Residual Brain Infection in Relapsing-Fever Borreliosis

Transcription

1 MAJOR ARTICLE Residual Brain Infection in Relapsing-Fever Borreliosis Diego Cadavid, 1,2 Marie Sondey, 1,2,a Edwin Garcia, 1,2,a and Catherine L. Lawson 3 1 Department of Neurology and Neuroscience and 2 Center for the Study of Emerging Pathogens, University of Medicine and Dentistry of New Jersey New Jersey Medical School, Newark, and 3 Department of Chemistry and Chemical Biology, Rutgers University, Piscataway, New Jersey Background. Neurological involvement is common in the spirochetal infection relapsing fever (RF) in both humans and experimental animals. RF is best known for antigenic variation caused by the sequential expression of variable outer membrane lipoproteins of 2 sizes, variable small (Vsp) and variable large (Vlp) proteins. Less understood is the persistence of RF borreliae in the brain after they are cleared from the blood, referred to as residual brain infection (RBI). Our goal was to investigate the phenomenon of RBI in RF. Methods. We studied RBI in immunocompetent mice by culturing blood and perfused brain samples 1 month after intraperitoneal inoculation with Borrelia turicatae serotype 1 (Bt1). Mice deficient in Toll-like receptor 2 (TLR2 / ) or in B and T cells (scid) were included for comparison. Results. All scid mice had persistent infection in blood and brain. RBI was found in 3 (19%) of 16 immunocompetent and TLR2 / mice. RBI was caused by either persistence of the original serotype (Bt1) or newly emerged Vsp ( n p 1, renamed Bt3) or Vlp serotypes. The Vsp of Bt1 (Vsp1) and Bt3 (Vsp3) were 75% identical. Conclusions. RBI in RF is relatively frequent and can occur by persistence of the original or newly emerged serotypes. Brain infection is a characteristic feature of many pathogenic spirochetes, notably Treponema palllidum and the several Borrelia species that cause Lyme disease (LD) and relapsing fever (RF). The neurological complications of Borrelia infections are collectively referred to as neuroborreliosis [1, 2]. The most common manifestations of neuroborreliosis in RF are meningitis, facial nerve palsy, radiculitis, and encephalopathy (reviewed in [3]). In outbreaks of tickborne RF, neuroborreliosis can be the dominant clinical feature [4, 5]. Our laboratory has been studying the pathogenesis of neuroborreliosis in animal models using nonhuman primates infected with LD borreliae and mice infected with RF borreliae [6 8]. RF borreliae are best known for the phenomenon of anti- Received 24 August 2005; accepted 9 December 2005; electronically published 4 April Potential conflicts of interest: none reported. Financial support: Foundation of the University of Medicine and Dentistry of New Jersey (grant to D.C.); Heritage Affiliate of the American Heart Association (Scientist Development Grant T to D.C.). a Present affiliations: Department of Biochemistry, University of Medicine and Dentistry of New Jersey New Jersey Medical School, Newark (M.S.); New York College of Osteopathic Medicine, Old Westbury (E.G.) Reprints or correspondence: Dr. Diego Cadavid, University of Medicine and Dentistry of New Jersey New Jersey Medical School, 185 S. Orange Ave., MSB H506, Newark, NJ (cadavidi@umdnj.edu). The Journal of Infectious Diseases 2006; 193: by the Infectious Diseases Society of America. All rights reserved /2006/ $15.00 genic variation, which allows them to spontaneously change their serotype by switching the expression of variable major proteins (VMPs) [9, 10] of 2 sizes, large (Vlp) and small (Vsp). A switch in VMP allows RF borreliae to escape killing by the host s serotype-specific antibody response. Less known is the phenomenon of residual brain infection (RBI), which refers to the tendency of RF borreliae to persist in the brain after they disappear from the blood (reviewed in [1]). The goal of the present study was to investigate the phenomenon of RBI in RF. The results revealed that RBI is relatively frequent in both immunocompetent and Toll-like receptor 2 deficient (TLR2 / ) mice. RBI can be caused by the persistence of spirochetes in the brain, after their clearance from the blood, of either the original infecting serotype or new serotypes that spontaneously emerge during infection. MATERIALS AND METHODS Strains and culture conditions. B. turicatae serotypes 1 (Bt1; formerly serotype A) and 2 (Bt2; formerly serotype B) have been described elsewhere [7, 11 13]. Borreliae were cultured in BSK II medium [14] and counted in a Petroff-Hauser chamber [10]. When mouse tissue samples were cultured, 50 mg/ml rifampin and 100 mg/ ml phosphomycin were present in the medium. Plas- Residual Brain Infection in RF Borreliosis JID 2006:193 (15 May) 1451

