INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Lyme disease (53), or Lyme borreliosis, is caused by a group of related tick-borne spirochetes classified as Borrelia burgdorferi sensu lato (including B. burgdorferi sensu stricto, B. afzelii, and B. garinii). Although much progress has been made in the characterization of the organism, spirochetal factors responsible for infectivity, immune evasion, and disease pathogenesis remain largely obscure. The most-studied B. burgdorferi membrane protein is outer surface protein A (OspA), a lipoprotein antigen expressed by borreliae in resting ticks and the most abundant protein expressed in vitro by most B. burgdorferi sensu lato isolates (2, 32). Experimental OspA vaccines have demonstrated efficacy in several animal models (11, 19, 24, 33), and OspA vaccines for human use are the subjects of current clinical evaluations (35, 36, 57).

During tick engorgment, OspA expression by borreliae diminishes (15) while expression of other proteins, exemplified by OspC, increases (51). By the time of B. burgdorferi transmission to hosts, spirochetes in the tick salivary glands express little or no OspA. This diminished expression of OspA appears to explain the weak early immune responses to this antigen in experimental or natural infection following tick-borne infection (26, 34, 47, 49) or following inoculation with minimal infectious doses of cultured spirochetes (3, 47). OspA antibodies are sometimes detected in later stages of infection (34, 50), implying that this antigen may be reexpressed by at least some spirochetes during infection. For these reasons OspA-specific antibodies are ineffective in eliminating infection when they are administered after infection is established by syringe inoculation (18, 48) or tick bite (15). Additionally, OspA immunity is circumvented by challenges of in vivo-adapted borreliae in the form of transplants of skin from infected donors into OspA-immunized mice (7). OspA immunization has been shown to mediate killing of spirochetes directly in the midgut of feeding ticks and therefore has a primary mode of action at the vector stage (15, 24). To be efficacious, OspA vaccines must elicit protective levels of antibody that must be maintained throughout periods of tick exposure in order to block spirochete transmission from the vector.

Vaccines against pathogens other than Borrelia are often based on in vivo-expressed antigens that boost anamnestic responses upon infection, potentiate the action of immune effector cells and complement, and inhibit key virulence mechanisms. Many laboratories have examined the vaccine potential of B. burgdorferi antigens that are immunogenic during infection and therefore presumably expressed in vivo. OspC is expressed during infection (37) and elicits protective immunity in rodents (27, 40, 42). However, OspC-immunized mice appear not to be protected against challenge with heterologous B. burgdorferi isolates (41), and in some cases OspC immunization failed to protect mice against challenge with even homologous B. burgdorferi isolates (6, 13). OspE and OspF are immunogenic during infection, but OspF elicits only partial protection against tick-borne or low-dose (10² borreliae) intradermal challenge and OspE is ineffective as a protective immunogen (38). Other immunogenic in vivo-expressed antigens that have been evaluated as targets for protection, including 41-kDa flagellin, P30, P39, P55, P83, IpLA-7, the OspE homolog P21, and the OspF homolog pG, have thus far failed to show efficacy in the prevention of infection (14, 14a, 17, 20, 27, 42, 59, 60). Identification of in vivo-expressed B. burgdorferi antigens that are capable of eliciting broadly protective immune responses has remained elusive.

In mammalian hosts, B. burgdorferi infection is initiated when the tick vector deposits the spirochetes into the dermis during feeding. At both initial and later stages of infection, B. burgdorferi is commonly found in association with collagen fibers in the extracellular matrix (5, 58). Recent findings (29) suggest that colonization of these collagenous tissues may be mediated by spirochetal surface adhesins binding specifically to the collagen-associated proteoglycan decorin (12). These adhesins were expressed at low-to-moderate copy numbers on cultured borreliae (29). Genes for two adhesins of B. burgdorferi recognizing decorin, lipoproteins with apparent masses of 18 to 20 kDa called decorin binding proteins A and B (DbpA and DbpB, respectively), have been partially characterized (28). In the current study, we examined the immunogenicity of DbpA and DbpB during infection, examined their accessibility to antibodies and efficacy as potential vaccine candidates, and evaluated the serological conservation of these proteins.

(This work was presented, in part, at the 14th annual meeting on Modern Approaches to the Control of Infectious Diseases, 9 to 13 September 1996, Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., and in a preliminary report [10a] resulting from that meeting.)

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

B. burgdorferi isolates. Isolates of B. burgdorferi sensu lato were donated by the laboratories of A. Barbour, S. Barthold, R. Johnson, J. Leong, and S. Norris or from the Rocky Mountain Laboratories Microscopy Branch collection. The biologic and geographic origins of these isolates are listed in Table 1. The molecular typing method described by Postic et al. (39) was used to confirm the phylogenetic designation of each of these isolates. Spirochetes were propagated in tightly closed containers at 34°C in modified Barbour-Stoenner-Kelly (BSKII) medium (1) overlaid with a 5% O₂-5% CO₂-90% N₂ gas mixture. Cell densities of these cultures were determined by dark-field microscopy at a magnification of ×400. Batches of BSKII were qualified for spirochete propagation and infection testing by confirming that they supported the growth of 1 to 5 cells of isolate B31. Infectivity of the B. burgdorferi sensu lato isolates for mice was assessed as described below. The following isolates used for immunologic protection studies were determined to have the median infectious dose (ID₅₀) values indicated: B31, 6 × 10¹; Sh-2-82, 6 × 10²; N40, 3 × 10²; 297, 3 × 10³; 25015, 6 × 10¹; HB19, 3 × 10³; and CA-3-87, 3 × 10¹.

