Outer surface lipoproteins of Borrelia burgdorferi vary in their ability to induce experimental joint injury.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 50, No. 7, July 2004, pp 2360–2369 DOI 10.1002/art.20337 © 2004, American College of Rheumatology Outer Surface Lipoproteins of Borrelia burgdorferi Vary in Their Ability to Induce Experimental Joint Injury Stephen Batsford,1 John Dunn,2 and Michael Mihatsch3 Objective. To examine the ability of bacterial lipoproteins from the spirochete Borrelia burgdorferi to cause in vivo tissue injury (arthritis). Methods. Outer surface proteins (OSPs) from B burgdorferi were used in a rat model of antigen-induced allergic arthritis. Intraarticular challenge with recombinant OspA, OspB, and OspC in nonlipidated (peptide) and lipidated forms was performed in the left knee joint; the contralateral joint received buffer as control. Inflammation was monitored by technetium scintigraphy and histology. Results. Nonlipidated (peptide) OspA, OspB, and OspC did not induce arthritis; the only exception was polymerized OspA, which was tested in preimmunized rats. Lipidated OspA from 2 different strains and lipidated OspC induced severe arthritis, whereas lipidated OspB failed to induce injury. A synthetic analog of the OSP lipid modification, lipopeptide Pam3Cys-SerLys4-OH, either alone or coupled to bovine serum albumin, also failed to induce injury. Injury did not develop in control groups that were given the appropriate buffers or lipopolysaccharide. This showed that lipidated borrelial OSPs can be potent arthritogens but vary greatly with respect to their injury-inducing potential. The possession of a lipid modification is essential but is not sufficient to render an OSP arthritogenic. Conclusion. This is the first study to demonstrate that individual lipoproteins from B burgdorferi can induce experimental joint injury in vivo. These results may help elucidate the pathogenesis of Lyme arthritis and, above all, underline the importance of bacterial lipoproteins as major virulence factors. Lipoproteins are now being recognized as prime virulence factors of bacteria. For many decades this group of molecules played the role of poor relation to the lipopolysaccharides (LPS), but particularly the discovery that lipoproteins can act as ligands for the Toll-like receptor (TLR) complex has rekindled interest (1). In earlier studies, a number of naturally occurring bacterial lipoproteins and their synthetic lipopeptide analogs were shown to possess impressive biologic activities. The Braun protein, a prototype lipoprotein, can stimulate lymphocytes (2), and the corresponding synthetic lipopeptide is in fact now used as an adjuvant (3). Membrane lipoproteins and the corresponding lipopeptides from mycoplasma could induce mononuclear phagocytes to secrete cytokines; a particular lipopeptide from Mycoplasma fermentans, macrophage activating lipopeptide 2, was effective even at concentrations in the picogram range (4). The outer surface lipoproteins (OSPs) of Borrelia burgdorferi have been shown to exhibit similar stimulatory effects (5–8). Results of recent in vitro studies indicated that the lipid moiety plays a key role, and that the CD14/ TLR-2 receptors can be involved (9–11). Lipoproteins have been shown to be capable of inducing tissue injury in animals, mimicking certain common clinical manifestations of the corresponding infection in man. First, synthetic lipopeptide analogs of treponemal and borrelial lipoproteins caused acute dermal lesions in mice (12). Second, an OSP complex extracted from B burgdorferi induced severe, chronic arthritis in rats (13). In the current study, 3 major OSPs (OspA, OspB, and OspC) derived from B burgdorferi were studied. Their ability to induce arthritis in a rat model of antigeninduced allergic arthritis (14) was analyzed in detail, using a battery of purified recombinant antigens. The injury-inducing potential of the lipid modification and 1 Stephen Batsford, PhD: Albert Ludwigs University, Freiburg, Germany; 2John Dunn, PhD: Brookhaven National Laboratory, Upton, New York; 3Michael Mihatsch, MD: Institute of Pathology, Kantonspital Basel, Basel, Switzerland. Address correspondence and reprint requests to Stephen Batsford, PhD, Department of Immunology, Institute of Medical Microbiology, University of Freiburg, Hermann-Herder-Strasse 11, D-79104, Freiburg, Germany. E-mail: email@example.com. Submitted for publication July 21, 2003; accepted in revised form March 5, 2004. 2360 INDUCTION OF EXPERIMENTAL JOINT INJURY IN VIVO BY OSPs FROM B BURGDORFERI the peptide regions was evaluated, and the role of the immune response to the OSPs in inducing inflammation in vivo was studied. Because the pathogenesis of organ manifestations of B burgdorferi infection is still quite unclear, this information may provide valuable insights into the mechanisms involved. MATERIALS AND METHODS Animals. Male Wistar rats (Harlan, Borchen, Germany) were used throughout. Body weight ranged from 100 to 120 gm at the beginning of the experiments. Animal experiments were approved by the local government agency (license no. G-00/43). Antigens. The following borrelial OSPs were prepared as both truncated nonlipidated (peptide) and full-length lipidated forms (the strain[s] of origin of the sequence is shown in parentheses): OspA (B31 and K48, both B burgdorferi sensu stricto), OspB (B31), and OspC (B31-lipidated using the leader sequence from OspB of B31, K48-truncated). The details of plasmid construction, expression systems, and purification of recombinant products have been published elsewhere (15). Expression of OspC with the original leader sequence could not be achieved due to the occurrence of severe toxicity problems in the expression system, which were overcome by replacing the OspC leader with OspB (Dunn J, Lade B: unpublished