|Home | About | Journals | Submit | Contact Us | Français|
Infection with Borrelia burgdorferi elicits robust yet disparate antibody responses in infected individuals. A longitudinal assessment of antibody responses to multiple diagnostic antigens following experimental infection and treatment has not previously been reported. Our goal was to identify a combination of antigens that could indicate infection at all phases of disease and response to antibiotic treatment. Because the rhesus macaque recapitulates the hallmark signs and disease course of human Lyme disease, we examined the specific antibody responses to multiple antigens of B. burgdorferi following infection of macaques. Five macaques infected with strain B31 and 12 macaques infected with strain JD1 were included in the analysis. Approximately half of these animals were treated with antibiotics at 4 to 6 months postinoculation. Antibody responses to several B. burgdorferi recombinant antigens, including OspC, DbpA, BBK32, OspA, and OppA-2, were measured at multiple points throughout infection. We have previously shown a decline in the response to the C6 peptide following antibiotic treatment. Responses to OspA and OspC, however, were variable over time among individuals, irrespective of antibiotic treatment. Not every individual responded to BBK32, but anti-DbpA IgG levels were uniformly high and remained elevated for all animals. All responded to OppA-2, with a decline posttreatment that was slow and incomplete. This is the first demonstration of B. burgdorferi OppA-2 antigenicity in nonhuman primates. The combination of DbpA, OspC, OspA, and OppA-2 with the C6 diagnostic peptide has the potential to detect infection throughout all disease phases.
The etiologic agent of Lyme disease, Borrelia burgdorferi, is transmitted to humans predominantly by ticks of the genus Ixodes, resulting in a disease of protean manifestations due to the organisms' systemic spread and promiscuous organ tropism. Infected individuals may exhibit a skin rash, arthritis, carditis, extreme fatigue, myalgia, and neurological dysfunction. In 2009, over 30,000 confirmed cases were reported to the Centers for Disease Control and Prevention. Hence, Lyme disease is the most commonly reported vector-borne disease in the United States (http://www.cdc.gov/lyme/stats/chartstables/casesbyyear.html). Treatment with antibiotics is generally effective, especially when they are administered soon after onset (4). As such, early and reliable diagnosis is critical for effective cure of Lyme disease patients.
Laboratory diagnosis of Lyme disease is typically made by the detection of patient serum antibodies specific for B. burgdorferi antigens. Detection of antibodies instead of antigens is necessitated by the absence of detectable spirochetes in the bloodstream once the organism has disseminated. Detection of the bacteria or bacterial antigens from blood or skin biopsy specimen is both invasive and of relatively low sensitivity (2). Antigen can sometimes be detected in urine and cerebrospinal fluid, but these tests are neither reliable nor recommended (45). Two of the most commonly used tests for diagnosis in North America are (i) the two-tier test, which includes an enzyme-linked immunosorbent assay (ELISA) and Western blotting using antigen derived from whole-cell lysates, and (ii) the C6 test, which detects antibodies to a specific peptide within a conserved region of the B. burgdorferi antigen VlsE (5, 30, 32). The C6 test has also been used experimentally for evaluation of treatment efficacy (36, 37).
While the two-tier and C6 tests are suitable for a majority of patients, neither is completely specific or sensitive enough to diagnose all patients. The C6 test, for instance, is more sensitive for human patients with early or late disseminated disease than for patients in the localized phase (5, 16). While the C6 peptide is highly conserved, other ELISA and Western blotting tests utilize whole-cell lysates or recombinant proteins from one species/strain/isolate, despite the enormous potential for antigenic variability that can preclude recognition by antibody. Western blotting will also specifically detect antibodies that recognize the antigen in a fully or partially denatured state, so those antibodies that target conformational epitopes are missed. Furthermore, the potential for patient serum cross-reactivity with antigens shared with other bacteria has led to stringent diagnostic criteria that can confound Western blot interpretation and thus hinder accurate diagnosis (6).