Expression and purification of recombinant immunogens. Several recombinant forms of DbpA, schematically represented in Fig. 1, were expressed, purified, and used for immunologic analyses and immunizations. Molecular cloning was accomplished by standard methods (46). A library (a gift of Robin Isaacs) of approximately 2-kb Sau3A I fragments of total genomic DNA from a low-passage culture of B. burgdorferi 297 was constructed in Lambda Zap II (Stratagene Cloning Systems, La Jolla, Calif.), and plaques were plated and blotted onto nitrocellulose membranes by standard methods (46). Phage was recovered from plaques binding digoxigenin-conjugated decorin (29), and B. burgdorferi DNA inserts were excised as clones in the phagemid vector pBluescript II (pBsII) SK() according to the manufacturer's protocol. One of these clones, recombinant plasmid pBG26 (ATCC 69791), harbored a 2.5-kb B. burgdorferi DNA insert containing two 561-bp open reading frames (28). Both of these two genes expressed products with decorin binding activity and were subsequently named dbpA (GenBank accession no. U75866) and dbpB (GenBank accession no. U75867). The nucleotide sequences of dbpA and dbpB shared 50% identity. Single homologs of the dbpA and dbpB genes have been identified recently in the whole genome sequence of B. burgdorferi B31, residing on the linear plasmid lp54, which also contains the ospAB operon (25a). The B31 homologs share 94 and 100% similarity, respectively, with the original dbpA and dbpB genes from B. burgdorferi 297. Recombinant clone pBG29 was constructed from pBG26 by HindIII partial digestion, with deletion of the 5' half of the dbpB gene and retaining full-length dbpA and expression of its product. DbpA₂₉₇ was purified from Escherichia coli JM101/pBG29 by affinity chromatography on decorin-conjugated Sepharose. Decorin purified (12) from fetal bovine skin, a kind gift of L. C. Rosenberg (Montefiore Medical Center, New York, N.Y.), was covalently linked to CNBr-activated Sepharose 4B (Pharmacia, Uppsala, Sweden) according to the manufacturer's instructions. A 1-liter culture of JM101/pBG29 in LB medium (46) plus 100 µg of ampicillin per ml was shaken until an A₆₀₀ value of 0.6 to 0.8 was reached, isopropylthiogalactoside (IPTG) was added to 0.2 mM, and shaking was continued for 2 to 3 h. Cells were suspended in 10 ml of phosphate-buffered saline (PBS) and lysed in a French pressure cell, and the soluble fraction was collected following centrifugation at 200,000 × g. The supernatant fluid was passed through a 0.45-µm-pore-size filter, and 5 ml of the filtrate was passed through a 2-ml decorin-Sepharose column that bound the DbpA. After the column was washed with 20 ml of PBS, DbpA was eluted from the column with 1 M NaCl and recovered at >90% homogeneity as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). DbpA remained soluble after dialysis against PBS. Amino-terminal sequencing showed that this soluble DbpA retained the leader peptide and was the precursor form (pre-DbpA₂₉₇) of the mature lipoprotein.

View larger version (26K): [in this window] [in a new window]   FIG. 1.   Diagrammatic representation of the recombinant Dbp immunogens evaluated for vaccine efficacy. All proteins included the complete DbpA or DbpB coding sequence following Cys at the site of the presumed posttranslational processing. Lpp1:DbpA297 is a fusion of this mature DbpA protein sequence to a modified lpp leader peptide. Fusions of DbpA and DbpB to a (His)6 tag were made within, or in place of, their respective leader peptides. Lpp2:DbpAN40His has both a N-terminal lpp leader and a C-terminal (His)6 fusion.