observations). The 83-kd structure protein (synonyms, 94- and 100-kd protein) of B burgdorferi sensu stricto (strain GeHo) was prepared in recombinant form as previously described (16). All of the truncated OSP preparations (OspA, OspB, OspC) and the 83-kd protein were water soluble; all lipidated OSP preparations were solubilized in phosphate buffered saline (PBS) containing 0.1% Triton X-100. All preparations were screened for contamination with LPS by the Limulus coagulation test (Acila, Frankfurt, Germany); concentrations found were between 0.1% and 1.0%. The water-soluble lipopeptide Pam3Cys-Ser-Lys4OH (molecular weight 1,620 daltons (Boehringer, Mannheim, Germany) was used. When required, it was covalently coupled to bovine serum albumin (BSA), as described below. LPS, prepared from Escherichia coli EH100, was used as an aqueous solution at 200 g/ml. Protein polymerization. Truncated recombinant OspA and OspB (both derived from strain B31) were polymerized with dimethyl suberimidate, a bifunctional iminoester that crosslinks lysine side chains without altering the net charge (17,18). The OSP preparation was brought to a concentration of 10 mg/ml in 0.1M NaHPO4 (pH 10.0), and then 2 mg of dimethyl suberimidate per milliliter of protein solution was added in portions over a 30-minute period, with stirring, at 20°C. The reaction mixture was dialyzed against PBS at 4°C, centrifuged at 1,000g for 10 minutes, and the supernatant was stored at ⫺20°C. The synthetic lipopeptide Pam3Cys-Ser-Lys4-OH was coupled to BSA to create an analog lipoprotein containing a nonarthritogenic peptide moiety. For this purpose, BSA at concentrations of 10 mg/ml and 20 mg/ml was coupled with Pam3Cys-Ser-Lys4-OH at concentrations of 0.5 mg/ml and 1 2361 mg/ml, respectively, as described above. Before use, the conjugate was diluted to a BSA concentration of 1 mg/ml. Size-exclusion chromatography. Polymerized OspA and OspB were separated on a HiLoad 26/60 Superdex 75 prep grade column run on an FPLC system (Amersham Biosciences, Freiburg, Germany). The Superdex column was equilibrated in PBS, samples containing 10 mg of protein were loaded and run in PBS at a flow rate of 3 ml/minute, 5-ml fractions were collected, and the effluent was monitored at 280 nm. Selected fractions were pooled and concentrated in an ultrafiltration cell using YM10 membranes (Millipore, Bedford, MA). Analysis by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (see below) showed that both polymerized OspA and polymerized OspB could be separated into monomeric, dimeric, and oligomeric (trimer and larger) fractions. SDS-PAGE. Protein preparations were analyzed in 12% acrylamide gels and stained with Coomassie blue R250, using molecular mass standards in the range of 14–93 kd (Amersham Biosciences) (19). Arthritis induction. Test substances (usually 50 g of test protein in 50 l of buffer) were injected directly into the left knee joint; the contralateral right knee joint received 50 l of buffer alone as a control. In the case of LPS, the dose was reduced to 10 g. In some experiments, rats were preimmunized by injecting 100 g of antigen, emulsified in Freund’s complete adjuvant, subcutaneously into the dorsal region to avoid joint interference. Two booster injections with the same dose were given at 2-week intervals, and experiments were initiated 1–2 weeks later. Measurement of delayed-type hypersensitivity. Delayedtype hypersensitivity was measured in the experimental groups by injecting 5 g of antigen in 5 l of buffer intracutaneously into the pinna of the ear; the contralateral ear was injected with PBS as a control. Increases in ear thickness were measured with an engineer’s micrometer after 48 hours, and results were expressed as the mean ⫾ SEM increase (mm ⫻ 10⫺2). 99m Tc scintigraphic uptake measurements. Joint inflammation was assessed by comparing the uptake of 99mTcpertechnetate between knee joints (20). Each rat received 2 MBq of 99mTc in 0.2 ml of PBS, subcutaneously into the neck region. Gamma radiation emanating from both knee joints of the hind legs was measured using a collimated photoscintillation crystal. The rats’ hind legs were kept in a defined position, and their bodies were shielded with a lead screen. Radioactivity was measured 3 times over each joint in an alternating manner, for a duration of 0.5 minutes, and the mean values were used to calculate the left:right (L:R) ratio. Joint retention of 131I-labeled OSP preparations. Onemilligram portions of nonlipidated OspA, lipidated OspA, and lipidated OspB were labeled with 131I (Amersham Biosciences) by the chloramine-T–method (21). Free 131I was removed by passage through a Sephadex PD-10 column (Amersham Biosciences). Before use, 131I-labeled samples were mixed with unlabeled, chloramine-T–treated protein, and each rat received 25 g of the 131I-labeled sample (2 ⫻ 104 Bq) directly into the left knee joint; the right knee joints received PBS at 1, 10, and 60 minutes and at 2, 6, and 24 hours. Radioactivity was measured over both knee joints with a collimated counter, in a manner similar to that used to measure 99mTc uptake (see above). 