europe and Asia, but has not beenisolated in the Belgrade area.

Analysis of the test specificity showed that IBwas the most specific assay (96%). In order toqualify for use as a confirmatory assay in step-wise diagnostics, the IB must yield a specificityof at least 95% (5). ELISA showed higher speci-ficity compared to IFA .

Sensitivity of the test varied with stage of thedisease. In early LB the sensitivity of the ELISAwas almost twice that of IFA. The difference insensitivity of the ELISA and IFA in late-stage

IMPROVED SERODIAGNOSIS OF EARLY LYME BORRELIOSIS

Fig. 1. Positive immunoblot strips of K1 strain incubated in nine sera from patients with EM (from the left:first line reactivity of IgM, and second reactivity of IgG). The molecular mass standards are indicated inkilodaltons (14.4–94) far left with negative control serum.

disease was about 8%. IB was the most sensi-tive assay in early (93%) and in late disease(94%).

Engstrom et al. (21) detected 72% of personswith physician-diagnosed EM by using wholecell sonicate ELISA to test sera collected 1–5days after treatment. When IB analysis wasdone, sensitivity for early disease after treat-ment was ∂80%.

The sensitivity of the in-house ELISA wassimilar for sera from patients with EM, but oursamples were collected 4 to 6 weeks after thetick bite. Our ELISA showed higher specificityand sensitivity than IFA and seems to be a goodcandidate for screening assay for serodiagnosisof LB.

In-house ELISA showed somewhat highersensitivity than ELISA Behring. Further exami-nations will reveal the significance of using thelocal Borrelia strains as an antigen in ELISA.

The high sensitivity of our IB in early andlate disease suggests that the choice of local B.afzelii strain for antigen preparation for IB af-fects the level of reactivity. Due to its specificity

1057

and sensitivity, the IB should be applied whenELISA is negative, especially in early LB.

Hauser et al. (22) showed that IB with PKo(B. afzelii) was the most sensitive, and this strainis recommended for use in WB for the serodi-agnosis of LB throughout Europe.

Whole-cell lysate IB offers the advantage thata greater number of immunoreactive proteinscan be detected. The humoral immune responseto Borrelia antigens was analysed by IB in thesera from patients with early and late LB. BothIgM and IgG antibodies from sera of patientswith EM and late LB most frequently demon-strated reactivity to the OspC protein. A num-ber of investigators showed that OspC is a po-tent immunogen in early Lyme disease (5, 23).Wilske et al. (5) concluded that in late Lymedisease detection of IgG against OspC is quiterare. However, IgM antibodies against OspCmay persist for months in some patients, evenafter successful therapy. Seinost et al. (24) re-ported that OspC, as a diagnostic antigen, gavehighly specific but not sensitive results due to itsvariability. In our study, Osp C was the most

JOVICIC et al.

important protein in early and late disease andwe postulate that these results are the outcomeof performance of IB based on the local B.afzelii strain abundantly expressing OspC.

The reactivity to p39 was high in IgM andIgG IB for all stages of the disease. Borreliamembrane protein A (BmpA) – p39 has beendescribed as an important immunogenic proteinfor the serodiagnosis of LB. However, sensitivityof p39 in IgM IB varies depending on the assaysystem used and may also be influenced by themanifestation of the disease. In contrast to ourresults, Roessler et al. (12) reported that p39was not a useful antigen for IgM antibody de-tection. Our investigation showed that p39 is asensitive and specific marker for serodiagnosisof LB, as detected by IB based on local B.afzelii. The anti-p41 IgM response was lessprominent than the anti-p39 IgM response inpatients with EM, and that phenomenon is sig-nificant for serological diagnosis in our geo-graphical region due to cross-reactivity with p41(only strong bands were considered positive).

In our study, IgG antibodies against Osp17were detected in a high percentage of sera frompatients with EM and late manifestations of dis-ease. Ulbrandt et al. (25) recently concludedthat Osp17 and decorin-binding protein A(DbpA) from B. afzelii strain PKo are the sameprotein. The Osp17 antigen from this particularborrelial strain has been proposed as a compo-nent in a new recombinant IgG IB (13), withseropositivity of 24% to 36% in patients withearly disseminated borreliosis and up to 85% inpatients with late disease (13, 26). Investigatorsfrom Finland (27) very recently demonstratedthat DbpA seems to be a sensitive and specificantigen for the serodiagnosis of LA or NB, butnot EM, provided that variants from all threepathogenic borrelial species are included in thecombined set of antigens. Our IB analyses sug-gest that an Osp17 protein is an immunodomin-ant protein for LB, predominantly recognizedby IgG antibodies in early and late disease.