Antibody reagents. Antisera were raised against recombinant preDbpA₂₉₇ and His-DbpB₂₉₇ in New Zealand White rabbits. Antisera were also generated in rabbits against recombinant B. burgdorferi B31 OspA lipoprotein purified from the E. coli clone containing plasmid pOA1 (16) and against recombinant pneumococcal surface protein A (PspA) purified from the E. coli clone containing plasmid pJY4306 (62). The predicted sequences of the lipoprotein OspA's expressed by plasmids pOA1 and pSO3 are identical, but their products used for mouse and rabbit immunizations, respectively, were purified by different chromatographic procedures. Each animal received a primary subcutaneous (s.c.) immunization of 200 µg of protein emulsified with complete Freund's adjuvant (CFA), followed by two or three booster injections of 100 µg with incomplete Freund's adjuvant (IFA) at 3- to 4-week intervals. Titers of rabbit sera against their homologous recombinant borrelial antigens were 128,000 to 256,000 by enzyme-linked immunosorbent assay (ELISA). Rabbit anti-His-DbpB₂₉₇ reacted weakly with pre-DbpA₂₉₇ (titer of 4,000), but reactivity of rabbit anti-preDbpA₂₉₇ with His-DbpB₂₉₇ was at background levels. These proteins were previously observed to have minimal serologic cross-reactivity (28) consistent with their limited sequence homology (40% amino acid identity and 56% similarity). Monoclonal antibody (MAb) H5332 (2) against OspA, an immunoglobulin G1 (IgG1), was a gift of Alan Barbour. The DbpA-specific MAb 7D2B.3G6, an IgG2b, was obtained from a cloned hybridoma of splenic B cells from preDbpA-immunized BALB/cByJ (BALB; The Jackson Laboratory, Bar Harbor, Maine) mice fused with P3X63Ag8U.1 myeloma cells by standard methods (61).

Scanning immunoelectron microscopy. In vitro-passage-5 B. burgdorferi B31 was collected from 0.5 ml of a mid-log-phase culture by centrifugation for 5 min at 1,000 × g. Sedimented cells were gently resuspended in 0.2 ml of BSKII containing 12 µl of immune or preimmune mouse serum and incubated for 30 min at room temperature with mixing. Spirochetes were then collected by centrifugation and washed twice in BSKII. The bacteria were labeled by incubation in 0.2 ml of BSKII containing 16 µl of goat anti-mouse IgG plus IgM conjugated to 30-nm-diameter gold particles (Ted Pella, Inc., Redding, Calif.). Labeled spirochetes were collected and gently washed twice in Tyrode's buffer and allowed to settle onto 0.1% poly-1-lysine-coated coverslips. After 10 min, the buffer was replaced with fixative containing 2.5% glutaraldehyde in 0.2 M sodium cacodylate (pH 7.2). Following fixation, the coverslips were washed twice in cacodylate and postfixed in 1% osmium tetroxide in cacodylate. After being washed in water and dehydrated in ethanol, the samples were critical-point dried from ethanol through carbon dioxide, lightly sputter coated with chromium, and observed with a Hitachi S4500 field emission scanning electron microscope equipped with an ultralow-voltage backscattered electron detector (GW Electronics, Norcross, Ga.).

Immunization of mice and challenge with B. burgdorferi. For passive immunizations, undiluted rabbit antiserum, or antiserum diluted in PBS, was administered intraperitoneally (i.p.) to 6- to 8-week-old female C3H mice within 1 h prior to challenge (day 0) or at various times afterward. For active immunizations, C3H or BALB mice were injected i.p. at week 0 with 20 µg of DbpA or DbpB, 5 µg of OspA, 20 or 5 µg of soluble or detergent extract of E. coli or PBS emulsified with CFA, given a similar booster immunization in IFA at week 4, and challenged at week 6. B. burgdorferi spirochetes were diluted in BSKII from exponentially growing cultures, and mice were injected s.c. at the base of the tail with 0.1 ml of these dilutions (typically 10⁴ borreliae unless noted otherwise). Isolates used for challenge were passaged fewer than six times in vitro. To assess infection, mice were sacrificed at 14 to 17 days postchallenge, and specimens derived from ears, urinary bladders, and tibiotarsal joints were placed in BSKII-1.4% gelatin-13 µg of amphotericin B per ml-1.5 µg of phosphomycin per ml-15 µg of rifampin per ml, and borrelial outgrowth at 2 or 3 weeks was assessed by dark-field microscopy. Mice were scored as infected when any of the three tissues were culture positive. In some instances, seroconversion for protein P39 reactivity was also used to confirm infections (see below). Median effective dilutions (ED₅₀) of passively administered antisera and ID₅₀ of borreliae were assessed in groups of three C3H mice per dose and were calculated by a standard method (43). Immune serum was also obtained from C3H mice persistently infected with B31.

In vitro growth inhibition assay. A microwell antibody titration assay (45) was used to evaluate the growth inhibition properties of antisera for various isolates of B. burgdorferi sensu lato. Briefly, 10⁵ spirochetes in 100 µl of BSKII were added to serial twofold dilutions of antisera in 100 µl of BSKII in 96-well plates, and the plates were covered and incubated at 34°C in a 5% O₂-5% CO₂-90% N₂ gas mixture for 72 h prior to quantification of borrelia growth by dark-field microscopy. The highest serum dilution showing growth inhibition of >90% reduction in the number of cells and motility was designated as the end point growth inhibition titer. Since, in most cases, antisera diluted less than 1:100 gave only partial growth inhibition, a titer of 100 was defined as positive inhibition.

