Diagnostic and Biological Significance of Anti-p41 IgM Antibodies against Borrelia burgdorferi

Transcription

1 Scand. J. Immunol. 53, 416±421, 2001 Diagnostic and Biological Significance of Anti-p41 IgM Antibodies against Borrelia burgdorferi E. ULVESTAD,* A. KANESTRéM,* L. J. SéNSTEBY,* R. JUREEN,* T. OMLAND,* B. EDVARDSEN,* J. LUNDERVIK,* E. KRISTOFFERSEN* & A. P. VAN DAM² *Department of Microbiology and Immunology, The Gade Institute, Haukeland University Hospital, 5021 Bergen, Norway, and ²Department of Medical Microbiology, Academic Medical Centre, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands (Received 15 November 2000; Accepted in revised form 29 December 2000) Ulvestad E, Kanestrùm A, Sùnsteby LJ, Jureen R, Omland T, Edvardsen B, Lundervik J, Kristoffersen E, van Dam AP. Diagnostic and Biological Significance of Anti-p41 IgM Antibodies against Borrelia burgdorferi. Scand J Immunol 2001;53:416±421 We performed a prospective study to investigate the biological significance and diagnostic specificity of anti-p41 immunoglobulin (Ig)M antibodies against Borrelia burgdorferi. During a 1-year interval 2403 patients were referred to our department for B. burgdorferi serology. Sixty-three patients had repetitive positive tests for IgM anti-p41 antibodies and negative tests for anti-p41 IgG antibodies. Ten of the 63 patients recently had symptoms of erythema migrans. A confirmatory IgM Western blot gave a positive reaction in 5 patients out of 53 patients with little or no clinical evidence of B. burgdorferi infection. The remaining 48 patients were negative in this test and were considered as false-positives. Two whole cell enzyme-linked immunosorbent assay (ELISAs), two immunofluorescence assays and Western blotting were not useful as confirmatory tests. Sera from 330 blood donors and 72 cord sera were also screened for anti-p41 IgM. Five blood donor sera and five cord sera showed an IgM reactivity against p41. Based on our data we hypothesize that up to 1.5% of the population may have natural IgM antibodies against p41 in their sera. We observed that six out of nine sera with such antibodies could immobilize a B. afzelii reference strain in vitro. Whether anti-p41 IgM antibodies are capable of inactivating infective spirochetes and thereby prevent infection in vivo is, however, not yet clarified. The paradoxical conclusion that anti-p41 IgM antibodies may be a sign of resistance to infection rather than a sign of infection should be given consideration. Dr E. Ulvestad, Department of Microbiology and Immunology, The Gade Institute, Haukeland Hospital, Armauer Hansen Building, N-5021 Bergen, Norway. elling.ulvestad@haukeland.no INTRODUCTION Lyme disease is a multisystem infectious disease caused by at least three spirochetes in the Borrelia burgdorferi genospecies complex: B. burgdorferi sensu stricto, B. garinii, and B. afzelii. Following transmission through the bite of B. burgdorferiinfected Ixodes ticks, most patients initially develop a distinctive rash, erythema migrans, accompanied by fatigue, headache, ache in joints and muscles, and fever. Symptoms and signs of the disseminated disease, especially the neurological, cardiac, and articular disease, may develop weeks to months after the initial exposure. Whereas erythema migrans is pathognomonic for the Lyme disease [1], the confirmation of a late disseminated disease requires laboratory support in addition to clinical evidence of an infection [2]. Because conventional microbiological methods such as cultivation and polymerase chain reaction (PCR) have low diagnostic sensitivities in the later stages of the disease, the disseminated Lyme disease is mostly diagnosed using serology. Patients with Lyme disease usually develop IgM antibodies against the 41 kda flagellar antigen (p41) of the spirochete during the early infection, followed by the development of IgG antibodies against p41 and various outer surface proteins [3]. Unconditional acceptance of the anti-p41 IgM antibodies as an early immunological manifestation of infection is, however, problematic. The diagnostic specificity of anti-p41 IgM antibodies may be confounded by several factors, some of which should be considered prior to the diagnostic or therapeutic decision-making. Firstly, anti-p41 IgM antibodies may persist for up to 5 years following treatment [4]. Secondly, cross-reactive q 2001 Blackwell Science Ltd

