OspE-Related, OspF-Related, and Elp Lipoproteins Are Immunogenic in Baboons Experimentally Infected with Borrelia burgdorferi and in Human Lyme Disease Patients

Abstract

Presently, the rhesus macaque is the only nonhuman primate animal model utilized for the study of Lyme disease. While this animal model closely mimics human disease, rhesus macaques can harbor the herpes B virus, which is often lethal to humans; macaques also do not express the full complement of immunoglobulin G (IgG) subclasses found in humans. Conversely, baboons contain the full complement of IgG subclasses and do not harbor the herpes B virus. For these reasons, baboons have been increasingly utilized as the basis for models of infectious diseases and studies assessing the safety and immunogenicity of new vaccines. Here we analyzed the capability of baboons to become infected with Borrelia burgdorferi, the agent of Lyme disease. Combined culture and PCR analyses of tick- and syringe-infected animals indicated that baboons are a sufficient host for B. burgdorferi. Analysis of the antibody responses in infected baboons over a 48-week period revealed that antibodies are generated early during infection against many borrelial antigens, including the various OspE, OspF, and Elp paralogs that are encoded on the ubiquitous 32-kb circular plasmids (cp32s). By using the baboon sera generated by experimental infection it was determined that a combination of two cp32-encoded lipoproteins, OspE and ElpB1, resulted in highly specific and sensitive detection of B. burgdorferi infection. An expanded analysis, which included 39 different human Lyme disease patients, revealed that a combination of the OspE and ElpB1 lipoproteins could be the basis for a new serodiagnostic assay for Lyme disease. Importantly, this novel serodiagnostic test would be useful independent of prior OspA vaccination status.

Lyme disease, the most common arthropod-borne disease in North America (49, 50, 70), is a multisystem disorder characterized by dermatologic, cardiac, neurologic, and arthritic manifestations (68, 69). Lyme disease pathogenesis and Borrelia burgdorferi infectivity have been studied in numerous animal models; however, the disease manifestations observed vary widely among host species (9). The murine model of Lyme disease has been the most intensively investigated and is presently the preferred animal model for Lyme disease research. The mouse model has allowed researchers to gain valuable insight into the effects of various components of the immune system in relation to Lyme disease pathogenesis (10, 13, 57, 64, 68, 69). However, there are drawbacks to the mouse model of Lyme disease. For instance, the mouse immune system and how it responds to B. burgdorferi infection can differ from the human immune response to this organism. Furthermore, not all disease manifestations observed in humans also are observed in mice, especially the erythema migrans and neurologic symptoms typically associated with Lyme disease (74).

Presently, the rhesus macaque (Macaca mulatta) provides the closest animal model to human Lyme disease. Infection of these animals with B. burgdorferi results in an almost complete spectrum of human disease and the various clinical presentations (16, 51-54, 56, 60). However, like the mouse model of Lyme disease, there are drawbacks to the rhesus macaque model. The major detractors of this nonhuman primate model are the facts that macaques (i) can carry the herpes B virus, which is lethal to humans, and (ii) do not have the opsonizing antibody subclass immunoglobulin G₃ (IgG₃) (18, 23, 61, 65). Additionally, rhesus monkeys have become difficult to obtain due to their limited supply and extensive use as the preferred nonhuman primate model for AIDS research investigations. In contrast, the baboon (Papio hamadryas), another nonhuman primate, is widely available and cost efficient, mainly due to their ability to breed well in captivity. Also, unlike macaques, baboons do not harbor the lethal herpes B virus and contain the full complement of IgG subclasses (61, 65). Baboons have increasingly been utilized for assessing the safety and immunogenicity of candidate human vaccines and for use as an animal model of numerous bacterial, protozoal, and viral infections (e.g., group A and B Streptococcus, group B meningococcus, Haemophilus influenzae type B, Schistosoma mansoni, Trypanosoma cruzi, human immunodeficiency virus, human T-cell leukemia virus, hepatitis B virus, Epstein-Barr virus, and rotavirus) (5-8, 29, 34-38, 41, 46, 48, 55, 66, 75, 77). Therefore, one of the primary goals of this study was to determine if B. burgdorferi could establish and maintain an infection in baboons so that it could be used as an alternative to the present rhesus macaque model of Lyme disease.

Several laboratories, including our own, have recently delineated the ontogeny of the antibody response against three different families of immunogenic circular plasmid-encoded lipoproteins, designated OspE-related, OspF-related, and Elps (4, 32, 39, 47, 71, 73, 76). Given that the various OspE, OspF, and Elp paralogs are ubiquitous among all Lyme disease spirochetes identified, we utilized the production of antibodies against all nine different OspE, OspF, and Elp paralogs to confirm that baboons could become chronically infected with B. burgdorferi. When this serological analysis of the OspE, OspF, and Elp paralogs was expanded to include human serum samples, we observed that a combination of the OspE and ElpB1 lipoproteins could specifically and sensitively diagnose human Lyme disease. The combined data not only provide evidence that baboons may be useful as an alternative nonhuman primate animal model for Lyme disease but also indicate that OspE and ElpB1 may be new candidates for Lyme disease serodiagnosis.

MATERIALS AND METHODS

Bacterial strains.