Along with C6 or VlsE, several other B. burgdorferi proteins that are known to elicit antibody responses in natural infections and have been incorporated into immunoblotting-based diagnostics include outer surface protein C (OspC) (44), the fibronectin-binding protein BBK32 (29), decorin-binding protein A (DbpA) (18), flagellar protein, FlaB, and outer surface protein A (OspA). The temporal induction and magnitude of the B cell response to each of these, characterized primarily in mice, are different. OspC, for example, is highly immunogenic, contains antigenic regions that vary among isolates, and is expressed early in infection and then repressed at the advent of humoral immune responses (28). Antibodies (IgM and IgG) to OspC often appear early in infection (24, 34). DbpA is expressed within the first few days of infection and continues postdissemination, so the antibody response can remain, even in late disease (18). FlaB is constitutive and immunogenic but holds the potential to be detected by cross-reactive antibodies from other bacterial species, especially when denatured as in Western blotting (17). OspA is expressed when B. burgdorferi is in the tick, but its expression is repressed as the spirochetes traverse to the host during tick feeding (40). However, studies indicate that expression of OspA and subsequent antibody responses may return during long-term infections in Lyme arthritis patients (3, 26). As a model for human Lyme disease, B. burgdorferi-infected nonhuman primates (NHP) are known to elicit bold and specific antibody responses to multiple antigens (33). Importantly, the responses may differ from those produced by mice (33). We also report here that nonhuman primates develop antibodies to the peptide transporter protein OppA-2. This response appears to decline with antibiotic treatment but more slowly and incompletely than the anti-C6 response (15).
This report presents the first assessment of temporal changes in levels of specific antibodies to multiple antigens following experimental infection with B. burgdorferi. By examining the responses in nonhuman primates before and after antibiotic treatment, we empirically determined antigens that would be appropriate for inclusion into a multiplex test to detect infection at multiple phases and as an indicator of treatment outcome.
Practices in the housing and care of animals conformed to the regulations and standards of the PHS Policy on Humane Care and Use of Laboratory Animals and the Guide for the Care and Use of Laboratory Animals. The Tulane National Primate Research Center is fully accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care-International. The Institutional Animal Care and Use Committee (IACUC) of the Tulane National Primate Research Center approved all animal-related protocols, including the infection, treatment, and sample collection from nonhuman primates. All animal procedures were overseen by veterinarians and their staff. For blood collection, monkeys were anesthetized with ketamine (10 mg/kg) by intramuscular injection. The infection and treatment of these animals were described previously (15). The experimental design reflects the assessment of antimicrobial efficacy against early disseminated infection (B31-infected macaques) or late disseminated infection (JD1-infected macaques). The repository of serum samples collected over time from these animals afforded the opportunity to assess serological responses to multiple antigens before, during, and after antibiotic treatment.
Five male rhesus macaques (Chinese origin), 3 to 4 years of age, were given three inoculations in the ventral midline of in vitro cultured late log-phase B. burgdorferi strain B31, isolate 5A19 (39), as follows: two subcutaneous 1.0-ml injections and one intradermal 0.1-ml injection, each containing 1 × 108 organisms diluted in sterile Hanks balanced salt solution (HBSS), for a total inoculum of 3 × 108 organisms. At 4 months postinoculation (p.i.), three of the five animals received antibiotic treatment consisting of one 50-mg tablet of doxycycline (Bio-Serv) twice daily for 28 consecutive days. This dose corresponded to >12 mg/kg of body weight/day to ensure that an effective blood level was achieved. Blood was collected at the following time points: 0, 2, 6, 10, 14, 16, 18, 22, 26, 28, 34, 40, and 47 weeks p.i.
Briefly, 24 rhesus macaques (Chinese origin) were each given an inoculation of 3.2 × 108 spirochetes of the JD1 strain (38) via needle and syringe. Four additional animals were sham inoculated as uninfected controls. At 27 weeks p.i., 12 of the infected animals and 2 of the controls were treated with ceftriaxone, followed by doxycycline. Animals received intravenous ceftriaxone at 25 mg/kg once daily for 30 days, followed by 60 days of oral doxycycline at 2 mg/kg twice daily. Twelve infected and two control animals were sham treated. Blood was collected prior to inoculation and every week for more than a year until the time of euthanasia (54 to 60 weeks p.i.). The serum antibody responses of half of the animals (randomly selected) in each test group were included in our assessment.