2 Borrelia burgdorferi Anti-p41 IgM Antibodies 417 anti-p41 IgM antibodies may be elicited by other spirochetes owing to the presence of conserved epitopes on their flagellar proteins [5]. Finally, anti-p41 IgM antibodies have also been detected in patients without the evidence of past or current Lyme disease [6] and in human cord blood [7]. The diagnostic uncertainties associated with the early B. burgdorferi infection have led clinicians to interpret positive results for anti-p41 IgM antibodies as evidential support for the disease. Accordingly, they have treated their patients with antibiotics. However, such reasoning commits the fallacy of equating the statement `anti-p41 IgM antibodies occur early during Lyme disease' with the statement `anti-p41 IgM antibodies are diagnostic for early Lyme disease'. The second statement, which can neither be logically nor biologically deduced from the first statement, has been firmly falsified [4±7]. In order to gain information on the diagnostic utility of antip41 IgM antibodies we studied the clinical diagnosis of a large number of patients with anti-p41 IgM but no IgG antibodies in their sera. In addition, we undertook a study of the immunological reactivity of these sera and further investigated the biological function of anti-p41 IgM antibodies in patients with and without borreliosis in an immobilization assay. MATERIALS AND METHODS Patients and controls. All sera submitted for B. burgdorferi serology during a 1-year interval were routinely tested with the IDEIA IgM and IgG test (DAKO Diagnostics Ltd, Cambridge, UK). New sera from all patients positive for IgM anti-p41 antibodies were obtained at a minimum of 2 weeks following the first blood sample. Patients with a sustained IgM anti-p41 response that did not develop anti-b. burgdorferi IgG antibodies were included for further study. Additional information on clinical symptoms and signs relevant to B. burgdorferi infection, especially manifestations involving the skin, the joints and the nervous system, was obtained from the patient's physician. Tick bites and the subsequent development of erythema migrans were noted, as was evidence of other infections or autoimmunity. Antibiotic therapy was also noted. Sera from 127 blood donors were collected at the Bloodbank, Haukeland University Hospital, Bergen, Norway, located in an area where borrelia-infected Ixodes ricinus ticks are prevalent. Sera from 203 blood donors were also collected from the Bloodbank, Tromsù University Hosptial, Tromsù, Norway, located in an area devoid of the Ixodes ricinus tick. Seventy-two cord sera were obtained from the Department of Obstetrics and Gynaecology at the Haukeland University Hospital following normal births. Serological analysis. All serology tests were performed strictly as recommended by the manufacturers. No attempts were made to redefine the given cut-off values. We utilized the IDEIA ELISA test which separately detects IgM and IgG antibodies to B. Burgdorferi 41 kda flagellar antigen as the primary test. The test for IgM anti-p41 utilizes the m-capture principle in which patient IgM antibodies are selectively captured by anti-igm antibodies coated to microwells. Thereafter the anti-p41 IgM antibodies are selectively identified by their reactivity with peroxidase-conjugated flagellum. The IgG test is an indirect ELISA in which microwells are coated with purified native borrelial flagellum. A selection of patient and blood donor sera were tested for reactivity against an antigenic extract of the PKo B. afzelii strain using the Enzygnost Borreliosis IgM and IgG indirect ELISA test (Behring Diagnostics, Frankfurt, Germany). Patient sera were also tested for reactivity against a mixture of antigens from B. burgdorferi sensu stricto (IRS) and B. afzelii (VS 461) using the Freka Borrelia IgG and IgM indirect ELISA (Fresenius, Bad Homburg v.d. HoÈhe, Germany). Before application of these two tests, sera were adsorbed with Treponema phagedenis to reduce possible cross-reacting antibodies. IgM antibodies against intact B. burgdorferi sensu stricto (IRS) spirochetes were also analyzed using indirect immunofluorescence (Freka Borrelia