The other significant protein in the early stageof the disease detected by IgM and IgG IB wasOspB. The exlusion of OspB from IB criteriaresulted in a negative serodiagnosis in 2.5% ofour patients with early Lyme disease. Hilton etal. (28) reported that the exclusion of OspA andOspB from the diagnostic criteria resulted in theunderdiagnosis of LB in 8% of patients.

1058

In our study the IgG antibodies from sera ofLyme disease patients also reacted againstOspD and OspE. Further investigation willshow the role of outer surface proteins in theserodiagnosis of LB.

In conclusion, the choice of a local B. afzeliistrain as antigen in IB affects the level of reac-tivity detected in the sera from patients with LB,especially in the early course of the disease. Theresults suggest that the screening serodiagnostictest should be ELISA instead of IFA, and thesecond confirmatory test should be IB. Thisstudy is the first in this geographical region andwill improve the local protocol for serodiagnosisof Lyme disease.

We are grateful to Nina Bulajic and Aleksandra Jaks-ic for valuable suggestions and review of the manu-script . We thank Sanja Jokic for help with the manu-script.

REFERENCES

1. Burgdorfer W, Barbour AG, Hayes SF, BenachJL, Grunwaldt E, Davis JP. Lyme disease – a tickborne spirochetosis? Science 1982;216:1317–9.

2. Steere AC. Lyme disease. N Engl J Med 2001;12:115–25.

3. Stiernstedt GT, Granström M, Hederstedt B,Sköldenberg B. Diagnosis of spirochetal menin-gitis by enzyme-linked immunosorbent assay andindirect immunofluorescence assay in serum andcerebrospinal fluid. J Clin Microbiol 1985;21:819–25.

4. Mitchell PD, Reed KD, Aspeslet TL, Vander-mause MF, Melski JW. Comparison of four im-munoserologic assays for detection of antibodiesto Borrelia burgdorferi in patients with culture-positive erythema migrans. J Clin Microbiol1994;32:1958–62.

5. Wilske B, Zöler L, Brade V, Eiffert H, Göbel UB,Stanek G in cooperation with Pfister HV. MiQ12 Lyme borreliosis. In: Mauch H, Lütticken R,Gatermann S, editors. Quality Standards for theMicrobiological Diagnosis of Infectious Disease.Urban & Fischer Verlag, München Jena 2000.

6. Centers for Disease Control and Prevention. Rec-ommendations for test performance and inter-pretation from the Second National Conferenceon Serologic Diagnosis of Lyme Disease. Mor-bidity Mortality Weekly Report 1995;44:590–1.

7. Robertson J, Guy E, Andrews N, Wilske B, AndaP, Granström M, et al. A European multicenterstudy of immunoblotting in serodiagnosis ofLyme borreliosis. J Clin Microbiol 2000;38:2097–102.

IMPROVED SERODIAGNOSIS OF EARLY LYME BORRELIOSIS

8. Hansen K, Hindersson P, Pedersen NS. Measure-ment of antibodies to the Borrelia burgdorferiflagellum improves serodiagnosis in Lyme dis-ease. J Clin Microbiol 1988;26:338–46.

9. Oksi J, Uksila J, Marjamäki M, Nikoskelainen J,Viljanen MK. Antibodies against whole soni-cated Borrelia burgdorferi spirochetes, 41-kilod-alton flagellin, and p39 protein in patients withPCR or culture-proven late Lyme borreliosis. JClin Microbiol 1995;33:2260–4.

10. Nilsson I, Andersson von Rosen I. Serum anti-bodies against Borrelia afzelii, Borrelia burgdor-feri sensu stricto and 41-kiloDalton flagellin inpatients from a Lyme borreliosis endemic area:analysis by EIA and immunoblot. APMIS 1996;104:907–14.

11. Gomes-Solecki MJ, Dunn JJ, Luft BJ, CastilloJ, Dykhuizen DE, Yang X, et al. Recombinantchimeric Borrelia proteins for diagnosis of Lymedisease. J Clin Microbiol 2000;38:2530–5.