Each B. burgdorferi antigen was produced as a glutathione S-transferase (GST) fusion protein by inserting the gene of interest into the pGex 4T-1 vector (GE Healthcare). Genes were cloned by primer insertion of BamHI and XhoI restriction enzyme sites for digestion and ligation into the expression vector. The full open reading frames of DbpA and BBK32 were amplified from B. burgdorferi strain B31 template DNA using the following primers: 5′-CATCGGATCCATGATTAAATGTAATAATAAAAC-3′ (DbpA forward), 5′-CATCCTCGAGTTAGTTATTTTTGC-3′ (DbpA reverse), 5′-CATCGGATCCATGAAAAAAGTTAAAAGC-3′ (BBK32 forward), and 5′-CATCCTCGAGAATATAATTTAATAAGGTC-3′ (BBK32 reverse). ospA genes from strains B31 and JD1 were cloned by insertion of BamHI and XhoI restriction enzyme sites so as to omit the leader sequence, beginning with the 20th amino acid of the coding sequence, using the conserved forward primer 5′-CATCATGGATCCAATGTTAGCAGCCCTGACGAG-3′ and reverse primers 5′-GACTAGCTCGAGTTATTTTAAAGCGTTTTTAATTTCATCAAG-3′ (B31) and 5′-GACTAGCTCGAGTTATTGTAAAGCGTTTTTAATTTCATCAAG-3′ (JD1); ospC was amplified beginning at the 20th amino acid, downstream of the leader sequence, using primers 5′-CATAGGATCCAATTCAGGGAAAGATGGGAAT-3′ (forward) and 5′-CATACTCGAGTTAAGGTTTTTTTGGACTTTCTG-3′ (reverse); the entire oppA2 gene was amplified with primers and cloned into the pGex 4T-1 vector using 5′-GATAGGATCCATGAAATTACAAAGGTC-3′ (forward) and 5′-GAGCCTCGAGTTATTTATTTTTTAATTTTAGCTG-3′ (reverse). Plasmids with productive insertions were used to transform Top10 cells and screened. To improve protein expression, BL21 competent cells were transformed with each plasmid. Protein expression was induced with isopropyl β-d-1-thiogalactopyranoside (IPTG) for 4 h or overnight. GST fusion proteins were extracted from Escherichia coli and purified with a Pierce B-PER Bacterial Protein Extraction Reagent GST Fusion Purification Kit (ThermoScientific). Protein expression and purity were confirmed by Coomassie blue gel staining and immunoblotting using the mouse monoclonal anti-GST antibody clone GST-2 (Sigma Life Science).
Specific IgG and IgM titers were measured and compared. Specificity was measured with preimmune serum and immune serum tested on GST alone purified from bacterial lysate, and none of the sera reacted.
The ELISAs were performed essentially as described previously (21). Briefly, plates were coated with 0.5 to 1.0 μg/well of recombinant protein. After a blocking step with nonfat milk, serum was added at a dilution of 1:200. Anti-GST antibody (Sigma) was used as a positive control, and peroxidase-labeled goat anti-monkey IgG or goat anti-monkey IgM (KPL, Gaithersburg, MD) secondary antibodies were used at a 1:1,000 dilution. Absolute optical density at 450 nm (OD450) was determined for each sample by subtracting the average value given by preimmune serum from that animal. The number of infected animals that were positive by each test (above the mean value plus 3 standard deviations [SD] for preimmune serum) versus those that were negative was compared by Fisher's exact test for significance.
For determining antibody titers, serum samples were diluted 2-fold until reactivity was absent. Specifically, endpoint titers were calculated as the dilution at which the sample reached <0.05 when the average preimmune serum values plus 2 SDs were subtracted. The reported values are the reciprocal of the EPDT. To compare changes in antibody titer with treatment, the percent change was calculated as follows: (peak titer − endpoint titer)/(peak titer).
Serum samples from macaques were tested for IgG antibodies against four different proteins (DbpA, BBK32, OspA, and OspC) that have been used in Lyme disease diagnosis by standard ELISA or immunoblotting. The data presented here are from five macaques followed for 11 months of infection; two were untreated, and three were treated with antibiotics. We have supplemented these results with other serum specimens from strain JD1-infected and treated macaques. A summary of the antigen reactivities among these groups can be found in Table 1.
In mice, DbpA is known to be a T-cell-independent antigen to which a high proportion of the antibody response is directed (9, 42). It has been tested as a vaccine immunogen (11, 20, 21) and as a diagnostic antigen in immunoblotting (45) and ELISA (35). The infected monkeys all produced high-titer, sustained responses to DbpA (Fig. 1). The anti-DbpA titers remained elevated in the five animals that were infected with strain B31 spirochetes (Fig. 1A and andB),B), while reactivity was more variable in animals infected with spirochetes of the JD1 strain (Fig. 1C and andD).D). The origin of the recombinant detection antigen (B31) may affect reactivity, but all animals showed a detectable response to DbpA. BBK32 is another extracellular matrix (fibronectin) binding protein that has been tested as both a vaccine (11) and a diagnostic antigen (35). Only three of the five B31-infected macaques produced strong responses to BBK32 though the titer dropped substantially for one of the treated monkeys (Fig. 2). Due to this low responsiveness, we did not test monkeys that were inoculated with JD1 spirochetes for responses to BBK32.