B. burgdorferi strain B31 was originally isolated from an infected Ixodes scapularis tick collected on Shelter Island, N.Y. (14). The culture utilized in these studies was obtained from the Centers for Disease Control and Prevention and was resurrected from frozen stocks in Barbour-Stoenner-Kelly (BSK)-H medium supplemented with 6% rabbit serum (Sigma Chemical Co., St. Louis, Mo.). Spirochetes were cultivated in vitro for no more than three serial passages before experiments were performed. Electrocompetent Escherichia coli strain DH5α (Gibco/BRL Life Technologies, Gaithersburg, Md.) was used for all transformations; all clones and transformants were grown by using tryptone-yeast extract agar or broth supplemented with the appropriate antibiotic.

Baboon infection.

Adult female baboons were housed in the primate center at the University of Oklahoma Health Sciences Center (OUHSC). Animals were anesthetized with ketamine prior to challenge, and whole-blood samples were obtained. Tick infection and rearing were performed as described previously (79). Two baboons (designated TI-1 and TI-2) each had 15 B. burgdorferi strain B31-CDC-infected ticks placed into capsules that were attached to three sites (5 ticks per site) on the shaved backs of the baboons. One baboon (NC-1) had 15 uninfected ticks placed into capsules similarly to serve as a control. Animals were fitted with tether jackets (to eliminate capsule removal), and ticks were allowed to attach and feed to repletion. Additionally, two baboons (SI-1 and SI-2) were syringe inoculated with 10³ B31-CDC organisms at three sites on their shaved backs, while a third (NC-2) was inoculated with BSK-H medium alone. At 2 weeks postinfection (p.i.) two skin biopsies were obtained from each animal approximately 8 cm from a site of infection or inoculation. One biopsy from each animal was placed in 10% buffered formalin for histological examination. The other biopsy was aseptically halved; one half of the specimen from each animal was placed into 2 ml of BSK-H medium containing rifampin (50 μg/ml) and amphotericin B (25 μg/ml) and was cultivated at 34°C, and the other half was used for DNA isolation and subsequent PCR (described below). Peripheral blood samples were taken prior to infection and at 1, 2, 6, 12, 24, and 48 weeks p.i. All animal protocols were approved by the OUHSC Institutional Animal Care and Use Committee.

Human sera.

A cohort of 39 known human Lyme disease serum samples with clinical summaries and time intervals between onset and blood collection was kindly provided by Martin Schriefer (Centers for Disease Control and Prevention). All patient sera were separated into early-stage (<12 weeks between onset and blood collection) and late-stage (>12 weeks) Lyme disease from the clinical summaries provided. This resulted in 21 different samples being considered early-stage Lyme disease and 18 being considered late-stage disease. Additionally, 29 blood samples from individuals in an area where Lyme disease is endemic (New Jersey) and 34 serum samples from areas where it is not endemic (Georgia, Utah, Arkansas, and Oklahoma) also were included as negative controls (provided by Sean Bauman; Immuno-Mycologics, Inc., Norman, Okla.).

Recombinant proteins.

A glutathione S-transferase fusion construct for B31 OspC was generated by directionally cloning a PCR-amplified product into the BamHI and EcoRI sites of the fusion vector pGEX-4T-2 by using ospC-forward (5′-GGCGGATCCTGTAATAATTCAGGGAAAGATGGG-3′) and ospC-reverse (5′-GCGGAATCCAGGTTTTTTTGGACTTTCTGCCAC-3′) primers. Fusion constructs for OspA, OspE, p21, OspF, BbK2.11, BbK2.10, ElpA1, ElpA2, ElpB1, and ElpB2 were generated as previously described (32).

ELISA.

Enzyme-linked immunosorbent assays (ELISAs) were performed as previously described (32) with minor modifications. B. burgdorferi whole-cell lysates (WCL) were prepared by cultivating organisms to mid-logarithmic phase (approximately 5 × 10⁷ bacteria per ml) and washing them three times in phosphate-buffered saline (PBS; pH 7.4). Pellets were resuspended in PBS (pH 7.4) prior to sonication. Two hundred-fifty nanograms of sonicated B. burgdorferi strain B31-CDC WCL or recombinant protein was diluted in 50 μl of PBS (pH 7.4) prior to plating in triplicate in 96-well Maxisorp Nunc-Immuno plates (Nalge Nunc International, Naperville, Ill.). Samples were allowed to coat overnight at 4°C. For endpoint titers, serum samples were initially diluted 1:50 and were serially diluted twofold to a final dilution of 1:51,200. For both human and baboon samples, secondary antibodies were horseradish peroxidase-conjugated sheep anti-human Ig, isotypes G (gamma chain)- or M (mu chain)-specific (The Binding Site, San Diego, Calif.), which were previously demonstrated to react with baboon antibodies (65). Antibody endpoint titers were determined as the highest dilution that the mean optical density at 405 nm (OD₄₀₅) was three standard deviations (SD) above the mean OD₄₀₅ of the same dilution of serum added to wells coated with blocking reagent alone. To determine specific reactivity, the average OD₄₀₅ of serum samples added to wells coated with blocking reagent alone was subtracted from the average OD₄₀₅ of the same dilution of serum samples added to wells coated with the antigen of interest. To determine the cutoff values for the identification of positive or negative human serum samples, the specific reactivity was determined for serum from healthy patients to each of the OspE, OspF, and Elp paralogs. For each protein, the specific reactivity from these patients was averaged and the SD was determined. Cutoff values for each of the proteins were determined as the average of the control patient serum samples plus three SD. All human serum samples were diluted 1:100 for ELISAs. A two-tailed unpaired t test was utilized to determine significant differences between OD₄₀₅ values.