IFA, Fresenius). The VIDAS Lyme Screen II test (biomeárieux, Marcy-l'Etoile, France), an enzyme-linked fluorescent immunoassay (ELFA), was used to investigate the combined IgM and IgG reactivity of patient and blood donor sera against a purified extract of B. burgdorferi sensu stricto strain B31. In order to investigate whether sera contained IgM and IgG antibodies against other than the p41 flagellar antigen, all patient sera were subjected to Western blotting (Lyme Blot IgM and IgG, DAKO). Antigens of B. garinii separated by polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes were incubated with patient sera. IgM and IgG antibodies against the various antigens were thereafter investigated using isotype specific secondary antibodies. For determination of IgM reactivity, sera were scored as positive if band 23 (OspC) alone, both band 39 (p39) and 41 (p41), or alternatively if all three bands were positive. Otherwise, sera were scored as negative. The IgG Western blot was scored as positive if reactivity to four or more of the bands 21, 23, 37, 39, 41, 45 and 93 kda was observed. We scored reactivity with the p41 band separately. To investigate for serological cross reactivity with Treponema Pallidum, sera were analyzed using a passive particle agglutination assay where gelatin particles are sensitized with purified Treponema Pallidum (Serodia-TP-PA, Fujirebio Inc., Tokyo, Japan). B. burgdorferi immobilization test. Nine sera with high optical density values (OD) in the IDEIA IgM test were tested in a B. burgdorferi immobilization test [8]. The immobilization study was performed with strain B31 (B. burgdorferi sensu stricto) and with strain PKo (B. afzelii). Fifty ml B. burgdorferi spirochetes in liquid culture medium were mixed with 25 ml normal human serum and 25 ml heat inactivated test serum. After 5 and 24 h incubation the percentage of nonmotile spirochetes was determined by dark-field microscopy. All experiments were performed in duplicate. More than 50 spirochetes were counted in a minimum of five fields. Earlier analysis of the assay showed that negative control sera immobilize less than 10% of the spirochetes after 5 h and less than 25% after 24 h (data not shown). Immunochemical analysis.. Concentrations of total serum IgG and IgM were quantified in a Behring nephelometer using monospecific antisera (Behringwerke AG, Frankfurt, Germany). RESULTS Patient characteristics Enduring anti-p41 IgM antibodies were identified in 63 patients, 30 males and 33 females, from a total of 2403 sera tested (2.6%). The median interval between the first sample and the follow-up sample was 32 days (range 13±156 days). The median age of the patients was 32 years (range 8±80 years). The patients were initially classified into three groups based on the clinical information obtained. Patients with erythema migrans (n ˆ 10) were classified as `borrelia infection`. For

3 418 E. Ulvestad et al. Table 1 Percentage IgM positive sera against various species and strains of B. burgdorferii within the three clinical categories Clinical category (n ˆ patients) Borrelia Burgdorferii PKo Strains IRS and VS461 Strain B31 Strain IRS B. garinii Borrelia infection (n ˆ 10) Borrelia infection unlikely (n ˆ 24) No borrelia infection (n ˆ 29) Total (n ˆ 63) Reactivity against B. afzelii Pko strain was measured using Enzygnost Borreliosis ELISA, against the mixture of B. burgdorferi sensu stricto IRS and B. afzelii VS461 strains by Freka Borrelia ELISA, against B. burgdorferi sensu stricto strain B31 by VIDAS enzyme-linked fluorescent assay, against the B. burgdorferi sensu stricto strain IRS by Freka Borrelia indirect immunofluorescence, and against B. garinii by western blot (WB). For WB reactivity to the p41 band is scored. Borderline results are not included. all but two of these patients a physician had evaluated the skin lesion. All patients in this group had received treatment with penicillin. The category `borrelia infection unlikely' (n ˆ 24) included 12 patients with aches in muscles and joints, seven patients with cutaneous reactions other than erythema migrans, three patients with cardiac symptoms, one patient with palsy of the facial nerve, and one patient with optic neuritis. A history of tick bite was known for only four patients in this