For interpretation of the antibody response to OspA, it should be mentioned that these animals were infected by needle inoculation of in vitro cultured spirochetes, which express OspA. Because expression of OspA is downregulated in the feeding tick and remains as such in the mammalian host immediately after transmission, the initial responses are not representative of a natural infection. Thus, all animals responded to OspA, as predicted (Fig. 3). However, fluctuations in antibody titers ensued with time in both treated and untreated animals. Following an initial decline between weeks 2 and 10 for each B31-infected animal (Fig. 3A and andB),B), the titers either declined and rose again or remained elevated, as shown for the JD1-infected animals between week 27 and necropsy (Fig. 3C and andD).D). OspC is known to be an immunodominant antigen that is expressed upon transition to the mammalian host environment (19, 40). As such, responses to OspC arose very early (within 2 weeks for most animals) after infection and steadily declined over time among individuals, irrespective of antibiotic treatment (Fig. 4A and andB).B). Importantly, all animals generated an early response to OspC, and antibodies from JD1-infected animals recognized strain B31 OspC.
The oppA-2 gene belongs to a B. burgdorferi oligopeptide permease operon that is induced during changes in environmental conditions (43). A similar protein has also been used as a vaccine antigen with other bacterial species (41, 46). All infected animals generated anti-OppA-2 antibody responses (Fig. 5). A decline was evident for all three treated strain B31-infected animals shortly after treatment (Fig. 5A and andB).B). This decline was not marked by a return to background levels but, instead, leveled off, and may have even increased, with induction of IgM late in two of three treated animals. (see Fig. 7). An additional 12 animals that were infected with B. burgdorferi strain JD1 all had high-titer anti-OppA-2 antibodies as well (Fig. 5C to toF).F). Here, a decline was seen initially for most treated animals, whereas the response generally remained elevated in the untreated animals.
We also determined the reciprocal endpoint dilution titers of serum antibodies to each antigen from animals at the peak response (16 to 22 weeks) and at necropsy (47 weeks). Figure 6 shows a comparison between the EPDTs of antibodies to the five recombinant antigens and the C6 peptide. The levels of IgG produced in response to DbpA were the highest, and they were lowest for C6. The order of relative responses is DbpA > OspA/OspC/OppA-2 > C6. The percent decline (Table 2) in C6 titer for treated animals was 93.8% for animal GB56 and 98.4% for animals GA59 and FK38. The only other antigens against which antibody titers declined with treatment in all animals but not in both untreated animals were OppA-2 and OspC. For OppA-2, a decline in titers occurred for treated animals GA59 (87.5%) and FK38 (96.9%) but not GB56 (50%). The anti-OspC titers among JD1-infected animals declined for all that were untreated, so the decline among treated animals is not relevant. This will need to be tested on more animals to determine if these changes are statistically significant.
Longitudinal measurement of the IgM response to each full-length antigen was also included in the assessment (Fig. 7). The anti-OspA IgM responses followed a predictable pattern (Fig. 7A). However, fluctuations in IgM to the other antigens tested, including OspC (Fig. 7B), DbpA (Fig. 7C), and OppA-2 (Fig. 7D) were evident beyond primary infection. For example, treated animal GB56 generated increased anti-OspC IgM levels at weeks 22 and 40. Fluctuations in both the anti-DbpA and anti-OppA-2 IgM response were evident for each of the few animals tested, regardless of antibiotic treatment status.
The importance of rational design for diagnostic tests is predicated by diversity in the immune responses produced by infected individuals. The use of multiplex testing with cytometric bead-based technology has expanded vigorously in recent years and has become a platform for serological testing. In the case of Lyme disease, the C6 peptide has demonstrated superior reliability as a standalone antigen for diagnostic testing. However, the inclusion of multiple antigens into a serologic test may increase sensitivity. This must be accomplished without compromising test specificity.
To gain insight into antigens that may offer the potential for broad (encompassing all phases of disease) sensitivity, we examined longitudinal responses: (i) in the most appropriate animal model for human Lyme disease, the rhesus macaque, and (ii) both pre- and postantibiotic treatment to include assessment of responses that may be linked to disease resolution. In doing so, we discerned a combination of antigens that would be appropriate for inclusion in a multiantigen serologic test.