Histological examination of tissues.

Tissue biopsies were fixed in 10% buffered formalin and were embedded in paraffin. Sections were cut and stained with hematoxylin and eosin for histological examination.

DNA extraction and amplification.

DNA was isolated from baboon tissue biopsies and whole-blood samples by utilizing Qiagen QiaAmp tissue and blood kits, respectively, following the manufacture's instructions (Valencia, Calif.). Primary PCRs were performed by utilizing primers designed for ospA, flaB, and β-actin (ospA-5′ ATGAAAAAATATTTATTGGGAATAG, ospA-3′ TTTTAAAGCGTTTTTAATTTCATCA, flaB-5′ ATGATTATCAATCATAATACATCAGC, flaB-3′ TCTAAGCAATGACAAAACATATTGGGG, actB-5′ ATTGCCGACAGGATGCAGAA, actB-3′ GCTGATCCACATCTGCTGGAA). Two microliters and 10 μl of each DNA sample were used in separate PCRs. Nested PCR was performed on all ospA and flaB reaction mixtures by using 2.5 μl of the primary amplification reaction mixture as template. The primers utilized for nested PCR were ospA-5′ nest TGCTGAAAAAACAACATTGGTGG, ospA-3′ nest TTGGTGCCATTTGAGTCGTATTG, flaB-5′ nest AGAGCTTGGAATGCAGCCT, and flaB-3′ nest GGGAACTTGATTAGCCTGCG. Products were separated on agarose gels and were visualized by ethidium bromide staining. Positive PCR amplicons were cloned into the pCR 2.1 TOPO vector (Invitrogen, San Diego, Calif.) for nucleotide sequence verification. Nucleotide sequencing was performed at the OUHSC core facility by using an ABI model 373A automated DNA sequencer and PRISM ready reaction DyeDeoxy terminator cycle sequencing kit according to the manufacturer's instructions (Applied Biosystems Inc., Foster, Calif.).

Computer analysis.

All nucleotide sequence analysis and primer selection was performed by using the MacVector version 6.5.3 software package (Oxford Molecular Group, Cambell, Calif.). ELISA reactions were read by using SOFTmax Pro version 2.2.1 (Molecular Devices Co., Sunnyvale, Calif.) software, and subsequent data were transferred into Microsoft Excel 98 (Microsoft Co., Redman, Wash.) for statistical analyses.

RESULTS

Infection of baboons with B. burgdorferi strain B31.

To determine if baboons are susceptible to B. burgdorferi infection, two were infested with 15 infected ticks each (designated TI-1 and TI-2) and two were infected by syringe inoculation (designated SI-1 and SI-2). Additionally, two baboons were included as controls, one infected with naïve ticks (NC-1) and a second injected with BSK medium alone (NC-2). The first indication that baboons were infected with B. burgdorferi occurred during an examination of the sites of infection or inoculation 14 days p.i. During this inspection, a circular erythematous lesion approximately 6 cm in diameter was observed on baboon TI-2 (Fig. 1). No other lesions or localized sites of inflammation were noticeable. At the 14-day p.i. time point, two biopsies, approximately 8 cm from the sites of infection, were obtained from each animal to determine if spirochetes were present. As shown in Table 1, PCR amplification using ospA and flaB primers resulted in the identification of amplicons of the expected size from the TI-1, TI-2, SI-1, and SI-2 biopsies. Additionally, all biopsy samples were culture positive for spirochetes, a further confirmation that spirochetes were replicating and had disseminated from the original inoculation sites. As expected, no ospA and flaB amplicons or spirochetes were detected for the uninfected baboons (NC-1 or NC-2). To determine if organisms hematogenously disseminated in the baboons after infection, DNA was isolated from whole-blood samples taken at 0, 1, 2, 6, 12, 24, and 48 weeks p.i. for PCR analysis. DNA isolated from the 2-week time points were positive for ospA and/or flaB in baboons TI-2 and SI-1 (Table 1). Finally, histopathology on biopsy samples revealed that the skin biopsy from baboon TI-2 had a significant lymphoid cellular infiltrate showing moderate, multifocal, interstitial accumulation of lymphoid cells with most of the infiltrate being perivascular and involving the vessels of the dermis and superficial subcutis. Biopsy samples from TI-1, SI-1, and SI-2 also showed mild perivascular lymphocytic infiltrates; the NC-1 and NC-2 samples had no evidence of an inflammatory infiltrate, as expected (data not shown).

FIG. 1.

Identification of an erythematous lesion on baboon TI-2 at the site of feeding by a tick infected with B. burgdorferi strain B31-CDC. (A) The lesion observed on a tick-infected baboon (TI-2) where B. burgdorferi B31-CDC-infected ticks were placed. (B) The absence of lesion or inflammation on baboon NC-1 where uninfected ticks were placed. The bars represent 2 cm, and arrows indicate the site of tick attachment.

TABLE 1.

