Immunological Responses to the Relapsing Fever Spirochete Borrelia turicatae in Infected Rhesus Macaques: Implications for Pathogenesis and Diagnosis

The global public health impact of relapsing fever (RF) spirochetosis is significant, since the pathogens exist on five of seven continents. The hallmark sign of infection is episodic fever and the greatest threat is to the unborn.

ABSTRACT

The global public health impact of relapsing fever (RF) spirochetosis is significant, since the pathogens exist on five of seven continents. The hallmark sign of infection is episodic fever and the greatest threat is to the unborn. With the goal of better understanding the specificity of B-cell responses and the role of immune responses in pathogenicity, we infected rhesus macaques with Borrelia turicatae (a new world RF spirochete species) by tick bite and monitored the immune responses generated in response to the pathogen. Specifically, we evaluated inflammatory mediator induction by the pathogen, host antibody responses to specific antigens, and peripheral lymphocyte population dynamics. Our results indicate that B. turicatae elicits from peripheral blood cells key inflammatory response mediators (interleukin-1β and tumor necrosis factor alpha), which are associated with preterm abortion. Moreover, a global decline in peripheral B-cell populations was observed in all animals at 14 days postinfection. Serological responses were also evaluated to assess the antigenicity of three surface proteins: BipA, BrpA, and Bta112. Interestingly, a distinction was observed between antibodies generated in nonhuman primates and mice. Our results provide support for the nonhuman primate model not only in studies of prenatal pathogenesis but also for diagnostic and vaccine antigen identification and testing.

INTRODUCTION

Relapsing fever (RF) spirochetosis is a neglected global disease. In parts of Africa, RF spirochetosis is a common bacterial infection (1), and the disease is a significant cause of hospital admissions and child mortality (2–6). The causative agents are Borrelia species that are transmitted by the human body louse, or ixodid, and argasid ticks (1–4). The manifestation of disease in humans includes recurrent febrile episodes, rigors, vomiting, severe headache, neurological symptoms, muscle and joint aches, and tachycardia (1). Antibiotic treatment may result in the Jarisch-Herxheimer reaction, which is caused by a cytokine release leading to shock (5) and even death (6, 7). Mortality of tick-borne RF spirochetosis is 4 to 10% and is associated with the burden of spirochetes in the blood (8). RF borreliosis is particularly devastating on fetal and neonatal health (9, 10). For example, in Tanzania a perinatal mortality rate of 436/1,000 was reported for Borrelia duttonii (11). The disease also has a severe impact in developing countries because of the nonspecific, malaria-like clinical manifestation of the disease. Importantly, with the geographic distribution of RF spirochetes largely overlapping with malaria (12) and studies often indicating a misdiagnosis (13, 14), the true morbidity of RF is underappreciated.

The reduction in spirochete levels and eventual clearance have been shown in animal models to be a direct result of the antibody response, especially IgM and IgG3 isotypes (15, 16). The clearance by lymphocytic response was established by Newman and Johnson (17), who showed not only the importance of the B-cell response, but that of a T-independent B-cell response. Subsequent studies have demonstrated neutralization (18) and a directly bactericidal (19) role of serum IgM in controlling RF spirochetemia. The contribution of B-cell subsets to RF pathogen control has been further delineated in mice (16, 20, 21).

Rodent models have contributed immensely to the understanding of infectivity, host-pathogen interactions, and immune responses to infection (22–27). For example, transmission studies in Borrelia turicatae demonstrated that RF spirochetes enter the host within seconds of tick bite (28), indicating the importance of preventing early mammalian infection. Moreover, vaccination of mice with the Borrelia hermsii variable tick protein (Vtp) has guided vaccine strategies. Vtp is produced in the salivary glands of Ornithodoros hermsi and subsequently downregulated once the pathogens are detectable in murine blood (29). Vaccination studies with Vtp indicated that RF spirochete surface proteins produced in the tick salivary glands could be ideal immunological targets to prevent the establishment of infection (30).

Mice are natural reservoir hosts and may have limitations as models for testing intervention and therapeutic strategies. Thermoregulation in mice varies, and they are a limited model to further understand the Jarish-Herxheimer reaction. Mammals have evolved unique thermoregulatory mechanisms in defense against pathogens, with rodents typically remaining afebrile or decreasing body temperatures in response to bacterial challenge and endotoxin administration (31–35). Therefore, mice may not be ideal for the evaluation of vaccine candidates and therapeutics that prevent the clinical sign of fever, which is a hallmark feature of RF.

Nonhuman primates (NHPs) infected with RF spirochetes accurately mimic human disease. A 1938 report published by Edward Francis showed that NHPs infected with B. turicatae by tick bite exhibited morbidity and mortality commonly observed with human disease (36). We have also demonstrated human-like illness with this model. Four rhesus macaques were infected with B. turicatae by tick transmission, and radio telemetry was used to quantify the intricacies of infection (22). Multiple instances of febrile episodes, high spirochete densities in blood, and disruption of cardiac function were observed.