2 Table 1. Susceptibility to neuroborreliosis in mice of different genetic background inoculated with serotype 1 of Borrelia turicatae. Mouse Blood Infected tissue Brain Serotype in brain a Vestibular dysfunction b Swiss Websters (outbred) 0/4 1/4 Novel 1/4 C3H/HeJ 0/4 1/4 Serotype 1 0/4 SWR/J 1/4 0/4 0/4 C57BL/6 0/4 1/4 Novel 0/4 C57BL/6 TLR2 / 4/16 3/16 All novel 0/16 C57BL/6 scid 4/4 4/4 Serotype 1 in all 4/4 C3H/HeJ scid 4/4 4/4 Serotype 1 in all 4/4 CB17 scid 4/4 4/4 Serotype 1 in all 4/4 NOTE. Data are no. of positive cultures/no. of tissue or fluid samples cultured. Mice were 3 inoculated intraperitoneally with 1 10 cells and infection examined by culture of blood or perfused brain 1 month later. TLR, Toll-like receptor. a By Western blot with anti variable small protein 1 polyclonal antibody. b Spinning in circles when lifted off the ground by the tail. ma samples from infected mice were used to start broth cultures, which were collected by centrifugation at 9000 g at cell densities of cells/ml. The purity of Bt1 was assessed before inoculation by Western blot with anti-vsp1 polyclonal antiserum [11, 12]. Mouse infections. Female mice, 4 6 weeks old, of different genetic backgrounds, purchased from Jackson or Charles Rivers or bred in-house (C57BL/6 TLR2 / ), were inoculated intraperitoneally with Bt1 spirochetes in 300 ml of PBS. Groups of 4 8 mice each were used for all experiments. Mice sham-inoculated with PBS were used as controls. The housing and care were in accordance with the Animal Welfare Act. Infection was confirmed by mixing tail-vein blood with an equal volume of PBS followed by dark-field microscopy. scid and C57BL/6-TLR2 / mice were maintained in a germ-free environment before and after infection. Tissue and fluid collection. Mice were anesthetized with isoflurane for euthanasia and necropsied as described elsewhere [6]. Total body perfusion with 30 ml of PBS was performed, to minimize blood contamination of the brain. Whole brains were homogenized in BSK II medium and centrifuged for 3 s at 7000 g; then, the supernatant inoculated into 6-mL BSK II culture tubes. Any brain homogenates in which blood contamination was detectable were discarded. All cultures from brain and blood were examined for spirochetes by phase-contrast microscopy for a 2-week period. Histopathological studies. Brains were removed at necropsy, fixed in 4% paraformaldehyde for 48 h at 4 C, and embedded in paraffin. Hematoxylin-eosin (HE) staining was prepared using standard techniques. All tissue sections were examined with standard light microscopy by a trained neuropathologist (D.C.) blinded to the infection status. Inflammation was defined in accordance with characteristic morphological changes after HE staining. A 3-step streptavidin-peroxidase technique was used for immunohistochemical analysis, as described elsewhere [15, 16]. Paraformaldehyde-fixed sections were treated with 0.5 mg/ml protease type VIII (Sigma P-5380) for 10 min for antigen retrieval. Commercially available rat monoclonal antibody anti mouse macrophage F4/80 at a 1:1000 dilution (Serotec) was used for the identification of activated macrophages/microglia. The chromogen was 3-3 -diaminobenzidine tetrahydrochloride. Tissue sections from uninfected mice were used as negative controls. Purified rat IgG of the same species (Sigma), matched for concentration and isotype, was used as a negative control. Image analysis was done by scoring microscopic fields per sagittal brain section as follows: 0, no F4/80-positive cells; 1, 1 2 cells/field; 2, 3 10 cells/field; 3, 110 cells/field; and 4, diffusely positive. The microscopic fields were from leptomeninges or brain parenchyma, and the examiner was blinded to infection status and genotype. Protein analysis. Whole cells from harvested cultures were subjected to SDS-PAGE analysis with 12.5% acrylamide [17]. For Western-blot analysis, proteins were transferred to nitrocellulose membranes (Millipore) that were then blocked with 3% (wt/vol) dried nonfat milk in 10 mmol/l Tris (ph 7.4) and 150 mmol/l NaCl (milk/ts) for 4 h [18]. After the membranes were washed 3 times with 0.3% milk/ts, they were incubated with anti-vsp1 polyclonal antiserum diluted 1:50,000 in 0.3% milk/ts [15]. Alkaline phosphatase conjugated anti rabbit antibody (Sigma) served as the second ligand. The blots were developed with nitroblue tetrazolium chloride/5-bromo 4-chloro 3-indolylphosphatase p-toluidine salt as the substrate (Pierce). DNA methods. Plasmid-rich Borrelia DNA was prepared using the diethylpyrocarbonate method [6]. All expressed vsp 1452 JID 2006:193 (15 May) Cadavid et al.