2362 BATSFORD ET AL Table 1. Effect of intraarticular challenge with truncated, nonlipidated outer surface proteins (OSPs) Group 1 2 3 4 5 6 (n (n (n (n (n (n ⫽ ⫽ ⫽ ⫽ ⫽ ⫽ 5) 5) 5) 5) 4) 4) Technetium scintigraphy‡ Challenge preparation (sequence)* Immunized (DTH test)* Histology† Result (no./total) Value, L:R ratio Monomeric OspA (B31) Monomeric OspB (B31) Monomeric OspC (K48) Polymerized OspA (B31) Polymerized OspA (B31) Polymerized OspB (B31) Yes Yes No No Yes (42 ⫾ 5.3) Yes (22 ⫾ 7.5) Normal Normal Normal Normal Injury Normal Normal (5/5) Normal (5/5) Normal (4/5) Normal (5/5) Raised (4/4) Normal (3/4) 1.10 ⫾ 0.03 1.09 ⫾ 0.02 1.16 ⫾ 0.02 1.03 ⫾ 0.02 1.37 ⫾ 0.03§ 1.07 ⫾ 0.06 * Values for delayed-type hypersensitivity (DTH) testing are the mean ⫾ SEM increase in ear thickness (mm ⫻ 10⫺2) measured at 48 hours (for control ears, the increase was 2.9 ⫾ 1.3 mm ⫻ 10⫺2). Intraarticular challenge and immunization were with the same antigen for all B31 sequences. † In each group, 2 additional rats were killed for histologic examination on days 2 and 7, respectively; classification is for both rats. ‡ The L:R ratio represents 99mTc uptake in left:right knee joints on day 2 (number of rats with a ratio of ⬍1.2 [cutoff level] for normal and ⬎1.2 for raised values/number in group; in groups 3 and 6, one rat had borderline levels (1.2–1.25). Values are the mean ⫾ SEM. § P ⫽ 0.05 versus control groups 4 and 7 in Table 3. Histology. In each group studied (see Tables 1, 2, and 3), animals were killed on days 2 and 7 after intraarticular challenge, for histologic examination. Knee joints were detached and fixed in 4% buffered formalin, decalcified, and embedded in paraffin. Standard frontal sections of the joints were stained with hematoxylin and eosin. Statistical analysis. Data from scintigraphic measurements revealed varying inhomogeneity and were subjected to a log-log transformation followed by analysis of variance. Differences between groups were tested by Tukey’s studentized range test, and within-group differences between days 2 and 7 were tested by paired t-tests. P values less than 0.05 were considered significant. Statistical analysis was performed using the general linear model procedure plus t-test functions in the SAS program, release 6.12 (SAS Institute, Cary, NC). Data on retention of OSPs in knee joints were compared using Fisher’s t-test for unpaired data. RESULTS Effect of intraarticular challenge with nonlipidated (truncated) OSPs. The preparations tested, the numbers of rats tested and their status (naive or preimmunized), and the ability of the preparation to induce joint inflammation are shown in Table 1 (the 99mTc scintigraphic L:R ratios shown are for day 2 only). As can be seen in Table 1, monomeric, truncated (nonlipidated), water-soluble OspA, OspB, and OspC did not induce injury (groups 1, 2, and 3). Polymerized samples were also tested, because, in the model used, the size of the test molecule is known to be an important parameter influencing induction of arthritis (20,22). Polymerized Table 2. Effect of intraarticular challenge with full-length, lipidated outer surface proteins (OSPs) Technetium scintigraphy‡ Group 1 2 3 4 5 6 (n (n (n (n (n (n ⫽ ⫽ ⫽ ⫽ ⫽ ⫽ 6) 6) 4) 6) 5) 3) 7 (n ⫽ 6) Histology† Result (no./total) Value, L:R ratio Yes (56 ⫾ 7.9) No BSA No Yes (48 ⫾ 4.6) No Injury Injury Injury Injury Borderline injury Borderline injury Raised (6/6) Raised (6/6) Raised (4/4) Raised (4/6) Normal (5/5) Normal (3/3) 1.59 ⫾ 0.11§ 1.61 ⫾ 0.07§ 1.45 ⫾ 0.12§ 1.30 ⫾ 0.07§ 1.12 ⫾ 0.02 1.10 ⫾ 0.08 No Injury Raised (4/6) 1.34 ⫾ 0.12§ Challenge preparation (sequence)* Immunized (DTH test)* Lipidated OspA (B31) Lipidated OspA (B31) Lipidated OspA (B31) Lipidated OspA (K48) Lipidated OspB (B31) Lipidated OspB (B31), 150 g (triple dose) Lipidated OspC (B31) ⫺2 * Values for delayed-type hypersensitivity (DTH) testing are the mean ⫾ SEM increase in ear thickness (mm ⫻ 10 ) measured at 48 hours (for control ears, the increase was 2.9 ⫾ 1.3 mm ⫻ 10⫺2). Intraarticular challenge and immunization were with the same antigen, except where indicated otherwise. BSA ⫽ bovine serum albumin. † In groups 3, 4, and 6, two additional rats were killed (on days 2 and 7, respectively) for histologic examination; classification is for both rats. In groups 1, 2, and 5, four additional rats were killed (two on day 2, and two on day 7), and in group 6, three additional rats were killed (two on day 2, and one on day 7). See text for detailed description. ‡ The L:R ratio represents 99mTc uptake in left:right knee joints on day 2 (number of rats with L:R ratios of ⬍1.2 [cutoff level] for normal and ⬎1.2 for raised values/number in group). In groups 4 and 7, two rats had an L:R ratio of ⬍1.2. Values are the mean ⫾ SEM. § P ⫽ 0.05 versus control groups 4 and 7 in Table 3. INDUCTION OF EXPERIMENTAL JOINT INJURY IN VIVO BY OSPs FROM B BURGDORFERI 2363 Table 3. Effect of intraarticular challenge with synthetic lipopeptides, conjugates, and control preparations Technetium scintigraphy‡ Group 1 2 3 4 5 6 7 (n (n (n (n (n (n (n ⫽ ⫽ ⫽ ⫽ ⫽ ⫽ ⫽ 4) 4) 5) 4) 4) 4) 8) Challenge preparation (sequence)* Immunized (DTH test)* Histology† Lipopeptide Pam3CSK4 Pam3CSK4 BSA conjugate 83-kd protein (GeHo) Triton X-100, 0.1% Triton X-100, 1.0% Lipopolysaccharide Phosphate buffered saline No No Yes (36 ⫾ 5.4) No No No No Normal Normal Normal Normal Normal Normal Normal Result (no./total) Normal Normal Normal Normal Normal Normal Normal (4/4) (4/4) (5/5) (4/4) (3/4) (4/4) (8/8) Value, L:R ratio 1.12 ⫾ 0.02 1.05 ⫾ 0.02 1.04 ⫾ 0.01 1.04 ⫾ 0.03 1.11 ⫾ 0.06 1.14 ⫾ 0.02 1.01 ⫾ 0.02 * Values for delayed-type hypersensitivity (DTH) testing are the mean ⫾ SEM of the increase in ear thickness (mm ⫻ 10⫺2) measured at 48 hours (for control ears, an increase of 2.9 ⫾ 1.3 mm ⫻ 10⫺2). Intraarticular challenge and immunization were with the same antigen (all B31 sequences), except where indicated otherwise. Lipopeptide Pam3CSK4 ⫽ Pam3Cys-Ser-Lys4-OH; BSA ⫽ bovine serum albumin. † In addition to the rats used for scintigraphy, 1 rat was killed for histologic examination on day 2 and day 7, respectively; classification is for both rats. ‡ The L:R ratio represents 99mTc uptake in left:right knee joints on day 2 (number of rats with an L:R ratio of ⬍1.2 [cutoff level]/number in group). In group 5, one rat had borderline levels (1.2–1.25). Values are the mean ⫾ SEM. No significant differences between groups were observed. OspB (group 6) did not induce arthritis, even in immunized rats, whereas oligomeric OspA induced a transient arthritis (that