In the present study, we further characterized the immune responses of NHPs that were infected with B. turicatae by tick bite (22). We originally hypothesized that B. turicatae would induce a TH2 type immune response, with concomitant induction of B-cell proliferation and antibody production. Rather, we found that in peripheral blood cells, B. turicatae induced TH1 type cytokines (interleukin-1β [IL-1β] and tumor necrosis factor alpha [TNF-α]), and significant declines in B-cell populations were observed soon after infection. Changes in peripheral blood lymphocyte subsets, immune mediator production by stimulated peripheral blood mononuclear cells (PBMCs), and antibody responses reflect a distinct response to RF Borrelia in NHPs. We evaluated antibody responses to a known conserved surface protein, the Borrelia immunogenic protein A (BipA) (24, 25), and two newly identified surface proteins, Bta112 and the Borrelia repeat protein A (BrpA). Bta112 and BrpA are upregulated in the tick and were evaluated to determine their antigenicity once B. turicatae enters the mammalian host. Our results demonstrate differences in the host antibody specificity between mice and NHPs infected with B. turicatae and further indicate the significance of macaques as a model that most accurately represents human RF borreliosis.

RESULTS

Coculture of macaque PBMCs with B. turicatae elicits inflammatory response mediators.

Given the high numbers of RF spirochetes that are observed in the blood during febrile episodes, we sought to measure immune mediators produced by PBMCs in response to stimulation with borreliae. In the analysis of 23 cytokines produced in reponse to stimulation with B. turicatae, both commonalities and differences with B. burgdorferi (the Lyme disease [LD] causing agent) were observed. While both borrelial pathogens elicited TNF-α, IL-10, granulocyte colony-stimulating factor (G-CSF), and IL-12/23p40, B. turicatae induced a statistically significant higher level of IL-1β and soluble CD40 ligand (sCD40L) than B. burgdorferi (Fig. 1A and D, respectively). We tested stimulation of PBMCs derived from a naive, uninfected monkey (JD03) in addition to PBMCs derived from infected animals. To preserve the viability of the spirochetes and retain soluble factors, the stimulations were performed with spirochetes in their own growth media (shown as indicated as “Btmedia” and “Bbmedia,” respectively). However, components of the media also had a moderate stimulatory effect for some cytokines/chemokines. Figure 1A shows IL-1β responses of naive macaque PBMCs stimulated with borreliae, indicating a significant induction of this inflammatory cytokine specifically by B. turicatae. For IL-1β, significant differences were observed at the 12-hour time point when comparing B. turicatae to uninoculated BSK-H medium (BSK; P = 0.0231) and B. burgdorferi to BSK (P = 0.001). At 24 h, significant differences in these two groups were observed as well (B. turicatae versus BSK, P < 0.0001; B. burgdorferi versus BSK, P = 0.0047). In Fig. 1B, the effect on TNF-α production indicates that both Borrelia species induce production of this inflammatory cytokine by PBMCs. At 12 h, significance was observed when we compared B. turicatae versus BSK (P = 0.0007) and B. burgdorferi versus BSK (P < 0.0001). No difference was observed between the B. turicatae media and BSK, and yet the quantity of TNF-α induced by B. turicatae over that of B. turicatae media was significant (P = 0.0013), indicating that soluble factors do not drive the induction of TNF-α by B. turicatae. At 24 h, each of these differences remained significant (B. turicatae versus BSK, P = 0.0002; Bb versus BSK, P = 0.0003; B. turicatae media versus B. turicatae, P < 0.0001). Figure 1 also shows the specific differences in the induction of G-CSF, sCD40, and IL-12/23 from naive PBMCs stimulated with borreliae (Fig. 1C, D, and E, respectively). Significant changes in G-CSF production by PBMCs stimulated with B. turicatae were only observed at the 24-h time point. Here, stimulation with B. turicatae versus BSK was significant (P = 0.0005), as was stimulation with B. burgdorferi versus BSK (P = 0.0002). For the soluble CD40 ligand (sCD40L), significant differences were observed only at the 12-h time point, with stimulation of B. turicatae compared to B. turicatae media demonstrating significance (P = 0.0085), along with B. turicatae versus BSK (P = 0.0027) and B. turicatae versus BSK (P = 0.0047). For IL-12/23, significant differences were seen at both the 12- and the 24-h time points when we compared B. turicatae versus BSK (12 h, P = 0.0007; 24 h, P = 0.0013) and B. burgdorferi versus BSK (12 h, P = 0.0033; 24 h, P = 0.0343). Figure 2 shows IL-1β responses among infected monkeys. Stimulation of day 14 postinfection (p.i.) PBMCs with B. turicatae versus B. turicatae media alone resulted in a significant increase in this inflammatory mediator for all three monkeys that were infected. A 2- to 3-fold increase in quantity of IL-1β produced in response to B. turicatae compared to medium alone indicates the specific effect of the pathogen. Specifically, significant differences were observed at the 12- and 24-h time points for JB60 (12 h, P = 0.0070; 24 h, P = 0.0010), JB23 (12 h, P = 0.0022; 24 h, P = 0.0005), and IN57 (12 h, P = 0.0008; 24 h, P = 0.0005) when B. turicatae versus B. turicatae media was compared.