3 Table 2. Microglial/macrophage activation in brains of C57BL/6 wild-type or Toll-like receptor (TLR) 2 deficient (TLR2 / ) mice 1 month after inoculation of serotype 1 of Borrelia turicatae (Bt1) or PBS (as a control). Mouse strain, inoculation Brain parenchyma Leptomeninges C57BL/6 Bt (0 2.2) 7.5 ( ) PBS (0 1.7) C57BL/6 TLR2 / Bt ( ) PBS 0 0 NOTE. Macrophages and microglia were detected by immunostaining with rat anti-mouse F4/80 monoclonal antibody. Results are given as mean (95% confidence interval) sum score for microscopic fields. Scoring was done by an examiner who was blinded to infection status. The scoring system was as follows: 0, no F4/80-positive cells; 1, 1 2 cells/field; 2, 3 10 cells/field; 3, 110 cells/field; and 4, diffusely positive. genes were amplified and sequenced using a forward primer for the B. turicatae vsp promoter (5 -TTGAAAAATCGAAAT- TTTGAAATA-3, nt 9 32 on GenBank U85413) and a reversedegenerated primer for a conserved region on the 3 end of all known B. hermsii and B. turicatae vsp genes (5 -TGWAAGCK- SWKYDATTRCAG-3, nt on GenBank U854130) [11, 12]. When it is annealed at 42 C, this primer pair specifically amplifies expressed vsp and vlp genes from B. hermsii (Bh7 and Bh21) and B. turicatae (Bt1 and Bt2). Because only genes placed in front of the promoter are expressed, using a promoter specific primer allows the study of gene expression by polymerase chain reaction (PCR) instead of reverse-transcription PCR. DNA amplification was performed in a thermal cycler using 50-mL reaction volumes that contained 1.25 U of Thermus aquaticus DNA polymerase (Roche), 25 pmol of each primer, 200 mmol/ L each dntp, 10 mmol/l Tris (ph 8.3), 50 mmol/l KCl, and 4 mmol/l MgCl 2. The program was 40 cycles for 1 min at 94 C, for 1 min at 42 C, and for 2 min at 72 C, followed by a final extension cycle for 7 min at 72 C. Tubes with buffers and reagents but without DNA were included as negative controls for crosscontamination. DNA sequencing was performed 3 times on each strand using the cycle-sequencing method through commercial vendors (Retrogen). vsp3 was compared with previously characterized genes using WU-Blast2 software (version 2.0; Washington University). Vsp3 was compared to Vsp1 and Vsp2 using CLUSTAL W (version 1.82) multiple sequence alignment (available at: Statistical analysis. The 2-tailed Fisher s exact test was used to compare differences in percentages, and nonparametric tests were used to compare differences between means. P!.05 was considered to be significant. RESULTS RBI in immunocompetent mice. To study the susceptibility to RBI of immunocompetent mice, we examined outbred and inbred mice of different genetic backgrounds. All mice were inoculated intraperitoneally with Bt1 spirochetes, and the infection of blood and perfused brain was examined 1 month later by culture followed by SDS-PAGE and Westernblot analyses. The inbred strains were chosen arbitrarily on the basis of their known susceptibility to Lyme arthritis, another spirochetal disease. The strains chosen were C57BL/6 (H2b haplotype), which is moderately susceptible to Lyme arthritis [19], and C3H/HeJ (H2k) and SWR/J (H2q), which are highly susceptible [20]. Swiss Webster outbred mice were also included for the examination of RBI in a genetically heterogeneous population. RBI was defined as a positive brain culture in a mouse whose necropsy blood sample was culture negative. The results (table 1) showed that RBI was found in 3 (19%) of 16 immunocompetent mice. Although none of the inbred mice developed any signs of neurological dysfunction, 1 of 4 outbred mice developed severe head tilt, walking in circles, and spinning in the air when lifted off the ground by the tail. These manifestations of vestibular dysfunction were indistinguishable in time of onset and severity from those observed in scid mice (see below). Only 1 (6%) of 16 immunocompetent mice had positive blood cultures 1 month after inoculation (table 1). These results showed that RBI is relatively common in immunocompetent mice across different genetic backgrounds. Furthermore, they showed that, occasionally, immunocompetent mice develop vestibular dysfunction as a complication of RF borreliosis. Figure 1. Assay in 12.5% SDS-PAGE of brain (br) and blood (bl) culture lysates from 4 (1 4) C57BL/6-scid mice inoculated with serotype 1 of Borrelia turicatae (Bt1) and necropsied 1 month later. Molecular-weight markers are shown in kilodaltons on the left, and the whole serotype 1 spirochete lysate is shown on the right (Bt1), for comparison. The arrow points to 23-kDa variable small protein 1, the serotype-defining major outer membrane lipoprotein of Bt1. Residual Brain Infection in RF Borreliosis JID 2006:193 (15 May) 1453