persisted for a maximum of 7 days) in previously immunized rats (group 5). Examination of joint sections from group 5 revealed moderate swelling of the synovial membrane and limited polymorphonuclear neutrophil infiltration, without pronounced cartilage destruction. Due to the limited quantities available, OspC was not tested in the polymerized form. Effect of intraarticular challenge with fulllength, lipidated OSPs. The preparations tested, the numbers of rats tested and their status (naive or preimmunized), and the ability of the preparations to induce joint inflammation are shown in Table 2 (the 99mTc scintigraphic L:R ratios shown are for day 2 only). Lipidated OspA derived from the strain B31 sequence induced severe arthritis in both naive and preimmunized rats (Table 2 and Figure 1). The severity of injury induced was similar in naive rats (group 2), in rats preimmunized with lipidated OspA (group 1), and in rats preimmunized with BSA (group 3) (Table 2). Histologic examination revealed massive swelling and cellular infiltration of the synovial membrane, together with focal, complete destruction of the cartilage surface in all of the rats examined (Table 2). Some rats in these groups had persisting arthritis and were followed up. An example of the severe changes seen on day 29 is shown in Figures 2A and B. A second lipidated OspA derived from the strain K48 sequence, and lipidated OspC derived from the strain B31 sequence also induced joint inflammation in naive rats (groups 4 and 7, respectively; Table 2); preimmunized rats were not tested. In groups 4 and 7, the 99mTc scintigraphic values were lower than those in rats given lipidated OspA derived from the strain B31 sequence (groups 1–3); the differences did not reach significance, however, and the histologic changes were qualitatively similar but milder. Figure 1. Ratio of the uptake of 99mTc in the right and left knee joints (L:R ratio) of groups of rats injected with different lipidated (Lip.) outer surface protein (OSP) preparations on days 2 and 7. On day 2, values in rats receiving lipidated OspA (sequences B31 and K48) and lipidated OspC (B31) were significantly raised, and values in rats receiving lipidated OspB (B31) were not. By day 7, only values in groups receiving lipidated OspA (B31) and lipidated OspC (B31) were significantly raised. Values are the mean ⫾ SEM simple ratios. 2364 BATSFORD ET AL Figure 2. Histologic appearance of knee joints from rats challenged with lipidated outer surface protein (OSP) preparations. A, Example of chronic joint injury induced by lipidated OspA (B31) in a preimmunized rat 29 days after challenge, showing severe chronic inflammation of the synovial membrane with focal complete destruction of the cartilage. B, Higher magnification of area in A. C, Example of minimal damage induced by lipidated OspB (B31) in some rat knee joints 2 days after challenge. Low power magnification shows no significant abnormalities. D, Higher magnification of area in C shows discrete area of slight chronic inflammation of the synovial membrane. (Hematoxylin and eosin stained; original magnification ⫻ 16 in A and C; ⫻ 80 in B and D.) Intraarticular injection of lipidated OspB derived from B31 sequence induced only borderline injury in preimmunized rats (group 5). There was no increased uptake of 99mTc in the joint challenged at any time point examined (Figure 1), the histologic picture was normal in 2 of the 4 rats examined, and changes were minimal in the other 2 rats (Figures 2C and D). In an additional experiment in naive rats, increasing the dose of lipidated OspB injected from 50 g to 150 g (group 6) did not increase the uptake of 99mTc (Table 2), and histologic examination still revealed only minimal changes in 2 of the 3 rats examined (the other rat was normal). Effects of synthetic lipopeptides and LPS. The water-soluble lipopeptide Pam3Cys-Ser-Lys4-OH (see Materials and Methods) did not produce joint inflammation following intraarticular injection (group 1; Table 3). In a further experiment, lipopeptide Pam3Cys-SerLys4-OH was covalently coupled to BSA to produce an analog of a larger lipoprotein; this preparation also failed to induce joint injury (group 2; Table 3). No joint INDUCTION OF EXPERIMENTAL JOINT INJURY IN VIVO BY OSPs FROM B BURGDORFERI 2365 antigens used for immunization (see Tables 1, 2, and 3), without subsequent antigen challenge, could cause arthritis. Retention of lipidated OspA and OspB in joints. The kinetics of elimination of truncated nonlipidated and full-length lipidated OspA, as well as full-length lipidated OspB, are shown in Figure 3. No significant differences were observed within the 24-hour observation period. DISCUSSION Figure 3. External radioactivity measurements at various times after intraarticular injection of 25 g of 131I-labeled outer surface protein (OSP) preparations. Values are the mean ⫾ SEM of groups of 4 rats. There were no significant differences at any time point studied. injury was seen after intraarticular challenge with 10 g of LPS (group 6; Table 3). This quantity of LPS far exceeds that present as a contaminant in the OSP preparations (0.05–0.5 g of LPS per 50 g of OSP; see Materials and Methods). Control experiments. For comparison, a group of rats were immunized and challenged with recombinant 83-kd protein derived from strain GeHo of B burgdorferi sensu stricto (cytoplasmic structure protein [synonyms, 94-kd and 100-kd protein]), and no injury was induced (group 3; Table 3). Because the full-length lipidated OSP preparations were solubilized in 0.1% Triton X-100, it was important to exclude an effect of this detergent on the joint. A group of 4 rats were given intraarticular injections of 50 l of 0.1% Triton X-100, and no injury was produced (group 4; Table 3 and Figure 1). Increasing the concentration of Triton X-100 to 