category. The last category, `no borrelia infection' (n ˆ 29) consisted of patients with a probable viral disease and other patients without clinical signs or symptoms suggestive of borreliosis. All patients had a normal functioning humoral immune system, as indicated by measurements of Ig concentrations. The median IgM value was 1.3 g/l (range 0.3±5.1 g/l), and the median IgG value was 10.6 g/l (range 5.7±20.5 g/l). Out of 63 patients with a positive test in the IDEIA IgM test, 36 patients had positive IgM reactivity against the PKo strain (Enzygnost Borreliosis ELISA). Three patients had a borderline IgM reactivity and 23 patients were evaluated as IgM negative. Three patients had borderline IgG reactivity against the PKo strain; all other patients were found negative for IgG. When sera were investigated for IgM reactivity against a mixture of antigens from B. burgdorferi sensu stricto IRS strain and B. afzelii strain VS461 (Freka Borrelia ELISA), 15 patients were scored as positive, one patient showed an intermediate result, and 47 patients were negative. All patients were negative for IgG antibodies against the two strains. Sixteen sera gave a positive IgM reaction in indirect immunofluorescence against the B. burgdorferi sensu stricto IRS strain (Freka Borrelia IFA) (Fig. 1). Twelve sera were scored as weakly positive, and 33 sera were scored as negative. Serological reactivity All anti-p41 IgM positive sera were found negative when analyzed for reactivity against Treponema Pallidum. All patients with `borrelia infection' gave negative results in the IgM and IgG Western blots. Sera from four patients with `borrelia infection unlikely' and one serum from a patient with `no borrelia infection' reacted with p41 and either p39 or OspC, and were scored as positive in the IgM blot. These five sera had negative IgG blots. Clinical evaluation of the five patients gave no conclusive evidence for a B. burgdorferi infection. One patient reported a tick bite with erythema. However, the erythema was judged not to be erythema migrans and antibiotics were not given. The other four patients gave no conclusive history of tick bites. One patient had arthritis but no other signs or symptoms of borrelia infection. None of the patients had neurological symptoms. Percentage IgM positive sera against various species and strains of B. burgdorferii within the three clinical categories are given in Table 1. Fig. 1. Indirect immunofluorescence microphotograph showing an immunoglobulin (Ig)M reactivity against immobilized Borrelia burgdorferi spirochetes. Slides were examined in a BioRad MRC 1000 confocal laser scanning microscope using the 100 objective lens.

4 Borrelia burgdorferi Anti-p41 IgM Antibodies 419 Out of the 330 blood donor sera investigated in the IDEIA IgM test, five sera came out as IgM positive and IgG negative, giving a prevalence of anti-p41 IgM antibodies of 1,5%. Three of the blood donors belonged to the Haukeland area; the other two came from the Tromsù area. One of the sera also reacted with the B. burgdorferi sensu stricto strain B31. The sera were negative against all other species. One serum reacted with p41 in Western blot (Fig. 2). Out of 72 cord sera analyzed, five sera reacted serologically with the p41 antigen, with one serum showing IgM reactivity against the p41 band in Western blot (Fig. 2). The cord sera were all negative for IgG antip41 antibodies. B. burgdorferi immobilization Nine sera obtained from all three clinical categories were investigated for immobilizing activity against the two strains B31 and PKo. All sera showed IgM reactivity against the PKo strain, four sera also showed anti-p41 IgM reactivity in Western blot. None of the sera showed other bands in the IgM blot. Three sera showed reactivity against strain B31, two of these were positive for IgM anti-p41 in Western blot. The sera were negative for IgG in all diagnostic assays. After 5 h incubation of serum and spirochetes, an immobilizing activity against the PKo strain was repetitively observed in six sera (Table 2). After 24 h incubation, a serum derived from a patient with borrelia infection, immobilized 100% of the spirochetes. Interestingly, this serum also had the highest reactivity in ELISA. The other five sera immobilized between 30% and 38% of