4 RBI in TLR25/5 mice. The main virulence determinants of RF [21, 22] and LD [23] borreliae are lipoproteins. Because TLR2 is an important signal-transducing receptor for bacterial lipoproteins [24, 25], it is possible that TLR2 deficiency, by impairing recognition of the infection, might increase susceptibility to RBI in RF. To investigate this possibility, we cultured blood and perfused brain samples from TLR2 / mice 1 month after inoculation with Bt1. The results revealed RBI in 3 (19%) of 16 TLR2 / mice examined, similar to the frequency of persistent infection in blood samples (4/16 [25%]; table 1). Although the differences in persistent blood infection between immunocompetent (1/16 [6%]) and TLR2 / (4/16 [25%]) mice did not reach statistical significance, we cannot rule out a type II error, because the number of mice used was not large. It is also possible that TLR2 deficiency may influence the spirochetal load in other tissues that we did not study. None of the TLR2 / mice developed any signs of vestibular dysfunction. Because TLR2 plays an important role in the inflammatory response to LD borreliosis [25], we studied whether TLR2 deficiency had any effect in the inflammatory response to the infection in the brain of mice with RF borreliosis. For this, the brain from C57BL/6 wild-type and TLR2 / mice necropsied 1 month after intraperitoneal inoculation with Bt1 were examined microscopically. The examination of HE-stained slides showed minimal to mild meningitis in both groups (not shown). Immunohistochemical staining with an antibody to the glycoprotein F4/80, a marker of mouse microglia/macrophage activation [26], revealed increased signals, mainly in the leptomeninges. Image analysis of F4/80-immunostained sections revealed that the extent 1454 JID 2006:193 (15 May) Cadavid et al. Figure 3. Residual brain infection (RBI) in Toll-like receptor (TLR) 2 deficient (TLR2 / ) mice. A, SDS-PAGE; B, Western blot with anti variable small protein (Vsp1) rabbit polyclonal antibody of 3 TLR2 / mice showing RBI 1 month after intraperitoneal inoculation (lanes 2 4). Lane 1, Serotype 1 whole-cell lysate for comparison; lane 5, molecular-weight markers (in kilodaltons). B, Borrelia turicatae serotype 1 cell lysate but none of the brain culture lysates from TLR2 / mice reacting with anti-vsp1 antibody. Figure 2. Residual brain infection in immunocompetent mice. A, SDSPAGE and B, Western blot with anti variable small protein (Vsp1) polyclonal antibody of blood (lane 1) and brain (lanes 2 4) BSK II culture lysates from 4 immunocompetent mice (lanes 1 4) of different genetic backgrounds persistently infected with Borrelia turicatae (table 1). Molecular-weight markers are shown in kilodaltons on the left, and a lysate from serotype 1 (Bt1) is shown in lane 5 for comparison. B, Reactions only with the antivsp1 antibody (lanes 3 and 5). The lysate from B. turicatae serotype 2 spirochetes did not react with the anti-vsp1 antibody (data not shown). The arrow points to the 23-kDa Vsp1 in the Bt1-positive control. of macrophage/microglia activation was similar in wild-type and TLR2 / mice (table 2). These results revealed that TLR2 deficiency does not increase susceptibility to RBI or brain inflammation in RF borreliosis. Other factors can compensate for the lack of TLR2. RF neuroborreliosis in scid mice. Previous studies showed that CB17-scid mice are very susceptible to RF borreliosis, including brain infection, when they are inoculated with Bt1 [7, 16]. To investigate whether the susceptibility to brain infection in scid mice is influenced by other background genes, we compared neuroborreliosis in scid mice of 3 different genetic backgrounds CB17, C3H/HeJ, and C57BL/6 after inoculation with Bt1. In contrast to immunocompetent and TLR2 / mice, all scid mice developed persistent infection of the blood and brain and vestibular dysfunction, independently of their genetic background (table 1). SDS-PAGE analysis of positive blood and brain cultures from all scid mice showed the presence of a single VMP band of 23 kda (figure 1). As expected, Western-blot analysis with anti-vsp1 polyclonal antiserum confirmed that the band was Vsp1 (not shown). These results showed that the scid immunodeficiency significantly increases the frequency of blood and brain infection, compared with that in immunocompetent and TLR2 / mice, from just under 20% to 100% (P!.0001). Similarly, all scid mice developed clinical manifestations of vestibular dysfunction. Analysis of B. turicatae serotypes causing RBI. In contrast to the finding in scid mice, which all had a single VMP band of 23 kda (Vsp1), SDS-PAGE analysis of the 3 serotypes causing RBI in the immunocompetent mice showed that only 1 still had a 23-kDa VMP band (figure 2, lane 3). The other 2 (figure 2, lanes 2 and 4) had much larger VMP bands, corresponding to members of the Vlp families [27]. Similarly, the one serotype causing persistent blood infection in immunocompetent mice