Immunoblotting and decorin blotting. In a single-well format, samples (2 µg) of Lpp1:DbpA₂₉₇, OspA_B31, or MBP-P39_B31 were boiled in sample buffer containing 1% SDS-2.5% 2-mercaptoethanol before SDS-PAGE through 3% stacking and 12.5% acrylamide resolving gels. Gels were electroblotted onto nitrocellulose membranes, and lanes were probed with 1:50 dilutions of immune sera from infected mice or 1:2,000 dilutions of purified MAbs, followed by the appropriate secondary antibody-enzyme conjugate. Membranes used for DbpA and OspA comparative immunogenicity studies were stained with Ponceau S to confirm that equivalent amounts of purified protein had transferred following electroblotting. Immunoblot signals were detected by fluorography with ECL chemiluminescence reagents (Amersham Corp., Arlington Heights, Ill.).

ELISA. Wells of 96-well microtiter plates (Immulon 2; Dynatech, Chantilly, Va.) were coated with antigen by incubating 50 µl of a 1-µg/ml antigen solution in 0.1 M sodium carbonate buffer at pH 9.6. After unbound antigen was decanted, additional binding sites were blocked by incubating 200 µl of 3% nonfat milk in wash buffer (PBS-0.2% Tween 20 [pH 7.4]). After washing, duplicate serial twofold dilutions of sera in PBS-Tween 20-1% fetal bovine serum were incubated for 1 h and removed, and the wells were washed three times. The wells were incubated with horseradish peroxidase-conjugated goat anti-mouse IgG or goat anti-rabbit IgG, as appropriate. After three washes, bound antibodies were detected with H₂O₂ and 2,2'-azino-di-(3-ethylbenzthiazoline-6-sulfonate) (ABTS; Kirkegaard & Perry, Gaithersburg, Md.), and A₄₀₅ was quantified with a Molecular Devices Corp. (Menlo Park, Ca.) V_max plate reader. IgG levels that were less than or equal to twice the background level in serum samples from naive mice or rabbits were assigned the minimum titer of 1:100 or 1:500, respectively.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

DbpA is surface accessible. Determination of the subcellular location of B. burgdorferi proteins is complicated by the chemical and mechanical fragility of the spirochetal outer membrane (13). We used scanning immunoelectron microscopy to evaluate the location of DbpA and OspA on in vitro-grown unfixed B. burgdorferi sensu stricto isolate B31 and to confirm the integrity of the antibody-labeled cells. Antibodies against both DbpA and OspA bound to the surface of intact borreliae (Fig. 2). Anti-DbpA labeling was uniformly scattered along the spirochete surface (Fig. 2A). Occasionally spirochetes (fewer than 1 in 10) were observed to label heavily with anti-DbpA, suggesting variability in the expression or the accessibility of this protein among cells in this population (Fig. 2B). Anti-OspA labeling was typically more dense and tended to occur in aggregates, possibly reflecting antibody cross-linking of these target proteins (Fig. 2C). The heavier labeling of OspA was consistent with analyses by SDS-PAGE indicating that in vitro OspA levels were 5- to 10-fold-higher than DbpA (29) (data not shown). Periplasmic endoflagella, which are normally contained within the spirochetal outer membrane, did not protrude from these antibody-labeled cells, leading us to tentatively conclude that these membranes were intact and that DbpA and OspA were surface exposed. Endoflagellar filaments were observed protruding from other cells that were partially disrupted (data not shown). These observations do not preclude the possibility of additional quantities of these lipoproteins at a subsurface location, as suggested by others (13).

View larger version (153K): [in this window] [in a new window]   FIG. 2.   Detection of surface-directed antibody labeling by scanning immunoelectron microscopy. B. burgdorferi B31 was incubated in BSKII plus mouse anti-DbpA (A and B), mouse anti-OspA (C), or normal mouse serum (D) prior to labeling of bound antibodies with goat anti-mouse IgG plus IgM-conjugated colloidal gold particles. Typical electron micrographs for these samples are shown except for panel B, which represents spirochetes with atypically heavy anti-DbpA labeling. Bars, 0.2 µm.

Infection elicits strong antibody responses against DbpA and DbpB but not OspA. Dermal inoculation with low numbers of B. burgdorferi organisms near the ID₅₀ failed to elicit antibody responses to OspA but did elicit antibodies to other proteins early in the infection (3, 47). It has been proposed that such inoculations approximate the number of borreliae delivered by tick bites (47). We examined the kinetics and level of the antibody response of C3H mice to DbpA following s.c. inoculation with 10², 10³, or 10⁴ B31 spirochetes. Isolate B31 (uncloned) was chosen because it was one of the most infectious isolates and could be consistently cultured from multiple tissues (including bladder, ear, and joint) of infected mice. Three of 10 mice inoculated with 10² B31 became infected, and all mice were infected at the higher doses. Sera from infected mice within each dose group were pooled and analyzed for immune responses to specific borrelial proteins. Antibodies reactive with DbpA were detectable by immunoblotting within 2 weeks of infection with 10² borreliae, while OspA-specific antibodies remained at background levels within the first 8 weeks after infection (Fig. 3). Sera from culture-negative mice in the 10² dose group did not react with DbpA or OspA (data not shown), indicating that the amount of these proteins in 10² borreliae was subimmunogenic and suggesting that DbpA immune responses were against antigens synthesized in vivo during spirochete multiplication. At inoculum doses of 10³ and 10⁴ spirochetes, early IgG responses to both DbpA and OspA were observed, but anti-DbpA responses were strongest. This was particularly evident at 8 weeks, when anti-OspA reactivity waned and responses to DbpA increased (lanes 9 and 12). Transient anti-DbpA IgM was detectable at all inoculum doses, but anti-OspA IgM responses were much weaker relative to background (lane 3) and were detectable only at inoculum doses of 10³ and 10⁴ spirochetes (lanes 8 and 11).