1% resulted in only a borderline increase in 99mTc uptake and minimal histologic changes in 1 of 4 rats (group 5; Table 3). Rats given only PBS (group 7) showed no evidence of injury. The major control in the experimental system used was the contralateral knee joint, which received only buffer (PBS or 0.1% Triton X-100), and no signs of inflammation were seen in any of these joints when they were examined histologically. This latter finding also effectively excludes the possibility that induction of a strong immune response to the OSP Only limited information on the potential of bacterial lipoproteins to induce tissue injury in vivo is available. Synthetic lipohexapeptide analogs of treponemal and borrelial lipoproteins have been shown to cause acute skin lesions in mice (12). We previously demonstrated that a model of chronic arthritis could be induced in rats with a butanol-extracted, water-soluble OSP complex from B burgdorferi (13). That observation prompted the current set of experiments, which represent the first study in which the pathogenicity of individual recombinant versions of OSPs was directly studied in vivo. Differences in the arthritogenic potential of the 3 OSPs studied, OspA, OspB, and OspC, could be shown. The concept that borrelial lipoproteins may be key virulence factors in Lyme disease is also supported by in vitro data demonstrating that lipidated OspA induced proliferation and apoptosis of astrocytes (astrogliosis) (23), and might thus contribute to the pathogenesis of neuroborreliosis. The authors of that study obtained a similar result using a synthetic analog of the lipid modification Pam3Cys, showing that this moiety also played a central role in this context. The model used here, known as antigen-induced allergic arthritis, allows individual testing of selected substances. Cationic molecules are particularly effective in this experimental setting (24), which has, in part, been attributed to their prolonged persistence following attachment to anionic structures in the joint, as originally shown for chemically cationized antigens (20,22), and subsequently extended to a cationic antigen from the bacillus Yersinia enterocolitica (25). In the case of B burgdorferi, both individual OSPs and the OSP complex originally prepared by butanol extraction are cationic (26). The results presented here allow some conclusions to be drawn concerning properties of borrelial OSPs that may confer pathogenicity. Clearly, the lipid modification plays a central role, because its absence drastically reduced the severity of arthritis induced in 2366 rats. The observation that synthetic lipopeptide analogs of treponemal and borrelial lipoprotein induced dermal inflammation in vivo (12) provides direct experimental support for this viewpoint. In addition, as discussed above (23), a synthetic analog of the lipid moiety was shown to be effective in an in vitro test system (23), in which it could induce astrogliosis. However, in the current study, a synthetic analog of the lipid moiety alone failed to induce joint injury, which could indicate that the mechanisms involved are organ/structure specific. The latter observation cannot be simply explained by the difference in the size of the molecules (31 kd for monomeric lipidated OspA, and 1.6 kd for lipopeptide Pam3Cys-Ser-Lys4-OH). First, coupling the synthetic lipopeptide to a larger protein (BSA) (product in monomeric form ⬃69 kd) failed to produce an arthritogen. Second, lipidated OspB (34 kd) did not produce injury, even in preimmunized rats, in spite of possession of the same lipid modification. Even challenge with the 3-fold dose of lipidated OspB (raised from 50 g to 150 g) failed to cause significant injury. The leader sequence of OspB was, for technical reasons (see Materials and Methods), also used to produce recombinant, lipidated OspC. This latter molecule also induced arthritis, which indicates that the lipid modifications of OspA and OspB differ little in respect to injury-inducing potential. In contrast, the peptide portions of the OSP alone are not very effective in inducing injury. When nonlipidated, pure-peptide OspA and OspB were tested, only polymerized peptide OspA derived from strain B31 induced mild, transient injury and then only in preimmunized rats. All of these data indicate that the fatty acid modification is essential, but not sufficient, for inducing injury; the peptide region appears to contribute. The evidence suggests that the OspB peptide sequence may differ in a critical manner. On comparing some relevant properties of OspA and OspB, the kinetics of joint retention of both lipidated OspA and lipidated OspB were found to be similar and did not differ from the kinetics of retention of nonlipidated OspA. This finding excludes more rapid elimination of OspB as an explanation for its failure to induce injury. It is also unlikely that differences in NF-B activation are involved, because all lipidated OSPs (OspA, OspB, and OspC) were able to activate this major inflammatory pathway in selected target cell populations, whereas the nonlipidated versions were not effective (6,27) (R. Lohrmann, J. Dunn, S. Batsford. Abstract P304, VIII International Conference on Lyme Borreliosis, Munich, Germany, 1999). Genes coding for lipoproteins are proportionally BATSFORD ET AL more abundant in B burgdorferi compared with other bacterial pathogens (28). The B burgdorferi genome contains ⬎150 genes coding for putative lipoproteins, compared with a known 22 genes in another spirochete, Treponema pallidum (29). In B burgdorferi these genes account for ⬎5% of the chromosomal open reading frames and 14.5–17% of functionally complete open reading frames on plasmids. In contrast, chromosomal lipoprotein genes account for 2.1% and 1.3% of the open reading frames in the spirochete T pallidum and the bacillus Helicobacter