the spirochetes. A negative control serum showed 3% immobilization of spirochetes after 5 h incubation and 13% immobilization after 24 h. Sera from all three clinical categories were capable of Fig. 2. Western blot of sera with IgM reactivity against the anti-p41 band. Lane A shows a positive control with IgM reactivity against the 41, 39, 32 and 23 kda bands. Sera were obtained from a patient with infectious mononucleosis (B), from a blood donor (C), and from cord blood (D). Seven patients were positive for IgM and/or IgG antibodies against the B31 strain (Vidas ELFA). Eight patients showed borderline reactivity and 43 patients showed no reactivity against the B31 strain. Flagellin-reactivity was also separately analyzed in Western blot. Twenty-two patient sera belonging to all categories showed IgM reactivity to the p41 band only (Fig. 2). Two sera were Western blot positive for IgG anti-p41, one from a patient with `borrelia infection unlikely' and one from a patient with `no borrelia infection'. Table 2 Percentage immobilized spirochetes of the B31 and pko strains following incubation of spirochetes and sera containing anti-p41 IgM antibodies Infection OD values IDEIA IgM Immobilization 5 h incubation (%) B31 pko pko Yes Yes Yes nt* Unlikely nt Unlikely Unlikely Unlikely No No nt Control *Not tested. Immobilization 24 h incubation (%)

5 420 E. Ulvestad et al. immobilizing spirochetes of the PKo strain. In contrast, none of the sera were capable of immobilizing strain B31. DISCUSSION Our results show that 2.6% of the sera submitted for B. burgdorferi serology expressed anti-p41 IgM but not IgG antibodies. Because 84% of these anti-p41 IgM positive sera came from patients in which a B. burgdorferi infection was unlikely, most of the results should be classified as diagnostically false positive for these patients. Only 16% of the sera with anti-p41 IgM antibodies came from patients with likely B. burgdorferi infection. Notably, none of the sera from the patients with recent erythema migrans were positive in the IgM immunoblot. It is not clear why the patients with erythema migrans did not develop IgG anti-p41 antibodies or antibodies against other borrelia antigens on immunoblots. However, all these patients had been treated with penicillin early during infection. This may have reduced the antigenic load and thereby prevented the natural development of the immune reaction [9]. Penicillin does not exert a negative effect on antibody production when used in therapeutic doses and it is therefore not likely that such a treatment prevented the class switch [10]. The anti-p41 IgM antibodies in patients without a B. burgdorferi infection were probably not produced following an immune response to flagellins of cross-reactive microbial agents because IgG anti-p41 antibodies were not generated [11]. Furthermore, the combined observations that human cord blood contains IgM antibodies to the 41kd flagellar antigen of B. burgdorferi and that such antibacterial antibodies remain in the antibody repertoire into adulthood [12] support the hypothesis that anti-p41 IgM antibodies can occur as natural antibodies. Although the functional significance of natural antibodies has not been ascertained, recent data indicate that natural IgM antibodies have an important role in the immediate defence against severe bacterial infection [13]. The most efficient antibodies for microbial defence react with surface exposed epitopes on the invading microbe. Although located in the periplasmic space, flagellar antigens appear to be surface exposed [14]. This is in agreement with our results showing that the PKo strain was immobilized with anti-p41 IgM antibodies. Although our in vitro results would suggest that antibodies against flagellin could be protective, several data contrast this hypothesis. Firstly, some patients with successfully treated B. burgdorferi infection may later develop clinical reinfection [15]. Secondly, immunization of mice with flagellin has no protective effect [16], but, so far, this has only been proven for infections with B. burgdorferi sensu stricto. However, we have previously shown that there is variability in the sensitivity among spirochetes for complement-mediated killing by human serum [8]. Our present data indicate that also