View larger version (50K): [in this window] [in a new window]   FIG. 3.   Early antibody responses to DbpA and OspA elicited by low-dose challenge. Multilane immunoblot showing antibodies reactive with DbpA and OspA at weeks 2, 4, and 8 postchallenge (lanes 4 to 12) in sera pooled from mice infected by challenge with 102, 103, or 104 B. burgdorferi (B.b.) B31 spirochetes. Lanes: 1, anti-DbpA MAb, 1.0 µg/ml; 2, anti-OspA MAb, 0.2 µg/ml; 3, normal mouse serum. Samples for SDS-PAGE were 2 µg of purified Lpp1:DbpA297 or OspAB31.

View larger version (20K): [in this window] [in a new window]   FIG. 4.   Relative levels of IgG antibodies to DbpA, DbpB, and other selected borrelial antigens during chronic infection of mice. Sera were collected for a 1-year period from mice infected by challenge with 102 B31. Sera were pooled at each time point, and end point titers of IgG specific for each of four recombinant borrelial antigens were determined by ELISA. E. coli MBP, the fusion partner for recombinant P39, was used as a negative control protein.

Passive immunization of mice with DbpA and DbpB antisera. As we initially lacked an isolate of B. burgdorferi 297 that was infectious for mice, isolate B31 was chosen for our first protection studies since it was found to be sensitive to killing in vitro by both preDbpA and OspA antisera (Table 1). Additionally, we sought to determine if antiserum against DbpA would passively protect mice against challenge with a heterologous B. burgdorferi sensu stricto isolate. Rabbit antiserum against pre-DbpA₂₉₇ and antiserum against OspA_B31 were passively administered at the time of challenge with 10⁴ B. burgdorferi B31 spirochetes. Undiluted, both antisera protected mice from infection (Table 3; data not shown). The ELISA IgG titers of the rabbit pre-DbpA₂₉₇ and OspA_B31 antisera against their homologous recombinant antigens were comparable (~256,000). When 0.1 ml of antiserum was given at the time of challenge, the ED₅₀ values of the DbpA and OspA sera were comparable (1:11.2 versus 1:7.5). A second administration of OspA serum on day 2 had no effect, whereas a second administration of DbpA antiserum was much more effective than the single treatment on day 0 (Table 3). It is noteworthy that the higher ED₅₀ for the anti-DbpA₂₉₇ serum, relative to the anti-OspA_B31 serum, was obtained with the B31 isolate, which was heterologous to the DbpA₂₉₇ antigen. Passive immunization of C3H mice with rabbit His-DbpB₂₉₇ antiserum (ELISA IgG titer of 128,000) showed only partial protection against challenge with 10⁴ B. burgdorferi B31 spirochetes (two mice protected of five) and no protection against challenge with 10⁴ B. burgdorferi N40 spirochetes (no mice protected of five). Passive immunization against DbpB was not evaluated further in this study.

Elimination of infection by postinoculation administration of DbpA, but not OspA, antisera. Passive protection results suggested that DbpA may be accessible to antibodies for longer than OspA after challenge and that these relative changes in protein expression or accessibility happen in periods as short as 2 days. The possibility that DbpA is a target for protective antibodies after OspA is no longer accessible was addressed further by administering antisera against both of these antigens to mice at increasing intervals after challenge (Table 4). Cohorts of 30 or 36 C3H mice were inoculated with 10⁴ B. burgdorferi B31 spirochetes, and then groups of three mice each received antisera starting at 0, 2, 4, 5, 6, 7, or 10 days later. In two separate experiments, passive transfer of rabbit pre-DbpA₂₉₇ antiserum completely protected mice from infection with B. burgdorferi B31, even when serum administration was delayed as long as 4 days but not 5 or more days postinoculation. OspA_B31 antisera were protective when administered at the time of challenge, but not at later times, similar to observations of others (48). Our observations showed that DbpA, unlike OspA, was a target for antibody-mediated elimination of borreliae during the early stage of infection and provided further evidence that the relative in vivo levels or accessibility of DbpA and OspA change during this period.