Detection of B. burgdorferi in baboon whole blood or skin biopsy

Baboon	PCR analysis of whole blood (ospA/flaB)a at weeks p.i.:							Biopsy analysis from:
	0	1	2	6	12	24	48	Cultureb	PCRc
TI-1	−/−	−/−	−/−	−/−	−/−	−/−	−/−	+	+
TI-2	−/−	−/−	+/+	−/−	−/−	−/−	−/−	+	+
SI-1	−/−	−/−	+/−	−/−	−/−	−/−	−/−	+	+
SI-2	−/−	−/−	−/−	−/−	−/−	−/−	−/−	+	+
NC-1	−/−	−/−	−/−	−/−	−/−	−/−	−/−	−	−
NC-2	−/−	−/−	−/−	−/−	−/−	−/−	−/−	−	−

^aNested PCR was performed on DNA isolated from whole-blood samples using primers specific for ospA, flaB, and β-actin as listed in Materials and Methods. All samples were positive for β-actin. A plus or minus indicates ospA or flaB was or was not detected, respectively.

^bBiopsies were cultivated in BSK-H medium containing rifampin and amphotericin B. Cultures were examined and considered positive if spirochetes were observed at any point by dark-field microscopy.

^cPCR was performed on DNA isolated from one half of a skin biopsy by using primers specific for ospA, flaB, and β-actin as listed in Materials and Methods. All samples were positive for β-actin. Specific amplicons were detected (+) or were not detected (−) for both ospA and flaB.

Antibody response following tick or syringe inoculation of baboons.

We next examined the ontogeny of the baboon humoral immune response against B. burgdorferi after infection to help determine if animals can become chronically infected. The antibody response against WCL of B. burgdorferi and two well-characterized outer-surface proteins, OspA and OspC, were used for these experiments. As shown in Table 2, ELISA analysis revealed that IgM endpoint titers against WCL ranged from 1:200 to 1:400 during the first 2 weeks of analysis; however, IgM antibody titers increased (range of 1:800 to 1:6,400) in all infected baboons by 6 weeks p.i., suggesting the infection was progressing. Consistent with an ongoing and active infection, all baboons exhibited IgG titers to WCL in a range of 1:400 to 1:1,600 during the initial 2 weeks p.i. which then dramatically increased (1:6,400 to 1:51,200) and stayed elevated throughout the rest of the study period.

TABLE 2.

Reciprocal antibody endpoint titers of baboon sera to B. burgdorferi WCL

Weeks p.i.	Endpoint titer (IgM/IgG) for animal:
	Tick infected		Syringe infected		Negative controls
	TI-1	TI-2	SI-1	SI-2	NC-1	NC-2
1	200/1,600	400/800	200/400	400/1,600	400/1,600	400/800
2	400/1,600	400/1,600	400/800	400/1,600	200/800	400/1,600
6	1,600/>51,200	6,400/>51,200	800/25,600	800/6,400	200/800	200/800
12	800/>51,200	400/>51,200	800/25,600	800/12,800	200/400	400/1,600
24	400/>51,200	400/>51,200	400/12,800	400/12,800	400/800	200/1,600
48	400/51,200	400/>51,200	800/25,600	400/25,600	400/800	400/1,600

It has been shown that OspC is dramatically upregulated during the early stages of mammalian infection while OspA is downregulated (62, 63). Consistent with this reciprocal pattern of protein expression during the early stages of mammalian infection, IgM and IgG titers to OspC in baboons peaked at 6 weeks p.i. and then quickly decreased (Table 3), while antibodies to OspA were never found to be greater than 1:200 in tick-infected animals or 1:1,600 in syringe-infected animals (Table 4).

TABLE 3.

Reciprocal antibody endpoint titers of baboon sera to OspC

Weeks p.i.	Endpoint titer (IgM/IgG) for animal:
	Tick infected		Syringe infected		Negative controls
	TI-1	TI-2	SI-1	SI-2	NC-1	NC-2
1	<50/<50	<50/<50	<50/<50	<50/<50	<50/<50	<50/<50
2	100/100	<50/<50	100/100	<50/<50	<50/<50	<50/<50
6	1,600/25,600	3,200/12,800	800/800	400/3,200	<50/<50	<50/<50
12	200/6,400	400/6,400	200/6,400	200/3,200	<50/<50	<50/<50
24	50/3,200	100/3,200	100/3,200	50/1,600	<50/<50	<50/<50
48	50/1,600	50/1,600	100/800	50/800	<50/<50	<50/<50

TABLE 4.

Reciprocal antibody endpoint titers of baboon sera to OspA

Weeks p.i.	Endpoint titer (IgM/IgG) for animal:
	Tick infected		Syringe infected		Negative controls
	TI-1	TI-2	SI-1	SI-2	NC-1	NC-2
1	<50/<50	<50/<50	<50/<50	<50/<50	<50/<50	<50/<50
2	<50/<50	<50/<50	100/<50	200/<50	<50/<50	<50/<50
6	<50/<50	100/200	200/1,600	400/1,600	<50/<50	<50/<50
12	<50/<50	<50/<50	200/400	200/200	<50/<50	<50/<50
24	<50/<50	<50/100	100/200	<50/<50	<50/<50	<50/<50
48	100/<50	100/200	100/100	200/100	<50/<50	<50/<50

Antibody reactivity to the OspE-related, OspF-related, and Elp proteins.