FIG 1.

IL-1β (A), TNF-α (B), G-CSF (C), sCD40 (D), and IL-12/23 (E) responses of naive macaque PBMCs stimulated with borreliae. PBMCs obtained at day 0 from animal JD03 were stimulated with B. turicatae (Bt), B. burgdorferi (Bb), filtered BSK-H medium derived from B. turicatae or B. burgdorferi cultures (Btmedia and Bbmedia), or uninoculated BSK-H medium (BSK) or left untreated (no trt). Supernatants were collected at 12 and 24 h to measure the inflammatory mediators by an NHP-specific 23-plex cytokine bead assay.

FIG 2.

IL-1β production by the PBMCs of B. turicatae-infected macaques. Cells isolated from blood on day 14 post-tick feeding were incubated with B. turicatae at a 10:1 ratio of spirochetes to cells (+Bt), left untreated (no trt), incubated with BSK-H medium (+BSK), or incubated with filtered modified BSK medium derived from B. turicatae cultures (+Btmedia). Supernatants were collected at 12 and 24 h, for measurement of inflammatory mediators by an NHP-specific 23-plex cytokine bead assay.

Immune regulatory molecules in serum were also quantified. We collected blood droplets between days 0 and 14, but serum was collected beginning on day 14, since our intent was to evaluate antibody responses. Therefore, we used available sera to test in the 23-plex cytokine magnetic bead panel. The immune mediators that were elevated in serum at various time points included IL-10, sCD40L, IL-8, and MCP-1 (Fig. S2). Both IL-10 and MCP-1 were elevated at the earlier time points (14 and 28 days) but declined by 6 weeks p.i. In contrast, sCD40L and IL-8 appeared to be elevated throughout the infection period in monkeys inoculated with B. turicatae (IN57, JB60, and JB23) compared to levels in the animal fed upon by uninfected ticks (JD03).

Characteristic peripheral B-cell depletion during acute infection.

T- and B-cell phenotypic analyses were performed with PBMCs at the day 0, 14, and 70 p.i. time points from all four NHPs. With respect to the T-cell phenotype, only general CD4 and CD8 T-cell phenotypes were measured in all three time points after infection. Only one animal (IN57) had reduced CD3⁺ populations, detected at day 14 postinfection (Table S1 and Fig. 3), and the percentages of all four subsets of CD3 cells (CD4⁺ CD8⁻, CD4⁺ CD8⁺, CD4⁻ CD8⁺, and CD4⁻ CD8⁻) were reduced. The other three animals showed a moderate increase in peripheral T cells at 2 weeks p.i. that declined by day 70.

FIG 3.

Frequency of B and T cells in the peripheral blood after infection with B. turicatae. PBMCs were subjected to flow cytometry to detect the relative percentages of B cells (CD20⁺) and CD4⁺/CD8⁺ T-cell subsets. Each staining experiment was performed twice, and the standard deviation is indicated with error bars.

B-cell subsets were distinguished by a panel of markers that included CD5, CD20, CD21, CD138, IgM, IgD, CD27, and CD38. Notable reductions in the percentages of B cells were observed in the serum 14 days after infection, suggesting an infection-induced B-cell depletion (Table 1; see also Fig. S1 in the supplemental material). As shown in Table 1, the B-cell depletion was due to the loss of CD5 (B-1a cells, marker of naive or immature B-cells [32, 33, 37]), CD21 (marker of B-cell differentiation and maturation [38–40]), CD86 (activation marker [41, 42]), and CD138 (plasmablasts [37]). In a subsequent staining, we looked at IgM⁺ B cells and switched memory (CD27⁺ IgD^–), non-switched memory (CD27⁺ IgD⁺), naive (CD27⁻ IgD⁺), double negative (CD27⁻ IgD^–), and CD27^high CD38^high plasmablasts. A precipitous and global decline in peripheral B-cell populations was observed in all animals at day 14 p.i. (Table 1 and Fig. 3). The B-cell percentages and different B-cell subsets returned to near preinfection levels at day 70 postinfection in all animals. The percent drop in total B-cell frequency between days 0 and 14 was significant for all monkeys. Specifically, the CD20⁺ lymphocytes decreased by 43% for JD03, 84% for IN57, 80% for JB60, and 56% for JB23.