12. Roessler D, Hauser U, Wilske B. Heterogeneity ofBmpA (P39) among European isolates of Borreliaburgdorferi sensu lato and influence of interspeciesvariability on serodiagnosis. J Clin Microbiol1977;35:2752–8.

13. Wilske B, Habermann C, Fingerle V, HillenbrandB, Jauris-Heipke S, Lehnert G, et al. An im-proved recombinant IgG immunoblot for serodi-agnosis of Lyme borreliosis. Med Microbiol Im-munol 1999;188;139–44.

14. Barbour A G. Isolation and cultivation of Lymedisease spirochetes. Yale Biol Med 1984;57:521–5.

15. Grego E, Colic M, Jovicic V, Lako B. Productionand characterization of monoclonal antibodiesto antigens of Borrelia burgdorferi strain Ko-sutnjak – K1. In: Lukic ML, Colic M, Mostar-ica-Stojkovic M, Cuperlovic K, editors. Immuno-regulation in Health and Disease. AcademicPress. San Diego, London 1997:397–404.

16. Craft JE, Grodzicki RL, Steere AC. Antibody re-sponse in Lyme disease: evaluation of diagnostictests. J Infect Dis 1984;149:789–95.

17. Lowry OH, Rosebrough NJ, Farr AL, RandallRJ. Protein measurement with the Folin phenolreagent. J Biol Chem 1951;193:265–75.

18. Laemli UK. Cleavage of structural proteins dur-ing assembly of the head of bacteriophage T4.Nature (London) 1970;227:680–5.

19. Hauser U, Lehnert G, Lobentanzer R, Wilske B.

1059

Interpretation criteria for standardized Westernblots for three European species of B. burgdorferisensu lato. J Clin Microbiol 1997;35:1433–44.

20. Stanek G, O’Connell S, Cimmino M, Abere E,Kristoferitsch W, Granström M. EuropeanUnion concerted action on risk assessment inLyme borreliosis: clinical case definitions forLyme borreliosis. Wien Klin Wochenschr 1996;108:741–7.

21. Engstrom SM, Shoop E, Johnson RC. Immuno-blot interpretation criteria forserodiagnosis of early Lyme disease. J ClinMicrobiol 1995;33:419–27.

22. Hauser U, Lehnert G, Wilske B. Validity of inter-pretation criteria for standardized Western blots(immunoblots) for serodiagnosis of Lyme borre-liosis based on sera collected throughout Europe.J Clin Microbiol 1999;37:2241–7.

23. Vaz A, Glickstein L, Field JA, McHugh G, Si-kand VK, Damle N, et al. Cellular and humoralimmune responses to Borrelia burgdorferi anti-gens in patients with culture positive early Lymedisease. Infect Immun 2001;69:7437–44.

24. Seinost G, Dykhuizen DE, Dattwyler RJ, GoldeWT, Dunn JJ, Wang IN, et al. Four clones ofBorrelia burgdorferi sensu stricto cause invasiveinfection in humans. Infect Immun 1999;67:3518–24.

25. Ulbrandt ND, Cassatt DR, Patel NK, RobertsWC, Bachy CM, Fazenbaker CA, et al. Confor-mational nature of the Borrelia burgdorferi deco-rin binding protein A epitopes that elicit protec-tive antibodies. Infect Immun 2001;69:4799–807.

26. Pohl-Koppe A, Kaunicnik A, Wilske B. Char-acterization of the cellular and humoral immuneresponse to outer surface protein C and outersurface protein 17 in children with early dissemi-nated Lyme borreliosis. Med Microbiol Immunol2001;189:193–200.

27. Heikkilä T, Seppälä I, Saxen H, Panelius J,Yrjänäinen H, Lahdenne P. Species-specific sero-diagnosis of Lyme arthritis and neuroborreliosisdue to Borrelia burgdorferi sensu stricto, B.afzelii, and B. garinii by using decorin bindingprotein A. J Clin Microbiol 2002;40:453–60.

28. Hilton E, Devoti J, Sood S. Recommendation toinclude OspA and OspB in the new immuno-blotting criteria for serodiagnosis of Lyme dis-ease. J Clin Microbiol 1996;34:1353–4.