pylori, respectively (28). Expression of OSP gene products in the tick vector and the mammalian host has been studied previously (30). In unfed ticks, B burgdorferi usually expresses OspA and more rarely expresses OspC (31–33). Following attachment to the host and feeding, the spirochetes multiply, shed or down-regulate OspA, and up-regulate OspC within the tick (32,34). This pattern of OSP gene expression is probably maintained both during transmission to and in the early phase of infection in the mammalian host (29,32,33). In persisting infections it is thought that a reversal eventually occurs, with OspA being up-regulated and OspC down-regulated. The observation that antibody to OspC is a marker of acute infection in humans (35,36), whereas antibody to OspA appears at later stages (37), is consistent with this presumed pattern of gene expression in humans. A detailed analysis of 137 lipoprotein genes from B burgdorferi in a murine model revealed that 116 genes were transcribed within 10 days of transmission; thereafter, many genes were down-regulated, until finally only 40 of 137 were expressed (29). In contrast to the fluctuations seen with the OspA and OspC genes, the OspB gene appeared to be continuously expressed, which provides an interesting parallel to the relative abilities of these particular gene expression products to induce injury in the model described. B burgdorferi does not produce true LPS, and genomic analysis has shown that LPS biosynthetic genes are absent (28). The recombinant proteins used in the current study contained LPS in quantities comprising 0.1–1.0% of total protein, presumably arising as a contaminant during production and purification. Based on the standard 50-g intraarticular challenge protocol, rats would then have received 50–500 ng of LPS in addition to that in OSP sample. Many of these recombinant samples induced no injury despite the LPS contamination, and lipidated OspA (B31) had the lowest level of LPS (50 ng/50 g OspA), although it induced the most severe injury. Direct challenge with as much as 10 g of LPS also produced no detectable joint injury. INDUCTION OF EXPERIMENTAL JOINT INJURY IN VIVO BY OSPs FROM B BURGDORFERI Taken together, these results indicate that LPS did not contribute to injury in these studies. Local inflammation initiated by OSPs, particularly OspA, is a potential major mechanism in the acute phase of Lyme arthritis. In addition to the experimental evidence presented here and previously (13), development of arthritis in humans during B burgdorferi infection is associated with humoral and cellular immune responses to borrelial antigens, including OspA (37–39). A temporal relationship between the appearance of antibody to OspA and the onset of joint inflammation was previously reported (40). The available data do not support the idea that a vigorous immune response to OSP antigens alone may precipitate joint inflammation. As reported here, immunization of rodents with OspA and OspB did not induce arthritis unless a direct antigen challenge was made subsequently. This is consistent with a further study that showed that hamsters previously immunized with recombinant OspA developed severe arthritis only after challenge with whole infectious B burgdorferi organisms (41). Active immunization of humans with lipidated OspA (42) (see below) or dogs with a bacterin-type vaccine (43) has not yet produced evidence of arthritis. It should be stressed that, in the model described, acute articular inflammation was seen without a specific immune response. This is consistent with the observation that transgenic mice, tolerant to OspA and OspB, still developed arthritis and other organ manifestations during subsequent experimental infection in the absence of specific immunity (44). In Lyme arthritis the organism usually cannot be cultured from joint fluid specimens, yet the strong local inflammatory response suggests that antigen is present. Genetic material specific for B burgdorferi can be detected in samples of synovial fluid from almost all untreated patients. According to one report, OspA sequences were present in 96% of untreated patients (45). Interestingly, Persing et al (46) have shown that there is a molecular imbalance in Lyme arthritis, because plasmid-coded genes such as OspA are much more abundant than genomic genes. The following scenario may help explain the occurrence and perpetuation of Lyme arthritis: after infection, small quantities of spirochetes enter the joint space and shed OSPs, particularly OspA. Molecular imbalance might also help lead to increased local antigen production. These OSPs then attach to joint structures and initiate inflammation, probably via TLRs. Studies on the role of this receptor system in the induction of arthritis following intraarticular application of lipidated OspA in genetically manipulated mice are 2367 currently under way in our laboratory. The subsequent onset of a vigorous immune response (humoral and cellular) may lead to further aggravation and chronic inflammation. It has been suggested that autoimmunity could also develop at this stage (e.g., via molecular mimicry between an immunodominant T cell epitope of OspA and human lymphocyte function–associated antigen 1 [hLFA-1], an adhesion molecule expressed on T cells in the synovium) (47). This notion is controversial, and recently Th1 cell reactivity to hLFA peptides has been attributed to the promiscuous nature of the T cell receptor (T cell degeneracy) (48,49). The current studies are also relevant for vaccine safety considerations, because a vaccine based on lipidated OspA is licensed in North America (50,51), and lipidated OspC is a serious candidate antigen for further vaccine development (42,52). To date, use of the OspA vaccine has not been associated with significant numbers of untoward reactions in joints or