Active immunization of mice with DbpA protects against infection. C3H and BALB mice were actively hyperimmunized with three forms of recombinant DbpA₂₉₇ differing only at their amino termini, one form of DbpA_B31, and one form of DbpA_N40 and then challenged with both homologous and heterologous B. burgdorferi sensu stricto isolates (Table 2). Immunizations with DbpA or OspA preparations elicited high titers of serum IgG (ELISA end point titers were >10⁵). C3H mice immunized with either His-DbpA₂₉₇ or Lpp1:DbpA₂₉₇ were protected from challenge with the homologous 297 isolate as judged by culture and P39 seroconversion (experiment A). C3H mice immunized with Lpp2:DbpA_N40His were protected against N40 homologous challenge (experiment B). However, OspA_B31 did not confer protection against the heterologous N40 challenge (experiment B). BALB mice immunized with pre-DbpA₂₉₇ were protected from challenge with the heterologous B31 isolate (experiment C), but protection of C3H mice from B31 challenge by immunization with the various forms of DbpA was less complete (experiments C, D, and E). Although protection of C3H mice against B31 challenge was somewhat higher in one experiment with His-tagged versions of DbpA (experiment E), more-extensive studies are required to determine whether this form of the immunogen elicits a more potent immune response in mice. Preliminary experiments comparing immunizations with 5 or 20 µg His-DbpA indicated that fewer mice were protected by the lower DbpA dose (2 of 5 protected at 5 µg versus 4 of 5 at 20 µg). Using Freund's adjuvants, we did not observe any substantial difference in immunogenicities between the acylated Lpp1:DbpA₂₉₇ and nonacylated pre-DbpA₂₉₇ and His-DbpA₂₉₇. Such differences may be more apparent with other adjuvants that are clinically relevant, as has been observed for OspA (2, 16, 33).

Passive immunization of mice against diverse B. burgdorferi sensu stricto isolates. We next compared the abilities of rabbit DbpA and OspA antisera to protect C3H mice against challenge with five B. burgdorferi sensu stricto isolates and isolate 25015, which is classified as phylogenetically distinct by some typing methods (10, 39), to assess the conservation of protective epitope(s) on these antigens (Table 5). Antiserum against pre-DbpA₂₉₇ protected C3H mice against challenge with the three genetically heterologous B. burgdorferi sensu stricto isolates B31, Sh-2-82, and N40. Significantly, DbpA_N40 shares only 67% amino acid sequence identity (75% similarity) with DbpA₂₉₇ (44), demonstrating broad reactivity of rabbit DbpA antiserum. Pre-DbpA₂₉₇ antiserum gave broader protection than anti-OspA_B31 serum against these isolates (Table 5). Isolates 25015, HB19, and CA-3-87 were resistant to both passively administered pre-DbpA₂₉₇ and OspA_B31 antisera. With respect to OspA heterogeneity, the lack of cross-protection between OspA of B. burgdorferi sensu lato isolate 25015 and OspA from a B. burgdorferi sensu stricto isolate (N40) has previously been observed by others (21).

Conservation of DbpA epitope(s) binding growth-inhibitory antibodies. We next examined a large panel of B. burgdorferi sensu lato isolates with diverse geographic and biologic origins and representing six of the major phylogenetic groups (10, 39, 53) using a microwell-based growth inhibition assay (45) and by immunoblotting to evaluate the in vitro expression and serologic conservation of DbpA and OspA among these isolates (Table 1). To supplement immunoblotting, we also used a decorin blot assay (29) to evaluate expression of DbpA and DbpB. Sensitivities of these isolates to growth inhibition by rabbit pre-DbpA₂₉₇ or OspA_B31 antisera were assessed in parallel. Growth of 19 of the 35 B. burgdorferi sensu lato isolates (54%) was inhibited (defined as growth inhibitory titer of 100) by pre-DbpA₂₉₇ antiserum. This growth-inhibitory effect did not require addition of serum complement, similar to observations by others using this assay (45). Most isolates of B. burgdorferi sensu stricto (12 of 17 [71%]) and nearly half of the isolates of B. afzelii (2 of 6) and B. garinii (4 of 9) were inhibited by pre-DbpA₂₉₇ antiserum (range of inhibitory titers, 100 to 12,800). In some cases resistance to DbpA antiserum correlated with weak or negative reactions by immunoblotting and decorin blotting (HB19, CA-3-87, and 25015); these isolates apparently expressed little or no DbpA in vitro although PCR analysis indicated that each isolate contained a dbpA gene homolog (44). Some B. garinii isolates were inhibited weakly, or not at all, by pre-DbpA₂₉₇ antiserum but were positive by decorin blotting (PBi, B4 91, G2.22, and IP89). Isolate PBi was inhibited by pre-DbpA₂₉₇ antiserum but was negative by immunoblotting. For a few isolates (FRED and IPF), decorin binding activity appeared to be lost after SDS-PAGE, yet the isolates remained immunoreactive. Possibly differences in DbpA in vitro expression levels, in epitope structure, and in antibody accessibility, all contributed to differences in anti-DbpA inhibitory end point titers among these isolates. About half of the isolates (17 of 35 [49%]) were inhibited by OspA_B31 antiserum (range of inhibitory titers, 100 to 51,000) even though all 35 isolates were positive for OspA expression by immunoblotting. The OspA_B31 antiserum had negligible inhibitory activity against the nine B. garinii isolates. Thus, DbpA epitopes capable of binding growth-inhibitory antibodies appeared to be as conserved as, if not more conserved than, those of OspA. Most of the isolates (26 of 35 [74%]) were inhibited by at least one of the two antisera, suggesting that combinations of these two antigens might provide broader vaccine coverage than either one alone.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Many proteins of B. burgdorferi, including Osps A-F, 41-kDa flagellin, IpLA-7, P21, P30, P35, P37, P39, P55, P83, pG, and lp6.6, have been evaluated as targets for protection in animal models of Lyme borreliosis (14-17, 19-24, 27, 33, 36, 38, 40-42, 48, 54, 59, 60). Among these, only antibodies to OspA, OspB, and OspC have been shown to prevent infection completely. Combined administration of antisera against two in vivo-expressed proteins, P35 and P37, 1 day after homologous challenge also provided protection in a recent study (22) but only against 10², but not 10⁴, spirochetes. The duration of P35/P37 target accessibility beyond day 1, or the level of serologic conservation of these antigens, was not reported. We show here that DbpA is accessible to antibodies capable of aborting infection up to 4 days after inoculation and is serologically cross-reactive among many borrelial isolates. In contrast, antibodies to OspA were ineffective at aborting established infection, as previously reported by others (48).