Previous studies with mice have clearly documented that antibodies are generated to many of the OspE-related, OspF-related, and Elp lipoprotein paralogs within 6 to 8 weeks after tick or syringe infection of B. burgdorferi (4, 32, 39, 47, 71, 73, 76). To compare the ontogeny of the antibody response against these immunogenic lipoproteins in baboons with the prior mouse data, reactivity to all of the OspE, OspF, and Elp paralogs from B. burgdorferi strain 297 was assessed. Consistent with the fact that there are few OspF-related proteins in the B31 strain used to infect the baboons, little reactivity was detected against the recombinant OspF-related proteins (OspF, BbK2.11, and BbK2.10) used for the ELISA analysis (Fig. 2). In contrast to the low overall reactivity observed with the OspF-related lipoproteins, high titers were observed with the OspE-related proteins (OspE and p21); antibody titers to these proteins were especially high in tick-infected animals (Fig. 2A and B). In TI-1 and TI-2, titers of 1:12,800 to 1:51,200 were detected during the 6 to 48 weeks of infection for both OspE and p21, with the response against OspE being more robust. Although the syringe-infected baboons had a lower response to the OspE-related proteins compared to that of the tick-infected animals, SI-1 and SI-2 had titers ranging from 1:6,400 to 1:12,800 to both OspE and p21 throughout the study period (Fig. 2C and D). Although the OspE and p21 antigens were highly reactive, the highest observed antibody titer in all baboons was observed against ElpB1. Tick- and syringe-infected animals generated high titers of ElpB1 antibodies during the 6- to 24-week p.i. period (1:51,200 or greater for tick-infected animals and 1:25,600 for syringe-infected animals; Fig. 2A through D).

FIG. 2.

Antibody response of baboons to OspE-related, OspF-related, and Elp proteins. ELISAs to determine combined IgM and IgG antibody titers were performed with sera from TI-1 (A), TI-2 (B), SI-1 (C), and SI-2 (D) baboons from 6 to 48 weeks p.i. Reciprocal antibody endpoint titers were determined as described in Materials and Methods. Asterisks indicate that the titer was greater than 1:51,200.

Comparison of baboon reactivity to B. burgdorferi WCL and a combination of OspE and ElpB1.

Given the high titers of antibody observed to both OspE and ElpB1 in the infected baboons, we next examined whether these antigens could be used to serodiagnose infected animals. As shown in Fig. 3, when recombinant forms of OspE and ElpB1 were used as substrate for ELISA analyses, extremely low background reactivity was identified in both negative control animals (panels A and B) and in prebleed samples from infected baboons (panels C to F); in all cases OspE and ElpB1 background reactivity in infected baboons was lower than that observed against B. burgdorferi WCL (compare leftmost columns in panels C to F). Six weeks p.i., antibody reactivity to the combination of OspE and ElpB1 was significantly higher in all four infected baboons compared to that for WCL (mean OD₄₀₅ of 1.447 versus 0.631; P < 0.001). Despite the fact that the specific antibody reactivity to the combination of OspE and ElpB1 decreased during the next 24 to 48 weeks of infection in all four baboons, the mean reactivity (OD₄₀₅) of sera against OspE and ElpB1 and WCL was not significantly different, strongly suggesting that the multivalent OspE and ElpB1 substrate was equal to or better than WCL for determining infection with B. burgdorferi.

FIG. 3.

Comparison of baboon reactivity to B. burgdorferi WCL and a combination of the OspE and ElpB1 proteins. Combined IgM and IgG ELISA OD₄₀₅ values were measured by using sera from NC-1 (A), NC-2 (B), TI-1 (C), TI-2 (D), SI-1 (E), and SI-2 (F) baboons during the 48 weeks of analysis to compare the seroreactivity to OspE and ElpB1 and WCL as indicated in Materials and Methods. Values displayed are mean OD₄₀₅ values. Asterisks indicate that the mean OD₄₀₅ for OspE and ElpB1 was significantly higher than the mean OD₄₀₅ for WCL (P < 0.001).

Reactivity of serum from early and late Lyme disease patients to the OspE-related, OspF-related, and Elp lipoproteins.

To expand the above analyses, we next analyzed reactivity to the OspE, OspF, and Elp paralogs with 39 serum samples from known human Lyme disease patients. The reactivity of the various samples also were determined for OspC and FlaB, two proteins that are presently included among the proteins used to confirm Lyme disease by serodiagnosis (21, 22, 25). To begin this analysis, 29 serum samples from healthy patients from an area where Lyme disease is endemic and 34 samples from individuals from areas where Lyme disease is not endemic were utilized to determine the lower limit of OD₄₀₅ values for each antigen to consider it positive for reactivity to each of the OspE-related, OspF-related, and Elp proteins. The combined ELISA analyses revealed that no single antigen resulted in 100% identification of early or late Lyme disease patients, including OspC and FlaB. As shown in Table 5, consistent with the low reactivity observed for baboons against the OspF homologs (OspF, BbK2.10, and BbK2.11), OspF-related proteins appear to be poor indicators of B. burgdorferi infection in humans as well. However, similar to the high reactivity observed in the infected baboons for OspE and ElpB1 (see Fig. 2), the majority of the early-stage human samples tested were positive for OspE (71%) or ElpB1 (86%) reactivity. The percentage of samples positive for OspE or ElpB1 was similar to the percentage of early Lyme disease patients reactive with OspC or FlaB (76%). However, when the OspE and ElpB1 data were combined, the antigens together were able to identify 19 out of 21 (90%) patients as positive for early Lyme disease; identical numbers were also found when the OspC and FlaB data were combined (19 out of 21; 90%). As shown in Table 6, the overall percentages of individuals positive for the various OspE, OspF, and Elp paralogs during the later stages of infection were similar to those observed during early infection (Table 5), with a combined 89% (16 out of 18) being positive for the OspE and/or ElpB1 antigens.