TABLE 1.

Changes in peripheral B-cell subsets following infection with B. turicatae

Animal (days postinfection)a	Mean % of cells/total lymphocytes ± SEM
	CD20+	CD20+ CD5+	CD20+ CD21+	CD20+ CD86+	CD20+ CD138+	IgM+	Switch memory (CD27+ IgD–)	Nonswitched memory (CD27+ IgD+)	Naive (CD27− IgD+)	Double negative (CD27− IgD–)	Plasmablast (CD27high CD38high)
JD03u* (0)	26.3 ± 0.71	0.652 ± 0.032	12.87 ± 0.88	16.43	0.051 ± 0.008
JD03u* (14)	30.3 ± 0.85	0.809 ± 0.069	13.87 ± 2.48	14.67	0.113 ± 0.002
JD03 (0)	31.8	0.86	15.20	18.83	0.08	16.75	9.4604	15.8388	0.871	0.61908	0.1059
JD03 (14)	18.2 ± 5.09	0.927 ± 0.14	8.71 ± 0.73	8.13	0.233 ± 0.019	16.9332	7.7559	9.3627	7.3233	6.4581	0.11546
JD03 (70)	54.7 ± 20.93	2.39 ± 1.78	26.69 ± 8.24	18.87	0.240 ± 0.024	13.4506	6.5836	7.739	4.5562	2.9212	0.10268
IN57 (0)	23.9	1.01	12.02	12.45	0.30	54.0015	12.7185	37.947	14.039	4.8094	0.0915
IN57 (14)	3.915 ± 1.74	0.225 ± 0.177	1.95 ± 1.17	3.25	0.045 ± 0.006	2.05418	0.43089	0.9214	1.07316	0.28455	0.04293
IN57 (70)	58.25 ± 2.05	1.96 ± 0.22	26.76 ± 2.72	28.63	0.342 ± 0.318	42.6855	9.3729	25.8501	16.716	7.761	0.16642
JB60 (0)	43.8	1.61	24.66	24.05	0.11
JB60 (14)	8.97 ± 3.02	0.276 ± 0.12	3.99 ± 1.68	8.00	0.026 ± 0.029	4.04336	2.72517	2.58857	1.01767	0.496541	0.2394
JB60 (70)	54.4 ± 2.97	2.73 ± 0.23	27.92 ± 1.05	27.67	0.125 ± 0.141	40.3975	18.532	36.2165	1.0283	0.74015	0.47888
JB23 (0)	7.3	0.42	3.78	3.96	0.11
JB23 (14)	3.23 ± 0.52	0.132 ± 0.068	1.51 ± 0.41	2.43	0.0158 ± 0.018	1.39282	1.34706	0.94094	0.31746	0.252538	0.30615
JB23 (70)	24.05 ± 8.27	0.864 ± 0.35	10.46 ± 1.31	9.50	0.071 ± 0.062	13.7839	15.4583	12.8271	0.7774	0.81627	0.53768

^a*, Uninfected (fed upon by uninfected ticks).

Evaluation of Bta112 between strains of B. turicatae.

Bta112 was further evaluated as an antigen because computational analyses suggested the protein was exposed on the surface of RF spirochetes. The PROSITE InterPro database identified a predicted lipid attachment site at the N terminus of the protein (Fig. S3). PSIPRED and the Phobius prediction server suggested that the Bta112 was rich with alpha-helices and the C terminus of the protein was soluble and positioned toward the extracellular environment, respectively. Sequence analysis of Bta112 between B. turicatae 91E135, FCB, TCB1, TCB2, and 99PE-1807 indicated the presence of an intact gene that coded for a protein that was nearly identical in all B. turicatae isolates evaluated (Fig. S3). Given the presence of Bta112 in multiple B. turicatae isolates, we evaluated the protein further.

Expression of recombinant Bta112, temperature-mediated production, and surface localization of the native protein.

To evaluate serological responses to B. turicatae Bta112, the gene was expressed as a recombinant fusion protein. Bta112 was overexpressed in BL21 Star (DE3) cells (Fig. 4A), and rabbit immune serum was generated against the recombinant protein. Since native Bta112 is upregulated by B. turicatae during culture at 22°C relative to 35°C (43), spirochetes grown at both temperatures were evaluated to assess temperature-mediated protein production. Optical density analysis of immunoblots probed with the rabbit serum sample generated against recombinant Bta112 (rBta112) indicated 3.2-fold increase of the protein in B. turicatae grown at 22°C versus 35°C (Fig. 4B, upper panel). The rabbit’s preimmunization serum sample was used as a negative control (Fig. 4B, middle panel). Moreover, a serum sample generated against B. turicatae FlaB was used as a control to indicate that similar protein loads were electrophoresed in the immunoblotting assays (Fig. 4B, lower panel).

FIG 4.