anti-p41 IgM antibodies have a differential effect against various strains of B. burgdorferi. The hypothesis that natural anti-p41 IgM antibodies may confer protection against infection with some B. burgdorferi species and strains should therefore be thoroughly considered when evaluating the natural defence against B. burgdorferi infection. The variability of the immune response to B. burgdorferi and the lack of sensitivity and specificity of commercial assays for the detection of antibodies to the B. burgdorferi spirochete have made the serodiagnosis of the Lyme disease an arduous task, especially during the early disease. Although about 95% of patients with a disseminated disease eventually develop antibodies to B. burgdorferi, diagnostic sensitivity of serological tests during early disease is less than 50%, whereas falsepositive results frequently occur. In our study, none of the utilized tests allowed us to discriminate between patients with and without B. burgdorferi infection. It appears that some of these diagnostic problems may be caused by characteristics inherent to the normal functioning of the immune system. ACKNOWLEDGMENTS IreÁn Folkem (DAKO Norway), Lisbeth Rohde (Behring Diagnostika Norway AS), Mette Nerli (Bergman AS) and Inger Jensen (Orion Diagnostica AS) kindly provided the diagnostic tests. Dr Bjùrn Skogen kindly provided sera from the Tromsù blood donors. REFERENCES 1 Steere AC, Malawista SE, Hardin JA, Ruddy S, Askenase PW, Andiman WA. Erythema chronicum migrans and Lyme arthritis: the enlarging clinical spectrum. Ann Intern Med 1977;86:685±98. 2 Hofstad H, Matre R, Nyland H, Ulvestad E. Bannwarth's syndrome: serum and CSF IgG antibodies against Borrelia burgdorferi examined by ELISA. Acta Neurol Scand 1987;75:37±45. 3 Ma B, Christen B, Leung D, Vigo-Pelfrey C. Serodiagnosis of Lyme borreliosis by Western immunoblot: reactivity of various significant antibodies against Borrelia burgdorferi. J Clin Microbiol 1992;30:370±6. 4 Hilton E, Tramontano A, DeVoti J, Sood SK. Temporal study of immunoglobin M seroreactivity to Borrelia burgdorferi in patients treated for Lyme borreliosis. J Clin Microbiol 1997;35:774±6. 5 Coleman JL, Benach JL. Identification and characterization of an endoflagellar antigen of Borrelia burgdorferi. J Clin Invest 1989;84:322±30. 6 Cooke WD, Bartenhagen NH. Seroreactivity to Borrelia burgdorferi antigens in the absence of Lyme disease. J Rheumatol 1994;21:126±31. 7 Cooke WD, Orr AS, Wiseman BL, Rouse SB, Murray WC, Ranck SG. Human cord blood contains an IgM antibody to the 41 kd flagellar antigen of Borrelia burgdorferi. Scand J Immunol 1993;38:407±9. 8 van Dam AP, Oei A, Jaspars R et al. Complement-mediated serum sensitivity among spirochetes that cause Lyme disease. Infect Immun 1997;65:1228±36. 9 Golightly MG. Lyme borreliosis: laboratory considerations. Semin Neurol 1997;17:11±7. 10 van Vlem B, Vanholder R, De Paepe P, Vogelaers D, Ringoir S. Immunomodulating effects of antibiotics: literature review. Infection 1996;24:275± Berland R, Fikrig E, Rahn D, Hardin J, Flavell RA. Molecular

6 Borrelia burgdorferi Anti-p41 IgM Antibodies 421 characterization of the humoral response to the 41-kilodalton flagellar antigen of Borrelia burgdorferi, the Lyme disease agent. Infect Immun 1991;59:3531±5. 12 Mouthon L, Lacroix-Desmazes S, Nobrega A, Barreau C, Coutinho A, Kazatchkine MD. The self-reactive antibody repertoire of normal human serum IgM is acquired in early childhood and remains conserved throughout life. Scand J Immunol 1996;44:243± Boes M, Prodeus AP, Schmidt T, Carroll MC, Chen J. A critical role of natural immunoglobulin M in immediate defense against systemic bacterial infection. J Exp Med 1998;188:2381±6. 14 Luft BJ, Jiang W, Munoz P, Dattwyler RJ, Gorevic PD. Biochemical and immunological characterization of the surface proteins of Borrelia burgdorferi. Infect Immun 1989;57:3637± Hammers-Berggren S, Lebech AM, Karlsson M, Svenungsson B, Hansen K, Stiernstedt G. Serological follow-up after treatment of patients with erythema migrans and neuroborreliosis. J Clin Microbiol 1994;32:1519± Fikrig E, Barthold SW, Marcantonio N, Deponte K, Kantor FS, Flavell RA. Roles of OspA, OspB, and flagellin in protective immunity to Lyme borreliosis in laboratory mice. Infect Immun 1992;60:657±61.