TABLE 5.

Antibody reactivity of early (<12 weeks p.i.) Lyme disease patients to the OspE-related, OspF-related, and Elp lipoproteins

Patient no. (n = 21)	Antibody reactivitya
	OspE	p21	OspF	BbK2.10	BbK2.11	ElpA1	ElpA2	ElpB1	ElpB2	OspC	FlaB
902622	+ (0.176)	− (0.141)	+ (0.624)	− (0.086)	− (0.090)	− (0.117)	+ (0.664)	+ (0.436)	+ (0.134)	+ (1.845)	+ (0.865)
902668	+ (0.246)	− (0.186)	+ (0.186)	− (0.231)	− (0.178)	+ (0.155)	− (0.210)	+ (0.597)	+ (0.188)	+ (0.422)	+ (0.419)
910865	+ (0.316)	+ (0.971)	+ (0.249)	− (0.114)	+ (0.947)	− (0.129)	− (0.119)	+ (1.001)	+ (0.406)	− (0.221)	+ (0.543)
911104	− (0.123)	− (0.218)	− (0.113)	− (0.143)	+ (0.541)	− (0.087)	− (0.102)	+ (0.148)	− (0.117)	+ (1.417)	+ (0.881)
911222	− (0.103)	− (0.151)	− (0.095)	− (0.350)	+ (1.582)	+ (0.732)	+ (0.574)	+ (0.154)	+ (0.202)	+ (1.922)	+ (0.915)
911347	+ (0.238)	+ (0.510)	− (0.120)	− (0.347)	− (0.099)	+ (0.167)	+ (0.236)	+ (0.218)	+ (0.162)	+ (1.011)	+ (0.575)
911348	+ (1.146)	+ (0.512)	− (0.118)	− (0.157)	− (0.096)	− (0.151)	+ (0.396)	+ (1.201)	+ (0.131)	+ (1.629)	+ (0.714)
911349	+ (0.478)	+ (0.919)	+ (0.227)	− (0.165)	− (0.161)	+ (0.178)	+ (0.286)	+ (0.186)	+ (0.166)	+ (1.119)	+ (0.648)
911350	+ (0.179)	+ (0.339)	− (0.145)	− (0.227)	− (0.132)	+ (0.191)	+ (0.275)	+ (0.218)	+ (0.203)	+ (0.787)	+ (0.446)
911351	+ (0.379)	+ (0.264)	+ (0.279)	− (0.309)	− (0.123)	+ (0.195)	+ (0.294)	+ (0.222)	+ (0.284)	+ (0.700)	+ (0.779)
911352	− (0.111)	− (0.212)	− (0.097)	− (0.099)	− (0.074)	− (0.095)	− (0.110)	+ (0.162)	− (0.117)	− (0.220)	− (0.294)
911353	+ (0.252)	+ (0.439)	+ (0.270)	− (0.275)	+ (0.283)	+ (0.266)	+ (0.361)	+ (0.214)	+ (0.183)	+ (0.263)	+ (0.389)
911354	+ (0.252)	− (0.169)	+ (0.259)	− (0.104)	− (0.110)	+ (0.176)	− (0.216)	+ (0.192)	+ (0.141)	− (0.176)	+ (0.390)
911458	+ (0.240)	− (0.146)	− (0.128)	− (0.094)	− (0.097)	− (0.124)	− (0.169)	+ (0.180)	+ (0.173)	− (0.217)	− (0.213)
911846	+ (0.286)	− (0.191)	+ (0.181)	− (0.141)	− (0.130)	− (0.144)	− (0.187)	+ (0.171)	+ (0.141)	− (0.168)	− (0.307)
911847	+ (0.249)	+ (0.271)	− (0.154)	− (0.230)	− (0.106)	− (0.135)	+ (0.289)	+ (0.199)	− (0.125)	+ (0.470)	+ (0.569)
921941	− (0.135)	− (0.227)	+ (0.221)	− (0.104)	− (0.089)	− (0.146)	− (0.131)	+ (0.153)	− (0.104)	+ (0.275)	− (0.244)
931414	− (0.126)	− (0.091)	− (0.084)	+ (1.212)	+ (0.402)	+ (0.188)	− (0.181)	− (0.116)	− (0.121)	+ (1.294)	+ (1.429)
931426	− (0.154)	− (0.110)	− (0.082)	− (0.128)	− (0.081)	− (0.131)	+ (0.450)	− (0.100)	+ (0.150)	+ (0.312)	− (0.309)
902111	+ (1.471)	+ (0.288)	+ (1.043)	+ (0.420)	+ (0.961)	+ (0.288)	+ (0.326)	+ (0.792)	+ (0.148)	+ (0.421)	+ (1.325)
940880	+ (0.207)	− (0.141)	− (0.117)	− (0.112)	− (0.093)	+ (0.157)	− (0.171)	− (0.137)	+ (0.155)	+ (1.762)	+ (0.834)

^aSera with an average OD greater than or less than the cutoff OD was scored as (+) or (−), respectively. Average OD₄₀₅ for each sample is listed in parentheses. Cutoff for each protein was determined as the average OD of serum from 63 healthy donors plus 3 SD. The cutoff OD values (and percent positive) were the following: OspE, 0.174 (71%); p21, 0.263 (43%); OspF, 0.158 (48%); BbK2.10, 0.391 (10%); BbK2.11, 0.191 (29%); ElpA1, 0.153 (52%); ElpA2, 0.232 (52%); ElpB1, 0.139 (86%); ElpB2, 0.128 (76%); OspC, 0.245 (76%); FlaB, 0.309 (76%).