Expression of Bta112 as a recombinant protein, temperature-mediated protein production and surface localization of Bta112. (A) Bta112 was produced in E. coli and purified. E. coli samples were taken prior to induction (Tp0), 3 h after induction (Tp3), and after purification of the protein. (B) Immunoblots using rabbit serum samples generated against rBta112 indicated the protein’s increased production at 22°C compared to 35°C (upper panel). Preimmune serum samples (middle panel) and serum samples generated to the flagellin (FlaB) protein (lower panel) are also shown. Immunoblotting was also performed to evaluate the surface localization of Bta112. (C) Proteinase K (PK) was used at concentrations of 5 to 200 μg/ml. Membranes were probed with anti-Bta112 serum samples (upper panel), and anti-FlaB (lower panel) as a control, to indicate equal protein loads in the gels (50). Molecular weight markers (MWM) are shown on the left of the gel in panel A, and molecular masses are indicated on the left of each immunoblot.

Performing proteinase K and immunoblotting assays with B. turicatae grown at 35°C indicated that the Bta112 was surface localized (Fig. 4C). Bta112 was degraded after incubation with increasing concentrations (5, 50, and 200 μg/ml) of proteinase K for 15 min (Fig. 4C, upper panel). The relative density of the periplasmic protein FlaB in 5 μg/ml of proteinase K compared to the 0 μg/ml of proteinase K control was 101%. The relative densities of FlaB in 50 and 200 μg/ml of proteinase K were 93 and 90%, respectively, indicating that the spirochetes’ membranes remained intact (Fig. 4C, lower panel). Collectively, these results supported that Bta112 was surface localized and that the protein’s production was elevated at 22°C. Given these findings, the antigenicity of rBta112 was assessed.

Serological responses to B. turicatae surface proteins.

Given variations in humoral responses between mammalian species (44, 45), we compared the antigenicity of B. turicatae rBta112, rBrpA, and rBipA using serum samples from NHPs and mice that were infected by tick bite. Immunoblotting indicated various serological responses between NHPs and mice to the recombinant proteins (Fig. 5A to F). All four NHPs produced antibodies that bound to B. turicatae protein lysates, rBta112, and rBipA, while serological reactivity to rBrpA was only detected in JB60 (Fig. 5B). An immunoblot from two mice represented the eight remaining animals (Fig. 5E and F). All of the mice seroconverted to rBipA, and two of eight animals seroconverted to rBta112, while none of the animals seroconverted to rBrpA. Probing immunoblots with a monoclonal antibody for the six-histidine epitope was used to as a control for the expected molecular weight of each protein (Fig. 5G). These findings suggested various serological responses to RF spirochete antigens between NHPs and mice.

FIG 5.

Immunoblot analysis using serum samples from NHPs and mice to rBta112, rBrpA, and rBipA. Immunoblots of JB23 (A), JB60 (B), IN57 (C), and JD03 (D) are shown. Membranes were probed with preinfection serum samples (left blot) and serum samples collected at days 84 (JB23, JB60, and IN57) and 100 (JD03). (E and F) Immunoblots from two mice, which represent the remaining 10 animals, are shown. (G) Membranes were also probed with an anti-six-histidine monoclonal antibody and indicate the molecular weight of each recombinant protein. An asterisk is placed next to each recombinant protein that was antigenic. Molecular weights are indicated at the left of each immunoblot.

Since rBipA and rBta112 were immunogenic by immunoblotting in all four NHPs, we further evaluated their serological responses over the duration of the study using enzyme-linked immunosorbent assay (ELISA). Assessment of rBipA and rBta112 indicated the temporal persistence of IgG responses to the recombinant proteins (Fig. 6). JB23, JB60, and IN57 generated IgG responses for at least 84 days after the animals were infected with B. turicatae by tick transmission (Fig. 6A to C). These responses were statistically significant compared to the preinfection serum samples for each animal to a given recombinant protein. JD03 was a control animal, as described in our previous report (22), and evaluating IgG response at three time points (7, 27, and 43 days) after feeding uninfected ticks indicated that tick saliva did not generate cross-reactive antibody responses to rBipA and rBta112 (Fig. 6D). After the animal was infected by tick transmission, IgG responses to rBipA and rBta112 were detected 42 days after feeding (Fig. 6D, day 100 of the study). Statistically significant IgG responses to rBta112 from animal JD03 were no longer detected 84 days after infection. Collectively, these findings indicated temporal persistence of IgG responses to rBipA, while three of four animals generated prolonged IgG responses to rBta112.

FIG 6.

Evaluation of temporal serological responses to rBipA (circles) and rBta112 (squares) by ELISA. (A to C) Serum samples for animals JB23, JB60, and IN57 were evaluated prior to infection and at time points in the following 84 days after infection. (D) Serum samples from JD03 were collected prior to and 7, 27, and 43 days after feeding the animal with uninfected ticks. The animal was then infected with B. turicatae by tick bite, and serum samples were collected at days 58, 72, 100, and 142 of the study (infected ticks). Preinfection serum samples from each animal were used to establish a statistically significant threshold (P ≤ 0.003) for rBipA (dashed line) and rBta112 (dotted line).