5 Figure 4. Comparison of deduced protein sequences of variable small protein (Vsp) 1, Vsp2, and Vsp3 of Borrelia turicatae. CLUSTAL W (version 1.82) multiple sequence alignment (available at: showed that Vsp3 is very similar to Vsp1 and Vsp2 in the proximal and distal thirds, whereas the middle third is more heterogeneous. In the consensus line, an asterisk indicates strict conservation, meaning that the same residue is found in the considered position in compared amino acid sequences; a colon indicates conserved substitutions, meaning that chemically similar residues are found in the considered position in compared amino acid sequences (e.g., all basic residues [R or K], or all acidic residues [D or E]); and a period indicates semiconserved substitutions, meaning that residues found in the considered position are more variable than in the conserved case but are still considered to be chemically similar (e.g., all polar residues [R, K, D, E, S, or T]). (figure 2, lane 1) was also a Vlp member. Western-blot analysis with anti-vsp1 polyclonal antiserum revealed that the VMP band on lane 3 was Vsp1 (figure 2B). This was a brain culture from a C3H/HeJ mouse whose blood was found to be not infected at necropsy (table 1). This is, to our knowledge, the first time that RBI caused by the serotype originally inoculated has been demonstrated in RF. Analysis by SDS-PAGE of the 3 relapse serotypes causing RBI in TLR2 / mice (figure 3A, lanes 2 4) showed a similar situation: only 1 of the lanes (lane 3) had a VMP of 23 kda, identical to that of Vsp1 (lane 1). The other 2 (lanes 2 and 4) had VMP bands of 37 kda, a typical molecular size of members of the 5 Vlp families [28]. In contrast to immunocompetent mice, the 23-kDa VMP band in TLR2 / mice (figure 3B, lane 3) was not recognized by anti-vsp1 polyclonal antiserum. The implication of this finding is that a new member of the Vsp family of B. turicatae was responsible for RBI in 1 of the TLR2 / mice. This relapse serotype was cloned by limiting dilution in BSK II media [29], renamed serotype 3, and further characterized. Characterization of B. turicatae serotype 3. Serotype 3 spirochetes were found to be infectious for newly inoculated CB17-scid mice and were confirmed to be neurotropic by infecting the brain of all 3 scid mice examined (not shown). A PCR product of the predicted size was amplified from serotype 3 plasmid DNA using a forward primer specific for the B. turicatae vsp1 and vsp2 promoter region [11, 12] and a degenerated reverse primer based on a conserved sequence at the 3 end of all known vsp genes from B. hermsii and B. turicatae [27]. The deduced amino acid sequence (figure 4) revealed that Vsp3 is a member of the Vsp/OspC family of Borrelia proteins [30]. The proximal and distal thirds were highly conserved, whereas the middle third was hypervariable (figure 4). Analysis of the DNA sequence homology revealed the highest homology of vsp3 was with B. turicatae vsp1 (75%; table 3). vsp3 was 64% 74% similar to vmp genes from other RF borreliae and 62% 66% similar to ospc genes from LD borreliae (table 3). DISCUSSION The major findings of the present study were as follows. (1) RBI occurs in close to 20% of immunocompetent mice inoculated with a neurotropic strain of B. turicatae. (2) TLR2 deficiency does not increase susceptibility to RBI. (3) The scid mutation confers 100% susceptibility to RF neuroborreliosis, independently of background genes. (4) Vestibular dysfunction is the main clinical manifestation of RF neuroborreliosis in mice. (5) RBI can be caused by persistence of the original infecting serotype or by newly emerged serotypes that spontaneously arise in the population. Interest in the phenomenon of neurological involvement by Borrelia species dates to the early twentieth century, when neurological complications were frequent in patients with RF; however, research was almost abandoned after effective treatment with penicillin became available (reviewed in [1]). Earlier investigators in the field were especially interested in the ability of some RF spirochetes to persist in the brain for long time after they disappear from the blood, referred to as RBI [31, Residual Brain Infection in RF Borreliosis JID 2006:193 (15 May) 1455