TABLE 6.

Antibody reactivity of late (>12 weeks p.i.) Lyme disease patients to the OspE-related, OspF-related, and Elp lipoproteins

Patient no. (n = 18)	Antibody reactivitya
	OspE	p21	OspF	BbK2.10	BbK2.11	ElpA1	ElpA2	ElpB1	ElpB2	OspC	FlaB
910521	− (0.154)	− (0.098)	− (0.094)	− (0.108)	− (0.071)	− (0.103)	− (0.130)	− (0.096)	+ (0.145)	− (0.103)	+ (0.355)
910531	+ (1.820)	+ (0.809)	+ (0.871)	+ (1.517)	+ (1.300)	+ (0.270)	+ (0.503)	+ (1.700)	+ (0.887)	+ (0.293)	+ (1.192)
910532	+ (0.848)	+ (0.363)	+ (0.709)	− (0.347)	+ (0.476)	+ (0.284)	− (0.193)	+ (0.617)	+ (0.178)	+ (0.315)	− (0.299)
910533	+ (0.240)	+ (0.493)	− (0.133)	− (0.141)	− (0.164)	− (0.126)	− (0.132)	+ (0.273)	+ (0.136)	− (0.168)	+ (0.329)
910544	+ (0.309)	+ (0.264)	+ (0.175)	− (0.229)	+ (0.383)	+ (0.201)	+ (0.650)	+ (0.621)	+ (0.328)	+ (0.356)	+ (1.653)
910794	+ (0.191)	− (0.129)	+ (0.173)	− (0.210)	− (0.089)	− (0.149)	− (0.095)	− (0.111)	− (0.093)	− (0.202)	+ (0.937)
910900	+ (0.922)	− (0.218)	+ (0.469)	− (0.180)	+ (0.578)	+ (0.158)	+ (0.315)	+ (0.479)	− (0.104)	+ (1.502)	+ (0.914)
910943	− (0.143)	− (0.096)	− (0.087)	+ (0.537)	− (0.121)	− (0.116)	− (0.104)	+ (0.168)	− (0.104)	− (0.111)	+ (0.411)
911841	+ (0.300)	+ (0.271)	− (0.155)	− (0.128)	− (0.115)	+ (0.161)	+ (0.242)	+ (0.166)	+ (0.184)	+ (0.378)	+ (0.366)
911842	+ (1.223)	+ (0.296)	− (0.127)	− (0.104)	− (0.112)	− (0.141)	− (0.183)	+ (0.141)	− (0.128)	− (0.185)	+ (0.351)
911843	+ (0.186)	+ (0.321)	− (0.127)	− (0.176)	− (0.124)	− (0.151)	− (0.136)	+ (0.140)	+ (0.170)	+ (0.272)	− (0.230)
911844	+ (0.243)	− (0.176)	+ (0.161)	− (0.126)	+ (0.217)	+ (0.266)	+ (0.699)	+ (0.551)	+ (0.224)	+ (1.545)	+ (0.390)
911845	+ (0.427)	+ (0.332)	+ (0.311)	− (0.177)	− (0.143)	+ (0.237)	+ (0.233)	+ (0.218)	+ (0.164)	− (0.182)	+ (0.464)
920057	+ (0.975)	− (0.186)	− (0.126)	− (0.136)	− (0.108)	− (0.129)	+ (0.236)	+ (0.720)	− (0.125)	− (0.130)	+ (0.588)
921682	+ (0.300)	− (0.186)	− (0.135)	− (0.116)	− (0.100)	+ (0.165)	− (0.186)	+ (0.151)	+ (0.152)	− (0.142)	− (0.291)
921982	− (0.147)	− (0.161)	− (0.135)	− (0.078)	− (0.101)	− (0.118)	− (0.144)	− (0.110)	− (0.103)	+ (0.535)	− (0.269)
930206	+ (0.250)	− (0.207)	− (0.089)	− (0.075)	− (0.072)	− (0.135)	− (0.104)	− (0.111)	− (0.115)	+ (0.252)	+ (0.748)
930208	− (0.147)	− (0.180)	− (0.111)	− (0.228)	− (0.078)	+ (0.302)	+ (0.445)	+ (0.159)	+ (0.166)	− (0.135)	+ (0.351)

^aSera with an average OD greater than or less than the cutoff OD was scored as (+) or (−), respectively. Average OD₄₀₅ for each sample is listed in parentheses. Cutoff for each protein was determined as the average OD of serum from 63 healthy donors plus 3 SD. The cutoff OD values (and percent positive) were as follows: OspE, 0.174 (78%); p21, 0.263 (44%); OspF, 0.158 (39%); BbK2.10, 0.391 (11%); BbK2.11, 0.191 (28%); ElpA1, 0.153 (50%); ElpA2, 0.232 (44%); ElpB1, 0.139 (78%); ElpB2, 0.128 (61%); OspC, 0.245 (50%); FlaB, 0.309 (78%).