In mice, DbpA is known to be a T-cell-independent antigen to which a high proportion of the antibody response is directed (9, 42). It has been tested as a vaccine immunogen (11, 20, 21) and as a diagnostic antigen (35, 45). All NHP infected with either B. burgdorferi strain in this study also produced staunch antibody responses to DbpA. It is likely that this is a T-cell-independent antigen for NHP as well. BBK32 is another extracellular matrix binding protein that has been tested both as a vaccine (11) and as a diagnostic antigen (35). Though the titer dropped substantially for one treated B31-infected monkey, only three of the five macaques produced veritable antibody responses to this protein. Due to the strong and sustained responses to DbpA in all animals tested to date, it appears a valuable diagnostic antigen. BBK32 has a lower potential for sensitivity and may not add value to a multiplex test.
Despite the artificial induction of primary anti-OspA responses by inoculation of in vitro cultured spirochetes, we found the fluctuations in response over time compelling. We detected ospA transcript from persistently infected macaques (15), so we reasoned that their anti-OspA antibody responses may not decline predictably over time. In addition, OspA antibodies have been associated with arthritis during disseminated infection (3) and have more recently been shown to be elevated in patients with posttreatment Lyme disease syndrome (12). For its potential to detect infection at the late disseminated phase, inclusion of OspA may benefit a multiplex test. OspC is a well-characterized, highly immunogenic B. burgdorferi lipoprotein. Heterogeneity in ospC among different isolates and strains has been well documented. However, this does not appear to result in type-specific responses to OspC (23), so it remains a good candidate diagnostic antigen. Antibody responses to OspC may be expected to arise early and steadily decline, as this antigen's expression is known to be shut off after the advent of the humoral response (28). This is indeed what we observed.
oppA-2 transcript was detected in heart tissue of mice postantibiotic treatment (8). Over a decade ago, two-dimensional electrophoresis and immunoblotting revealed antibody responses of infected individuals that targeted an ABC transporter protein of Borrelia garinii (25). Human patients are known to produce antibodies to the oligopeptide permease A-type protein from B. burgdorferi (7, 14, 31). The OppA-2 protein was also demonstrated to be antigenic in rabbits (13) and in mice (7, 31). The gene is on the chromosome, whereas the gene from which C6 is derived is located on a plasmid that can be lost during infection (39) and possibly also posttreatment (10). OppA-2 transcript was shown to be produced by B. burgdorferi in mice postantibiotic treatment, as an indicator of their metabolic viability. Given the broad reactivity of macaques to OppA-2 and the decline in response, albeit moderate, posttreatment, this protein could prove to have utility as a diagnostic antigen both before and after treatment. While we did not see a response to OppA-2 using preimmune serum, the modest homology (27) that this protein may have with similar bacterial proteins necessitates its careful screening with multiple samples for specificity.
Continuous IgM responses beyond primary introduction of antigen are peculiar but have been observed in B. burgdorferi infection previously (1, 22). Primarily through the use of immunoblotting, the continual detection of B. burgdorferi antigen-specific IgM is well documented; however, fluctuations in specific antibody levels over time have not generally been presented. The induction of IgM-producing cells, particularly by T-cell-independent antigens (42), may provide a mechanism for such fluctuations and persistent IgM titers.
The results of this study indicate that the antigens OspA, OspC, DbpA, and OppA-2 may offer distinct benefits when they are combined with the C6 peptide into a multiantigen diagnostic test. However, there are limitations to this study, including the inoculum dose, route, and the small number of animals tested. While the inoculum dose used here may not accurately reflect the dose that naturally infected humans would receive, the dynamic responses to particular antigens may be comparable in a general sense. Use of an outbred animal model and two different B. burgdorferi strains yielded patterns of antibody response that were remarkably similar per antigen. We cannot assert that these responses reflect those which may be observed in humans, but future studies are aimed at testing the responses to these specific antigens from a broad panel of Lyme disease patients.
For a multiantigen quantitative diagnostic test, specifically the inclusion of OspC could allow for improved sensitivity in the localized phase and that of OspA has the potential to alert a physician about the possibility of development of late arthritis. Detection of OppA-2 increases in antibody levels posttreatment may be a sign of persistent spirochetes or recrudescence as all three treated animals in the strain B31-inoculated experimental group had indications of persistent B. burgdorferi infection (15). The early (~4 weeks p.i.) and strong response to DbpA by a majority of animals indicates that DbpA may also lend diagnostic value to a multiplex test. Observation of dynamic longitudinal responses to various antigens over time may provide insight for optimal diagnostic antigen combination in developing diagnostic tests for other infections with pathogens that possess multiple candidate diagnostic antigens.
This work was supported by a TNPRC Pilot Study Grant (M.E.E.), NIAID grant R01-AI042352 (M.T.P.), and NCRR grant RR00164.
We thank Lisa Bowers for technical assistance.
Published ahead of print 20 June 2012