DISCUSSION

In this study, we identified cytokine profiles associated with pathogenesis and characterized differences in antibody responses of NHPs and mice infected with B. turicatae. Evaluating cytokine production from B. turicatae-stimulated PBMCs identified mediators involved in disease manifestation. B. turicatae induced significant increases in TNF-α, IL-1β, sCD40, and IL-23 compared to levels found in medium controls. The observed TNF-α response has been linked to spirochete lipoprotein-induced Jarisch-Herxheimer reactions (14). Interestingly, B. turicatae also induced statistically significant higher levels of IL-1β than those found in cells incubated with Borrelia burgdorferi, the LD pathogen. IL-1β and TNF-α are known to play a primary role in triggering miscarriage and preterm labor in rhesus macaques (46) and in human patients (47, 48). If significant quantities of RF spirochetes cross the placenta, such a response could be induced in utero, and this pathogenic mechanism should be further evaluated. We did not detect elevated levels of these key inflammatory mediators directly in serum of infected monkeys; however, the response in PBMCs was detected between 12 and 24 h poststimulation. We therefore suspect that we missed the height of the inflammatory response with evaluation, commencing after 14 days of infection. In mice infected with B. hermsii, the plasma levels of IFN were elevated at the height of spirochetemia, whereas IL-1β was detected after clearance of the infection (49). Directly comparable experiments in mice and primates would be of benefit, but consistent detection of these inflammatory mediators in blood cells exposed to RF spirochetes indicates that they are likely important for pathogenesis.

Our findings suggest unique characteristics in antibody responses generated to B. turicatae antigens between mammalian species. BipA is known to be immunogenic in mice (24, 25), but previous work screening serum samples from a small cohort of mice naturally infected with B. turicatae by tick bite indicated that BrpA was not antigenic (50). The serological responses from the 10 mice that were evaluated in the present study supported previous findings with BrpA. Interestingly, one NHP produced a detectable response against rBrpA. Furthermore, while only two mice seroconverted to rBta112, the protein was antigenic in all four NHPs. While more animals are needed to definitively determine differences in antibody responses between mice and NHPs, these findings suggested that the immune responses between the two mammalian species were dissimilar. Future work should evaluate NHPs as a model for antigen discovery and vaccine development.

B cells drive the immune effort to control infection with RF spirochetes, and distinct subsets with roles in immunity have been delineated (19, 20, 51). Mature B cells can be divided into follicular B cells, present in the lymphoid follicles; marginal zone (MZ) B cells, located in the marginal sinus of the spleen; and B1 cells, predominantly found in the mouse peritoneum. These are subdivided into B1a and B1b cells. B1 and MZ B cells are known to engage in the T-cell-independent antibody response. In contrast to LD, which induces an expansion of MZ B cells upon infection, RF spirochetes induce a loss of MZ B cells (21). This may reflect the long-term presence of LD spirochetes in the spleen versus the periodic blood-borne expansion of RF spirochetes and/or the differential responses to antigens. The importance of the B1b cell component of the T-independent response to RF spirochetes was demonstrated by transfer of B1b lymphocytes from convalescent mice to Rag 1^–/– mice (lacking mature B and T lymphocytes). This subpopulation conferred protection that consisted of a specific IgM response which occurred when mice were challenged 60 days after the reconstitution, indicating that this population alone could confer memory and afford protection (20). Importantly, the identical counterpart of this particular B-cell subset has not been identified in humans (52), so it remains to be seen if the same mechanism to control infection occurs in RF patients. Our study shows a precipitous drop in all of the major B-cell subsets within the peripheral blood of RF spirochete-infected NHP during the height of bacteremia (2 weeks p.i.). While we did not examine lymph node populations, we surmise that the steep decline in peripheral B cells was met with migration to lymphoid organs. By day 70, no specific B-cell subset emerged at an increased frequency over the others. In addition, the specific antibody responses to recombinant proteins were of IgG isotype. Our attempts at screening IgM responses did not produce clear and specific binding to either recombinant proteins or B. turicatae lysates. This suggests that B-cell responses in primates may rely more on T-cell-dependent IgG subclass responses.