DISCUSSION

While numerous animal species have been demonstrated to be susceptible to B. burgdorferi infection, the only nonhuman primate previously used as a model for Lyme disease is the rhesus macaque (9, 24, 53, 56, 57, 60). The data presented here demonstrate that baboons can become infected with B. burgdorferi and, furthermore, it appears that this infection can persist within these animals for at least 48 weeks. This is supported by the findings that spirochetes were detected in peripheral blood and skin biopsy samples 2 weeks after infection and that robust antibody titers to B. burgdorferi were generated during infection and increased (or were maintained) for at least 48 weeks p.i. Although spirochetes were not detected in peripheral blood samples of baboons after 2 weeks of infection, this is not surprising given the extreme difficulty in detecting and isolating spirochetes from human Lyme disease patients during active infection (11, 12). The combined data support our contention that baboons can become persistently infected with B. burgdorferi. However, we recognize it will be necessary in future studies to include a full necropsy of an infected baboon(s) in order determine the pathological relevance of this animal model to human Lyme disease.

Analysis of the humoral immune response against B. burgdorferi antigens in mice has demonstrated that within 6 to 8 weeks of infection antibodies are generated for many of the OspE, OspF, and Elp paralogs (4, 32, 39, 47, 71, 73, 76). Although the extent of the antibody response can vary a small amount between individual mice, the development of antibodies directed against the OspE, OspF, and Elp paralogs are consistently seen within 12 weeks p.i., independent of whether mice are tick or syringe inoculated (32). In contrast, the patterns and degree of antibody reactivity to each of the OspE, OspF, and Elp lipoproteins within infected baboons was much more divergent compared to that with murine infection. For instance, although robust (1:12,800) antibody titers are generated against ElpA1, ElpA2, and BbK2.10 within 6 to 8 weeks p.i. in mice (33), infected baboons never generated antibody titers to these proteins greater than 1:400 throughout the 48 weeks of study. While it is important to note that two different strains of B. burgdorferi were utilized between the mouse studies and this baboon study (strains 297 and B31-CDC, respectively), both strains used are B. burgdorferi sensu stricto isolates and both encode highly similar OspE, OspF, and Elp paralogs (3, 17, 19, 20, 28). Since it has been shown that some of the OspE, OspF, and Elp paralogs are highly regulated by specific host factors (2, 4, 32, 33, 72, 73, 76), it is tempting to speculate that the different amounts of antibody being generated against some of the OspE, OspF, and Elp paralogs results from different amounts being expressed in the baboon (outbred population) and mouse (inbred population) environments. Along these lines, it was previously demonstrated with mice that the antibody response to most of the OspE, OspF, and Elp paralogs is greater following natural tick infection compared to that following syringe inoculation (32). Therefore, the adaptation process of the spirochete during natural transmission from the tick to the mammalian host may dramatically alter the expression of various proteins during the infectious process, which may have contributed to the differences in antibody reactivity observed.

One of the major obstacles in serodiagnosis of human Lyme disease, especially early disease, is the presence of cross-reactive antibodies in serum that bind B. burgdorferi WCL (67). Furthermore, the administration of the OspA-based Lyme disease vaccine has complicated the application of ELISAs based upon WCL to identify B. burgdorferi infection in vaccinated individuals (1, 26, 78). Recent attempts have been made to identify single antigens or a combination of antigens that result in sensitivity similar to that generated against B. burgdorferi WCL but with an increase in the specificity of reactive antibodies (15, 27, 30, 31, 40, 42-44, 58). In this study, a serological analysis of reactivity to the OspE-related, OspF-related, and Elp proteins was performed on patients with early and late Lyme disease, arbitrarily defined as less than 12 weeks p.i. or more than 12 weeks p.i., respectively. This analysis revealed that during early Lyme disease, reactivity to the OspE, OspF, and Elp paralogs was variable in degree and frequency. This was not altogether surprising as numerous other OspE-related, OspF-related, and Elp orthologs have been analyzed for their ability to detect human infection, yielding similar results (39, 43-45, 59, 71). At the same time, however, none of the prior studies had performed as comprehensive an analysis as the one performed here. Given our expanded analysis, which included all nine OspE, OspF, and Elp paralogs, we were able to determine that the majority of the human sera analyzed exhibited positive reactivity to OspE and/or ElpB1. This finding was especially exciting because the infecting B. burgdorferi strains are not known for any of the Lyme disease patients analyzed, indicating these two lipoproteins can identify many different North American isolates. Furthermore, since a format utilizing OspE and ElpB1 is not dependent on WCL, an added benefit is that it will not result in false-positive reactivity when individuals previously vaccinated for Lyme disease are screened. Additionally, while a combination of recombinant OspC and FlaB had a predictive capability similar to that of OspE and ElpB1, the overall sequence variability of OspC and potential cross-reactivity of FlaB would suggest that a combined OspE and ElpB1 immunoassay could help to increase serodiagnostic specificity for Lyme disease. Future studies with more human serum samples are needed to determine the applicability of a multivalent OspE and ElpB1 assay for diagnosing human Lyme disease.