6 Table 3. Genes most closely related to Borrelia turicatae vsp3. Borrelia species Disease Genes Similarity, a % B. turicatae Relapsing fever in the US vsp1 b 75 vsp2, vsp4, vsp5, vsp6 74 B. hermsii Relapsing fever in the US vsp3, vlp7, vlp21 68 vsp3, vsp22, vsp13, vsp26 67 vsp24, vsp6, vsp17, vsp25 66 B. duttoni Relapsing fever in Africa vmp (unnamed) 66 B. valaisiana Lyme disease like species from Eurasia c ospc 66 B. afzelii Lyme disease in Europe ospc 65 B. miyamotoi Relapsing fever like species from Japan c vmp (unnamed) 64 B. burgdorferi sensu stricto Lyme disease in the Western Hemisphere B. garinii Lyme disease in Europe ospc 62 ospc 62 a Determined using the Washington University Basic Local Alignment Search Tool (version 2.0). b The nomenclature for the variable major protein family of relapsing-fever borrelia genes has been changed from the original vmp to vsp (variable small protein) or vlp (variable large protein) genes, and letters have been replaced by nos. For example, Vsp1 was initially referred to as VmpA [11], then VspA [49]. c These species have been isolated from ticks and rodents, but no human infections have been reported. 32]. RBI was the subject of many investigations during the last century [1]. It was shown to occur not only in experimental animals but also in nature [33, 34]. RBI was documented in experimental animals for up to 3 years after inoculation [35]. Borreliae causing RBI are susceptible to the serum from the animal from which they are recovered, and they cannot reenter the blood from the brain [36]. In contrast to RF, RBI has not been documented in animal models of Lyme borreliosis [37 39]. In a recent study, we did not find a single case of RBI in 40 outbred mice 1 month after inoculation with several different strains of LD borreliae, most of which were isolated from human cerebrospinal fluid [40]. In contrast, brain involvement is a frequent occurrence with many species of RF borreliae [1, 6, 7, 41, 42]. A major recent advance in the field of RF has been the elucidation by Barbour et al. [43 45] that a switch in the expression of VMP proteins due to spontaneous gene conversion of silent vmp genes to a single expression locus in a linear plasmid is the underlying mechanism of antigenic variation in RF. However, whether the VMP switch influences brain infection remains to be determined. The results reported in the present article reveal that the main determinant of susceptibility to brain infection in RF is the ability to mount a specific immune response. This is in agreement with previous studies that showed that the ability to mount a specific antibody response is the main determinant of susceptibility to RF [6, 7]. Recent studies have shown that these antibodies are produced by B1b lymphocytes [46] and that they can be bactericidal in the absence of complement [47]. Although scid mice are deficient in both B and T cells, B cell deficiency is likely responsible for their increased susceptibility to RF borreliosis [48]. In support of this is our recent finding that Igh6- deficient mice, which are deficient only in B cells, are also highly susceptible to neuroborreliosis after inoculation with Bt1 (D.C. and H. Gelderblom, data not shown). In contrast to the scid mutation, we did not find any evidence that TLR2 deficiency increases susceptibility to RF neuroborreliosis: the frequency of RBI, cerebral microgliosis, and vestibular dysfunction was similar in wild-type and TLR2-deficient mice. This is in contrast to the case in Lyme borreliosis, in which TLR2 deficiency increases the severity of arthritis and the spirochetal load in some tissues [25]. Previous studies of scid mice inoculated with B. turicatae [7, 15] and Balb/c mice inoculated with Borrelia hermsii [6] showed significant differences in the ability of the individual serotypes to enter the brain. Because the only difference between isogenic Borrelia serotypes is their VMPs, this suggests that expressing certain VMPs (like Vsp1) may facilitate brain invasion. In other words, some RF VMPs may function as spirochetal neuroinvasins. This suggests that another consequence of VMP variation in RF may be a gain or loss of function with regard to brain infection. Sequence analysis of Vsp3 showed that it is closely related to Vsp1 (table 3 and figure 4). This suggests that a VMP switch from Vsp1 to Vsp3 may have allowed B. turicatae to escape killing while still preserving its ability to infect the brain. Sequence alignment of B. turicatae Vsp1, 2, and 3 showed they are very similar in the proximal and distal regions, with most of the variability concentrated in the middle third (figure 4). Localized polymorphisms in this middle region of Vsp [49] may account for differences of individual Borrelia serotypes in the ability to infect the brain. In previous studies of Balb/c mice inoculated with B. hermsii serotype 7, we found that it had been cleared not only from the blood but also from the brain by the time of the first relapse [6]. In this study, we also found that an established serotype 7 brain infection was cleared from the brain of irradiated Balb/ 1456 JID 2006:193 (15 May) Cadavid et al.