RF borreliosis is a major burden to maternal and fetal health, especially in resource-poor areas, and novel intervention strategies are needed. According to the World Health Organization, every year 45% of all deaths in children under 5 years are among newborn infants in their first 28 days of life or the neonatal period, and 25% of neonatal deaths result from infections (53). The major threat of RF borreliosis caused by both Old and New World species is pregnancy complications occurring during the perinatal period (∼20 weeks after gestation to 1 to 4 weeks after birth) (5, 9, 54–60). The use of rhesus macaques as an animal model resulted in the identification of novel antigens (BrpA and Bta112) and confirmed the immunogenicity of BipA. Future studies will determine whether these antigens offer potential targets for human vaccination or in diagnosis of RF. We will also expand on the developed NHP model and focus on understanding immunopathology of infected pregnant macaques to identify cytokines in amniotic fluid, which may reveal a mechanism of perinatal effects associated with RF infection.

MATERIALS AND METHODS

Ethics statement.

Practices in the housing and care of NHPs and mice conformed to the regulations and standards of the Public Health Service Policy on Humane Care and Use of Laboratory Animals (66). The Tulane National Primate Research Center (TNPRC) and Baylor College of Medicine (BCM) are fully accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care-International. The Institutional Animal Care and Use Committees at the TNPRC and BCM approved all animal-related protocols, including the infection and sample collection from NHPs and mice. All animal procedures were overseen by veterinarians and their staff.

B. turicatae strains used and animal infections by tick bite.

B. turicatae strains used in this study were 91E135, Florida canine Borrelia (FCB), 99PE-1807, Texas canine Borrelia 1 (TCB1), and TCB2 (61). Tick transmission studies were previously reported using a colony of Ornithodoros turicata that originated from Kansas (22). Briefly, four male Indian rhesus macaques (JB23, JB60, IN57, and JD03) 2.02 to 2.85 years of age were used. Animals were sedated with 5 to 8 mg/kg body weight Telazol by intramuscular injection, and ten third-stage nymphal ticks infected with B. turicatae were fed on each NHP (22). JD03 was initially fed upon by 10 uninfected ticks and monitored for 42 days as a control for tick-specific responses. This animal was subsequently fed upon by infected ticks.

Murine infection by tick bite was performed as previously described (25). Eight to ten infected third-stage nymphal O. turicata ticks were fed to repletion on 10 Institute of Cancer Research (ICR) mice, a Swiss derivative maintained at BCM. Infection was assessed by collecting a drop of blood from the animals and evaluating the specimen by dark-field microscopy for the presence of circulating spirochetes. Thirty days after infection by tick bite, the animals were exsanguinated, and serum samples were obtained.

Collection and processing of NHP blood.

To evaluate immune responses from NHPs, both whole blood and clotted blood for serum were collected. Animals were anesthetized (ketamine, 0.1 ml/kg, intramuscular), and blood was collected by venipuncture of the femoral vein into either clot tubes or EDTA tubes (whole blood). Blood for serum samples was collected at day 0 (prior to tick feeding), day 28, day 56, day 70, and day 85, as previously described (22). Whole blood for flow cytometry was collected at day 0, day 14, and day 70. Tubes containing clotted blood were centrifuged at 3,000 rpm for 10 min to obtain serum samples. PBMCs were isolated from whole blood using lymphocyte separation medium (MP Biomedicals) (62). The lymphocyte layer was washed once with sterile phosphate-buffered saline (PBS), resuspended in PBS–2% fetal bovine serum (FBS), and counted. Cells were again pelleted, resuspended in freeze medium (Invitrogen) at ≤1 × 10⁷/ml, and then cryopreserved in liquid nitrogen until staining.

Flow cytometry assay.

Cryopreserved PBMCs were thawed, washed in RPMI 1640 medium, counted with trypan blue exclusion staining, and adjusted to a concentration of 1 × 10⁷ cells/ml in RPMI 1640 medium with 10% FBS. Portions (100 μl) of cells were used for staining with different concentrations of monoclonal antibodies, followed by incubation for 25 min at room temperature, protected from light, as reported earlier (39, 62–64). The cells were further washed two times with 3 ml of flow wash buffer (PBS with 0.1% bovine serum albumin and 7 mM sodium azide) and centrifuged at 1,350 rpm for 7 min. After aspiration of supernatants from cell pellets, the cell pellets were resuspended in 350 μl of 1% paraformaldehyde buffer (in PBS). For antibodies conjugated with tandem dyes, the cell pellets were dissolved in FACS fixation and stabilization buffer (Becton Dickinson). For T-cell phenotyping, CD3-fluorescein isothiocyanate (FITC) (SP34-2; BD Biosciences), CD8-peridinin chlorophyll protein (PerCP) (SK1; BD Biosciences), and CD4-allophycocyanin (APC) (L200; BD Biosciences) were used. For B-cell phenotyping, anti-CD5-phycoerythrin (PE)-Cy5.5 (CD5-5D7; Invitrogen), anti-CD20-extracellular domain (ECD) (B9E9; Beckman Coulter), anti-CD21-APC (B-Ly4; BD Biosciences), anti-CD86-PECy5 (FUN-1; BD Biosciences), anti-CD138-FITC (MI15; BD Biosciences), anti-CD27-FITC (M-T271; BD Biosciences), anti-IgM (G20-127; BD Biosciences), and anti-IgD (purified polyclonal; Southern Biotech) antibodies were used. Anti-CD38 antibody (clone OKT10) was obtained from NIH NHP Reagent Resource. Data were acquired within 24 h of staining using either a BD Fortessa instrument (BD Immunocytometry System) or BD FACSVerse (BD Biosciences) and FACSDiva software (BD Immunocytometry System). For each sample, 50,000 events were collected by gating either on CD3⁺ T cells or CD20⁺ B cells. For B-cell phenotypic analysis, cells were first gated on singlets, followed by lymphocytes, and CD20⁺ B cells and CD20⁻ cells. CD20⁺ B cells were further gated for CD5/CD21/CD86 and CD138 expression. In cases where enough PBMCs were available, flow cytometry was repeated to give duplicate samples. The gating strategy, along with representative results from a single animal, is shown in Fig. S1.