DbpA and DbpB are relatively minor proteins of cultured B. burgdorferi compared to OspA (28, 29). The antigenic stimulus provided to mice by a minimal infectious dose (10²) of B. burgdorferi B31 was insufficient to elicit immune responses to these proteins in mice that failed to become infected, but persistent and high titers of DbpA and DbpB antibodies, and not OspA antibodies, were observed in culture-positive mice. These observations provide evidence that DbpA and DbpB are expressed in vivo during infection.

DbpA antibodies bound specifically to the surface of intact cultured spirochetes (Fig. 2) and were able directly to inhibit the growth of B. burgdorferi both in vivo and in vitro (Tables 1, 2, and 5). Similar observations were also made for OspA antibodies. However, the experiments evaluating the postchallenge protective efficacies of these antibodies were able to distinguish major differences in the expression level or surface accessibility of OspA, relative to that of DbpA, during the first 4 days after dermal inoculation of spirochetes (Tables 3 and 4). This decrease in expression or accessibility of OspA can be taken as evidence that the spirochetes are undergoing a host adaptation process at this stage of the infection. Methods capable of directly demonstrating the in vivo surface accessibility of B. burgdorferi outer membrane components are not currently available. The observation that B. burgdorferi remains sensitive to the growth-inhibitory effects of DbpA antiserum after a 4-day period of host adaptation provides evidence that DbpA, and not OspA, is surface accessible in vivo during this stage of infection. This is the first report demonstrating evidence of in vivo surface accessibility of a defined target in this manner. In order for decorin adherence (29) to play a role in the virulence of B. burgdorferi, decorin binding proteins would presumably require in vivo surface exposure.

Others have shown that passive transfer of immune sera from persistently infected mice to naive mice will eliminate spirochetes in newly infected mice at 4 days postinfection but will not eliminate infection when immune serum is administered at 12 days postinfection (4). Antigenic targets for this effect have not been identified, but DbpA may be such a target. In the present study, spirochetes remained sensitive to the effects of DbpA antiserum through the first 4 days of infection in mice but became resistant to DbpA antiserum administrations at later times. Several mechanisms for this resistance are possible: (i) DbpA expression becomes down-regulated during or after infection; (ii) DbpA becomes inaccessible on the spirochete to antibodies; or (iii) spirochetes are sequestered in cells or tissues inaccessible to antibodies. A model suggesting that surface-exposed lipoproteins may be globally down-regulated in vivo by spirochetes or differentially localized during infection as part of the host adaptation process has been recently proposed (13). Our observations are consistent with this model but suggest that this may happen in a stage-specific or tissue-specific manner and that changes in lipoproteins occur in a programmed, rather than global, fashion since OspA and DbpA lipoproteins appeared to lose antibody-mediated vulnerability at different times during early infection (Table 4). In C3H mice the dissemination of B. burgdorferi from the site of dermal inoculation by syringe begins about 3 to 4 days postinoculation (8). This coincides with the time at which DbpA antibodies were no longer effective at aborting infection in the present studies. For instance, decorin adherence may contribute to establishing localized infection in collagen-rich connective tissues yet may be dispensable during spirochete dissemination. B. burgdorferi shows tropism for the skin of infected animals (5, 8), and cutaneous manifestations of human Lyme borreliosis are seen in both early and late disease (53). Immune response-mediated disease resolution has been observed in the immunocompetent mouse model of Lyme disease (3, 4) and has been interpreted to explain intermittent remissions during human Lyme disease. Anti-DbpA responses may contribute to these effects and to the control of persistent infection. The sustained anti-DbpA, as well as DbpB and P39, IgG levels in persistently infected mice may be a result of an immune stimulus by spirochetes when they periodically exit a putative immunity-privileged niche and ultimately become eliminated by the host's immune response.