other organs (50,51). However, because vaccine efficacy is ⬃70–80%, concern must be raised concerning individuals who are sensitized to OspA but subsequently still become infected. At present, three 30-g intramuscular injections of lipidated OspA are recommended for protection. Using this latter regimen, it is not very likely that lipidated OspA could enter any joint in quantities sufficient to trigger local inflammation, although great caution should be used if the dose of antigen is significantly increased or if DNA vaccines are introduced. Experiments with lipoproteins derived from the bacterium B burgdorferi, the causative agent of Lyme disease, demonstrate that these molecules can be capable of inducing severe tissue injury in vivo. This class of protein should certainly be regarded as a major virulence factor. Particular lipoproteins derived from B burgdorferi, the surface-expressed OSPs A and C, could play an important role in the induction of arthritis, a frequent clinical complication of Lyme disease. ACKNOWLEDGMENTS We thank Jürgen Schulte Mönting (Institute of Medical Biometry and Informatics, Freiburg, Germany) for performing the statistical analysis. Recombinant 83-kd protein from B burgdorferi was provided by Sebastian Rauer (Department of Neurology, University Hospital, Freiburg, Germany). The LPS we used was a kind gift from Marina Freudenberg (Max-Planck Institute of Immunobiology, Freiburg, Germany). The skilled technical assistance of Oliver Schweier and Barbara Lade is acknowledged. 2368 BATSFORD ET AL REFERENCES 1. Vasselon T, Detmers PA. Toll receptors: a central element in innate immune responses. Infect Immun 2002;70:1033–41. 2. Melchers F, Braun V, Galanos C. The lipoprotein of the outer membrane of Escherichia coli: a B-lymphocyte mitogen. J Exp Med 1975;142:473–82. 3. Reitermann A, Metzger J, Wiesmuller KH, Jung G, Bessler WG. Lipopeptide derivatives of bacterial lipoprotein constitute potent immune adjuvants combined with or covalently coupled to antigen or hapten. Biol Chem Hoppe Seyler 1989;370:343–52. 4. Muhlradt PF, Kiess M, Meyer H, Sussmuth R, Jung G. Isolation, structure elucidation, and synthesis of a macrophage stimulatory lipopeptide from Mycoplasma fermentans acting at picomolar concentration. J Exp Med 1997;185:1951–8. 5. Ma Y, Weis JJ. Borrelia burgdorferi outer surface lipoproteins OspA and OspB possess B-cell mitogenic and cytokine-stimulatory properties. Infect Immun 1993;61:3843–53. 6. Wooten RM, Modur VR, McIntyre TM, Weis JJ. Borrelia burgdorferi outer membrane protein A induces nuclear translocation of nuclear factor- B and inflammatory activation in human endothelial cells. J Immunol 1996;157:4584–90. 7. Knigge H, Simon MM, Meuer SC, Kramer MD, Wallich R. The outer surface lipoprotein OspA of Borrelia burgdorferi provides co-stimulatory signals to normal human peripheral CD4⫹ and CD8⫹ T lymphocytes. Eur J Immunol 1996;26:2299–303. 8. Morrison TB, Weis JH, Weis JJ. Borrelia burgdorferi outer surface protein A (OspA) activates and primes human neutrophils. J Immunol 1997;158:4838–45. 9. Morr M, Takeuchi O, Akira S, Simon MM, Muhlradt PF. Differential recognition of structural details of bacterial lipopeptides by toll-like receptors. Eur J Immunol 2002;32:3337–47. 10. Akira S. Mammalian Toll-like receptors. Curr Opin Immunol 2003;15:5–11. 11. Hirschfeld M, Kirschning CJ, Schwandner R, Wesche H, Weis JH, Wooten RM, et al. Cutting edge: inflammatory signaling by Borrelia burgdorferi lipoproteins is mediated by toll-like receptor 2. J Immunol 1999;163:2382–6. 12. Norgard MV, Riley BS, Richardson JA, Radolf JD. Dermal inflammation elicited by synthetic analogs of Treponema pallidum and Borrelia burgdorferi lipoproteins. Infect Immun 1995;63: 1507–15. 13. Gondolf KB, Mihatsch M, Curschellas E, Dunn JJ, Batsford SR. Induction of experimental allergic arthritis with outer surface proteins of Borrelia burgdorferi. Arthritis Rheum 1994;37:1070–7. 14. Brackertz D, Mitchell GF, Mackay IR. Antigen-induced arthritis in mice: I. Induction of arthritis in various strains of mice. Arthritis Rheum 1977;20:841–50. 15. Dunn JJ, Lade BN, Barbour AG. Outer surface protein A (OspA) from the Lyme disease spirochete, Borrelia burgdorferi: high level expression and purification of a soluble recombinant form of OspA. Protein Expr Purif 1990;1:159–68. 16. Rauer S, Kayser M, Neubert U, Rasiah C, Vogt A. Establishment of enzyme-linked immunosorbent assay using purified recombinant 83-kilodalton antigen of Borrelia burgdorferi sensu stricto and Borrelia afzelii for serodiagnosis of Lyme disease. J Clin Microbiol 1995;33:2596–600. 17. Dutton A, Adams M, Singer SJ. Bifunctional imidoesters as cross-linking reagents. Biochem Biophys Res Commun 1966;23: 730–9. 18. Ludwig ML HM. Amidation. In: CHW Hirs, editor. Methods in enzymology. Orlando (FL): Academic Press; 1967. p. 595. 19. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680–5. 20. Gondolf KB, Batsford SR, Lassle G, Curschellas E, Mertz A. Handling of cationic antigens in the joint and induction of chronic allergic arthritis. Virchows Arch B Cell Pathol 1991;60:353–63. 21. McConahey PJ, Dixon FJ. A method of trace iodination of proteins for immunologic studies. Int Arch Allergy Appl Immunol 1966;29:185–9. 22. Van Lent PL, Dekker C, Mosterd J, van den Bersselaar L, van den Berg WB. Allergic arthritis induced by cationic proteins: role of molecular weight. Immunology 1989;67:447–52. 23. Ramesh G, Alvarez AL, Roberts ED, Dennis VA, Lasater BL, Alverez X, et al. Pathogenesis of Lyme neuroborreliosis: Borrelia burgdorferi lipoproteins induce both proliferation and apoptosis in rhesus monkey astrocytes. Eur J Immunol 2003;33:2539–50. 24. Van den Berg WB, van de Putte LB, Zwarts WA, Joosten LA. Electrical charge of the antigen determines intraarticular antigen handling and chronicity of arthritis in mice. J Clin Invest 1984;74: 1850–9. 