7 c mice by the intravenous administration of anti-vsp7 monoclonal antibodies of the IgG but not of the IgM isotype [6]. This demonstrated that some antibodies are better than others at clearing spirochetes from the brain. However, we have now found Bt1 in the brain of an immunocompetent mouse that had been inoculated with Bt1 1 month earlier. This is, to our knowledge, the first demonstration that the original infecting serotype can somehow escape the host s antibody response and survive in the brain. The implication is that, occasionally, immunocompetent hosts fail to clear spirochetes from the brain, resulting in RBI. The mechanism that allowed Bt1 to survive in the brain but not in the blood remains to be determined. Because the spirochetes cultured from the brain were expressing Vsp1, it is more likely that the antibody response of this particular C3H/HeJ mouse had difficulty reaching or killing the spirochetes that had entered the brain. Although the number of mice was small, it was also interesting that one of the outbred mice developed vestibular dysfunction as severe as that of scid mice. This also implies that some hosts, who apparently are immunocompetent, fail to control RF borreliosis, resulting in neurological complications. Contrary to the case in humans, in whom neuroborreliosis involves predominantly the seventh cranial nerve both in RF and LD [3], in mice, RF neuroborreliosis predominantly affects the vestibular system [7, 15]. The present results reveal that persistent brain infection in RF borreliosis is not only a feature of antibody-deficient mice but that it also occurs in a significant percentage of immunocompetent mice. RBI can be caused by the persistence in the brain of the original or relapse serotypes after they had been eliminated from the blood. Future studies may elucidate how variations in VMPs modulate the ability of RF borreliae to enter the brain. Acknowledgment We thank Dr. Shizuo Akira (Osaka University, Japan), for his gift of the Toll-like receptor 2 deficient mice. References 1. Cadavid D, Barbour AG. Neuroborreliosis during relapsing fever: review of the clinical manifestations, pathology, and treatment of infections in humans and experimental animals. Clin Infect Dis 1998; 26: Pachner AR, Cadavid D. Lyme neuroborreliosis. In: Griggs RC, Joynt RJ, eds. Clinical neurology. Philadelphia: Lippincott-Raven, 1998: Cadavid D. Lyme disease and relapsing fever. In: Scheld W, Whitley R, Marra C, eds. Infections of the central nervous system. Philadelphia: Lippincott-Raven, 2004: Taft W, Pike J. Relapsing fever: report of a sporadic outbreak including treatment with penicillin. JAMA 1945; 129: Bergeret C, Raoult A. Notes sur les formes nerveuses de la fievre recurrente fievre recurrente a tiques en Afrique Occidentale Francaise. Bulletin Medicale de l Afrique Occidentale Francaise 1948; 5: Cadavid D, Bundoc V, Barbour AG. Experimental infection of the mouse brain by a relapsing fever Borrelia species: a molecular analysis. J Infect Dis 1993; 168: Cadavid D, Thomas DD, Crawley R, Barbour AG. Variability of a bacterial surface protein and disease expression in a possible mouse model of systemic Lyme borreliosis. J Exp Med 1994; 179: Cadavid D, O Neill T, Schaefer H, Pachner AR. Localization of Borrelia burgdorferi in the nervous system and other organs in a nonhuman primate model of Lyme disease. Lab Invest 2000; 80: Plasterk RH, Simon MI, Barbour AG. Transposition of structural genes to an expression sequence on a linear plasmid causes antigenic variation in the bacterium Borrelia hermsii. Nature 1985; 318: Stoenner H, Dodd T, Larsen C. Antigenic variation of Borrelia hermsii. J Exp Med 1982; 156: Cadavid D, Pennington PM, Kerentseva TA, Bergstrom S, Barbour AG. Immunologic and genetic analyses of VmpA of a neurotropic strain of Borrelia turicatae. Infect Immun 1997; 65: Pennington PM, Cadavid D, Barbour AG. Characterization of VspB of Borrelia turicatae, a major outer membrane protein expressed in blood and tissues of mice. Infect Immun 1999; 67: Pennington PM, Cadavid D, Bunikis J, Norris SJ, Barbour AG. Extensive interplasmidic duplications change the virulence phenotype of the relapsing fever agent Borrelia turicatae. Mol Microbiol 1999; 34: Barbour AG, Hayes SF. Biology of Borrelia species. Microbiol Rev 1986; 50: Cadavid D, Pachner AR, Estanislao L, Patalapati R, Barbour AG. Isogenic serotypes of Borrelia turicatae show different localization in the brain and skin of mice. Infect Immun 2001; 69: Londoño D, Bai Y, Zuckert W, Gelderbrom H, Cadavid D. Cardiac apoptosis in severe relapsing fever borreliosis. Infect Immun 2005; 76: Barbour AG, Tessier SL, Stoenner HG. Variable major proteins of Borrelia hermsii. J Exp Med 1982; 156: Sadziene A, Rosa PA, Thompson PA, Hogan DM, Barbour AG. Antibody-resistant mutants of Borrelia burgdorferi: in vitro selection and characterization. J Exp Med 1992; 176: Schaible UE, Kramer MD, Wallich R, Tran T, Simon MM. Experimental Borrelia