Active immunization of C3H mice with DbpA conferred complete, or nearly complete, protection against challenge with 10⁴ spirochetes of the isolate homologous to the immunogen. Immunization with DbpA conferred complete protection against infection following challenge with a heterologous B. burgdorferi sensu stricto isolate in BALB mice and in some C3H mice and appeared to limit dissemination in the remaining C3H mice. Spirochetes were recovered only from bladders, and not from ears or joints, of the partially protected DbpA-immunized mice, and these spirochetes were highly moribund. Although C3H mice hyperimmunized with OspA_B31 were protected from dermal challenge with 10⁴ spirochetes of the homologous B31 isolate, OspA immunity can be overcome by challenge doses greater than 10⁵ (23) (data not shown). The spirochetes recovered from OspA-immunized mice in the study by Fikrig et al. (23) were shown to be ospA escape mutants and remained viable and infectious, unlike the spirochetes recovered from DbpA-immunized mice. Future experiments will evaluate the disease status of DbpA-immunized mice harboring reduced numbers of spirochetes at 2 weeks postchallenge and will determine whether they eventually eliminate their infections. It is not known at this time whether the mice partially protected from infection also had reduced severity of disease or duration of symptoms.

Passive immunization with rabbit antiserum against DbpA₂₉₇ appeared to elicit more-complete and broader protection of C3H mice than active immunization. This response included protection against isolate N40 expressing DbpA having only 67% amino acid sequence identity (75% similarity) with the DbpA₂₉₇ immunogen (44). Additionally, we found that two passive immunizations of as little as 4 µl of the rabbit pre-DbpA₂₉₇ antiserum completely protected C3H mice from challenge with 10⁴ B. burgdorferi B31 spirochetes (Table 3), but two passive immunizations of the equivalent of 4 µl of BALB or C3H mouse pre-DbpA₂₉₇ antiserum (with even higher DbpA IgG ELISA titers) did not protect naive mice from a similar challenge (data not shown). Differences in the potencies of the various sera may be due to differences in the responses of mice, which demonstrate a persistent B. burgdorferi infection (5, 8), and rabbits, which can clear a B. burgdorferi infection (25), to protective versus nonprotective epitopes or the relative avidities of the antibodies contained in the sera. We do not yet know how broad heterologous protection by DbpA antibodies may be in nonmurine species, but our in vitro results summarized in Table 1 provide encouragement.

By immunoblotting, and by blotting with tagged decorin, isolates HB19, CA-3-87, and 25015 were found to express little or no DbpA in vitro (Table 1), indicating that these borreliae had few DbpA targets at the time of challenge. Interestingly, sera from naive mice infected with HB19 or CA-3-87 reacted with purified Lpp1:DbpA₂₉₇ by immunoblotting (data not shown), suggesting that expression of dbpA by these isolates was up-regulated in vivo after infection began, possibly at the time when the passively administered DbpA₂₉₇ antiserum had diminished to levels below which protection may have otherwise been provided.

The high levels of antibodies to DbpB elicited during persistent infection of C3H mice suggested that, like DbpA, DbpB is also expressed in vivo (Fig. 4). However, the results of protection and in vitro growth inhibition studies suggest that DbpB is less vulnerable to antibodies or is expressed at lower levels than DbpA in vivo, or both. The DbpB sequences of the isolates examined in this study share >98% identity with DbpB₂₉₇ (44), suggesting that serologic variation is not an explanation for their lack of sensitivity to DbpB₂₉₇ antibodies. As we observed that rabbit anti-His-DbpB₂₉₇ was able to protect mice partially against a challenge with 10⁴ borreliae, it appears that accessibility of DbpB to antibodies may differ in vivo and in vitro. In this regard, others have shown that antibodies from OspC-immunized mice that were protected against infection had very weak in vitro growth-inhibitory activity against the same borrelial isolate used for challenge (41, 42). These preliminary experiments suggested that DbpB is a less effective target for protection than DbpA, but additional studies are required to confirm this possibility.

Like the initial reports showing protective immunity targeting OspA (19, 48) and OspC (40), we have utilized cultured B. burgdorferi to demonstrate the protective effect of anti-DbpA immunity. We have also shown that cross-protection among heterologous DbpAs is feasible. The next and most critical step in the evaluation of DbpA vaccines will be to assess whether DbpA immunity will also confer protection against the tick-borne route of infection as was eventually shown for OspA and OspC (24, 27). The protective effects of an anti-DbpA immune response may extend beyond the transmission-blocking effect of an OspA immune response, which is apparently effective only against the spirochete's arthropod stage in the tick midgut. If so, immune responses against DbpA may surpass the efficacy of those elicited by OspA-based vaccines by acting in concert with immune effector functions in vivo. Furthermore, mixtures of DbpA and OspA may be synergistic for efficacy. A multistage, multiantigen strategy is under consideration for vaccines against Plasmodium falciparum malaria, another arthropod-borne pathogen with developmental regulation of antigen expression (31, 56). This strategy may also be effective for protection against both infection and disease caused by B. burgdorferi.