25. Mertz AK, Batsford SR, Curschellas E, Kist MJ, Gondolf KB. Cationic Yersinia antigen-induced chronic allergic arthritis in rats: a model for reactive arthritis in humans. J Clin Invest 1991;88: 632–42. 26. Gondolf KB, Batsford SR, Vogt A. Isolation of an outer membrane protein complex from Borrelia burgdorferi by n-butanol extraction and high-performance ion-exchange chromatography. J Chromatogr 1990;521:325–34. 27. Ebnet K, Brown KD, Siebenlist UK, Simon MM, Shaw S. Borrelia burgdorferi activates nuclear factor- B and is a potent inducer of chemokine and adhesion molecule gene expression in endothelial cells and fibroblasts. J Immunol 1997;158:3285–92. 28. Porcella SF, Schwan TG. Borrelia burgdorferi and Treponema pallidum: a comparison of functional genomics, environmental adaptations, and pathogenic mechanisms. J Clin Invest 2001;107: 651–6. 29. Liang FT, Nelson FK, Fikrig E. Molecular adaptation of Borrelia burgdorferi in the murine host. J Exp Med 2002;196:275–80. 30. Hodzic E, Feng S, Freet KJ, Borjesson DL, Barthold SW. Borrelia burgdorferi population kinetics and selected gene expression at the host-vector interface. Infect Immun 2002;70:3382–8. 31. Fingerle V, Hauser U, Liegl G, Petko B, Preac-Mursic V, Wilske B. Expression of outer surface proteins A and C of Borrelia burgdorferi in Ixodes ricinus. J Clin Microbiol 1995;33:1867–9. 32. Schwan TG, Piesman J. Temporal changes in outer surface proteins A and C of the lyme disease-associated spirochete, Borrelia burgdorferi, during the chain of infection in ticks and mice. J Clin Microbiol 2000;38:382–8. 33. Leuba-Garcia S, Martinez R, Gern L. Expression of outer surface proteins A and C of Borrelia afzelii in Ixodes ricinus ticks and in the skin of mice. Zentralbl Bakteriol 1998;287:475–84. 34. Schwan TG, Piesman J, Golde WT, Dolan MC, Rosa PA. Induction of an outer surface protein on Borrelia burgdorferi during tick feeding. Proc Natl Acad Sci U S A 1995;92:2909–13. 35. Fung BP, McHugh GL, Leong JM, Steere AC. Humoral immune response to outer surface protein C of Borrelia burgdorferi in Lyme disease: role of the immunoglobulin M response in the serodiagnosis of early infection. Infect Immun 1994;62:3213–21. 36. Wilske B, Fingerle V, Herzer P, Hofmann A, Lehnert G, Peters H, et al. Recombinant immunoblot in the serodiagnosis of Lyme borreliosis: comparison with indirect immunofluorescence and enzyme-linked immunosorbent assay. Med Microbiol Immunol (Berl) 1993;182:255–70. 37. Batsford S, Rust C, Neubert U. Analysis of antibody response to the outer surface protein family in lyme borreliosis patients. J Infect Dis 1998;178:1676–83. 38. Lengl-Janssen B, Strauss AF, Steere AC, Kamradt T. The T helper cell response in Lyme arthritis: differential recognition of Borrelia burgdorferi outer surface protein A in patients with treatmentresistant or treatment-responsive Lyme arthritis. J Exp Med 1994;180:2069–78. 39. Vaz A, Glickstein L, Field JA, McHugh G, Sikand VK, Damle N, et al. Cellular and humoral immune responses to Borrelia burg- INDUCTION OF EXPERIMENTAL JOINT INJURY IN VIVO BY OSPs FROM B BURGDORFERI 40. 41. 42. 43. 44. 45. 46. dorferi antigens in patients with culture-positive early Lyme disease. Infect Immun 2001;69:7437–44. Kalish RA, Leong JM, Steere AC. Association of treatmentresistant chronic Lyme arthritis with HLA-DR4 and antibody reactivity to OspA and OspB of Borrelia burgdorferi. Infect Immun 1993;61:2774–9. Croke CL, Munson EL, Lovrich SD, Christopherson JA, Remington MC, England DM, et al. Occurrence of severe destructive lyme arthritis in hamsters vaccinated with outer surface protein A and challenged with Borrelia burgdorferi. Infect Immun 2000;68: 658–63. Wormser GP. Prospects for a vaccine to prevent Lyme disease in humans. Clin Infect Dis 1995;21:1267–74. Chu HJ, Chavez LG Jr, Blumer BM, Sebring RW, Wasmoen TL, Acree WM. Immunogenicity and efficacy study of a commercial Borrelia burgdorferi bacterin. J Am Vet Med Assoc 1992;201: 403–11. Fikrig E, Tao H, Chen M, Barthold SW, Flavell RA. Lyme borreliosis in transgenic mice tolerant to Borrelia burgdorferi OspA or B. J Clin Invest 1995;96:1706–14. Nocton JJ, Dressler F, Rutledge BJ, Rys PN, Persing DH, Steere AC. Detection of Borrelia burgdorferi DNA by polymerase chain reaction in synovial fluid from patients with Lyme arthritis. N Engl J Med 1994;330:229–34. Persing DH, Rutledge BJ, Rys PN, Podzorski DS, Mitchell PD, 47. 48. 49. 50. 51. 52. 2369 Reed KD, et al. Target imbalance: disparity of Borrelia burgdorferi genetic material in synovial fluid from Lyme arthritis patients. J Infect Dis 1994;169:668–72. Gross DM, Forsthuber T, Tary-Lehmann M, Etling C, Ito K, Nagy ZA, et al. Identification of LFA-1 as a candidate autoantigen in treatment-resistant Lyme arthritis. Science 1998;281:703–6. Benoist C, Mathis D. Autoimmunity provoked by infection: how good is the case for T cell epitope mimicry? Nat Immunol 2001;2:797–801. Gomes-Solecki MJ, Wormser GP, Dattwyler RJ. IFN␥ production in peripheral blood of early Lyme disease patients to hLFA␣L (aa326-345). BMC Musculoskelet Disord 2002;3:25–37. Steere AC, Sikand VK, Meurice F, Parenti DL, Fikrig E, Schoen RT, et al, and the Lyme Disease Vaccine Study Group. Vaccination against Lyme disease with recombinant Borrelia burgdorferi outer-surface lipoprotein A with adjuvant. N Engl J Med 1998; 339:209–15. Sigal LH, Zahradnik JM, Lavin P, Patella SJ, Bryant G, Haselby R, et al, and the Recombinant Outer-Surface Protein A Lyme Disease Vaccine Study Consortium. A vaccine consisting of recombinant Borrelia burgdorferi outer-surface protein A to prevent Lyme disease. N Engl J Med 1998;339:216–22. Bockenstedt LK, Hodzic E, Feng S, Bourrel KW, de Silva A, Montgomery RR, et al. Borrelia burgdorferi strain-specific Osp C-mediated immunity in mice. Infect Immun 1997;65:4661–7.