burgdorferi infection in inbred mouse strains: antibody response and association of H-2 genes with resistance and susceptibility to development of arthritis. Eur J Immunol 1991; 21: Barthold SW, Beck DS, Hansen GM, Terwilliger GA, Moody KD. Lyme borreliosis in selected strains and ages of laboratory mice. J Infect Dis 1990; 162: Barstad PA, Coligan JE, Raum MG, Barbour AG. Variable major proteins of Borrelia hermsii: epitope mapping and partial sequence analysis of CNBr peptides. J Exp Med 1985; 161: Vidal V, Scragg IG, Cutler SJ, et al. Variable major lipoprotein is a principal TNF-inducing factor of louse-borne relapsing fever. Nat Med 1998; 4: Fraser CM, Casjens S, Huang WM, et al. Genomic sequence of a Lyme disease spirochaete, Borrelia burgdorferi. Nature 1997; 390: Lien E, Sellati TJ, Yoshimura A, et al. Toll-like receptor 2 functions as a pattern recognition receptor for diverse bacterial products. J Biol Chem 1999; 274: Wooten RM, Ma Y, Yoder RA, et al. Toll-like receptor 2 is required for innate, but not acquired, host defense to Borrelia burgdorferi. J Immunol 2002; 168: Perry VH, Hume DA, Gordon S. Immunohistochemical localization of macrophages and microglia in the adult and developing mouse brain. Neuroscience 1985; 15: Hinnebusch BJ, Barbour AG, Restrepo BI, Schwan TG. Population structure of the relapsing fever spirochete Borrelia hermsii as indicated by polymorphism of two multigene families that encode immunogenic outer surface lipoproteins. Infect Immun 1998; 66: Burman N, Bergstrom S, Restrepo BI, Barbour AG. The variable antigens Vmp7 and Vmp21 of the relapsing fever bacterium Borrelia Residual Brain Infection in RF Borreliosis JID 2006:193 (15 May) 1457

8 hermsii are structurally analogous to the VSG proteins of the African trypanosome. Mol Microbiol 1990; 4: Barbour A. Isolation and cultivation of Lyme disease spirochetes. Yale J Biol Med 1984; 57: Carter CJ, Bergstrom S, Norris SJ, Barbour AG. A family of surfaceexposed proteins of 20 kilodaltons in the genus Borrelia. Infect Immun 1994; 62: Rothermundt M. Ueber die beziehungen zwischen virulenz und persistenz der rekurrensspirochaten im gehirn weisser mause. Arbeiten aus dem Staatsins fur Experimentelle Therapie und dem Georg-Speyer-Haus zu Frankfurt AM 1928: Levaditi C, Anderson T, Selbie F, Schoen R. L encephalite recurrentielle. Bull Acad Med 1930; 103: Diatta G, Trape J, Legros F, Rogier C, Duplantier J. A comparative study of three methods of detectin of Borrelia crocidurae in wild rodents in Senegal. Trans R Soc Trop Med Hyg 1994; 88: Horrenberger R. Spirochaeta hispanica chez les rats d Alger (nouvelle enquete). Arch Inst Pasteur Alger 1954; 32: Sergent E. Persistence de spirochaete hispanica pendant trois ans dans le cerveau d un cobaye. 3e note. Arch Inst Pasteur Alger 1945; 23: Steiner G, Steinfeld J, Schauder H. Zur frage der spirochatenpersistenzim zentralnervensystem bei experimenteller recurrens. Klinische Wochenschrift 1925; 4: Schaible UE, Gay S, Museteanu C, et al. Lyme borreliosis in the severe combined immunodeficiency (scid) mouse manifests predominantly in the joints, heart, and liver. Am J Pathol 1990; 137: Barthold SW, de Souza MS, Janotka JL, Smith AL, Persing DH. Chronic lyme borreliosis in the laboratory mouse. Am J Pathol 1993; 143: Pachner AR, Itano A. Borrelia burgdorferi infection of the brain: characterization of the organism and response to antibiotics and immune sera in the mouse model. Neurology 1990; 40: Pachner AR, Dail D, Bai Y, et al. Genotype determines phenotype in experimental Lyme neuroborreliosis. Ann Neurol 2004; 56: Nordstrand A, Shamaei-Tousi A, Ny A, Bergstrom S. Delayed invasion of the kidney and brain by Borrelia crocidurae in plasminogen-deficient mice. Infect Immun 2001; 69: Gebbia JA, Monco JC, Degen JL, Bugge TH, Benach JL. The plasminogen activation system enhances brain and heart invasion in murine relapsing fever borreliosis. J Clin Invest 1999; 103: Meier J, Simon M, Barbour A. Antigenic variation is associated with DNA rearrangements in a relapsing fever borrelia. Cell 1985; 41: Barbour A. Antigenic variation of a relapsing fever Borrelia species. Annu Rev Microbiol 1990; 44: Barbour AG, Burman N, Carter CJ, Kitten T, Bergstrom S. Variable antigen genes of the relapsing fever agent Borrelia hermsii are activated by promoter addition. Mol Microbiol 1991; 5: Alugupalli KR, Leong JM, Woodland RT, Muramatsu M, Honjo T, Gerstein RM. B1b lymphocytes confer T cell-independent long-lasting immunity. Immunity 2004; 21: Connolly SE, Thanassi DG, Benach JL. Generation of a complementindependent bactericidal IgM against a relapsing fever Borrelia. J Immunol 2004; 172: Newman K Jr, Johnson RC. T-cell-independent elimination of Borrelia turicatae. Infect Immun 1984; 45: Zuckert WR, Kerentseva TA, Lawson CL, Barbour AG. Structural conservation of neurotropism-associated VspA within the variable Borrelia Vsp-OspC lipoprotein family. J Biol Chem 2001; 276: JID 2006:193 (15 May) Cadavid et al.