Cytokine/chemokine array.

A portion of PBMCs derived from whole blood were also used for in vitro stimulation with B. turicatae. Cells isolated from blood collected from each animal (which included JD03 following control/uninfected tick feeding) on day 14 after tick feeding were resuspended in RPMI 1640–10% FBS at 1 × 10⁶/ml, and 0.5 ml was added to each well of a 24-well plate. Late-log-phase B. turicatae was diluted to 1 × 10⁷ spirochetes per ml, and 0.5 ml was added to appropriate wells for a 10:1 ratio of spirochetes to cells. Controls included untreated cells, cells incubated with B. burgdorferi, and cells incubated with BSK medium (Sigma). To determine the impact of soluble factors produced by the spirochetes, cells were incubated with 0.22-μm-filtered BSK medium derived from B. turicatae and B. burgdorferi cultures. Cultures were placed in a 37°C, 5% CO₂ incubator. Supernatants were collected at 12 and 24 h and stored at –20°C. Serum samples from days 14, 28, and 41 were also tested by the cytokine/chemokine array. Undiluted samples were analyzed using a Milliplex MAP nonhuman primate cytokine magnetic bead panel (Premixed 23 Plex; Millipore) according to the manufacturer’s instructions. The bead assay was performed by the Pathogen Detection and Quantification Core at the TNPRC and analyzed on a Bioplex 2000 suspension array system (Bio-Rad). Each analyte concentration was calculated by logistic-5PL regression of the standard curve. To determine the statistical significance between the means for two experimental groups, an unpaired, two-tailed Student t test was performed using the GraphPad software QuickCalcs. Differences with P ≤ 0.05 are reported as significant.

Computational analysis of Bta112.

Initially, the Bta112 gene was identified as a gene upregulated by B. turicatae in the tick and at 22°C (tick-like growth conditions) compared to spirochetes isolated from infected murine blood and spirochete grown at 35°C (mammalian-like growth conditions) (43). The protein was evaluated using the Basic Local Alignment Search Tool (BLAST) from NCBI, LipoP1.0, and ScanProsite. The Bta112 gene sequence was evaluated in B. turicatae 91E135, FCB, 99PE-1807, TCB1, and TCB2 through ongoing genome sequencing efforts of these isolates.

Recombinant proteins and generation of rabbit serum to recombinant Bta112 (rBta112).

Recombinant BipA (rBipA) and BrpA (rBrpA) were produced as six-histidine-linked proteins as previously described (25, 50). Recombinant Bta112 was also produced as a six-histidine-linked recombinant protein using the pEXP1-DEST expression vector (Thermo Fisher Scientific, Waltham, MA). The Bta112 gene was amplified by PCR from B. turicatae gDNA with Accuprime Pfx (Thermo Fisher Scientific) without its predicted signal sequence (bp 1 to 69, signal P 3.0 [http://www.cbs.dtu.dk/services/SignalP-3.0/]). The gene’s signal sequence was omitted from amplification using the primers SP/1779/70-1461 (5′-CAAACAAGTTTGTACAAAAATTTCAAAAGTCCAAAAGACGCTG-3′) and ASP/1779/70-1461 (5′-CGTATGGGTAAAGCTTATTACTACTTGCGGTACTATCTGCTG-3′). The amplicon was cloned by using In-Fusion (BD Clontech) into pEXP1-HA-DEST and digested with BsrGI and HindIII to create pEXP1-HA::bta112, and Top10 Escherichia coli samples were transformed. Plasmid DNA was isolated and submitted for sequencing to ensure that errors were not introduced by PCR. Vector NTI 11.0 (Thermo Fisher Scientific) was used to assess the Bta112 gene sequence. rBta112 was produced by transforming E. coli BL21 Star (DE3) cells (Thermo Fisher Scientific) with pEXP1-HA::bta112, and expression was induced with 0.5 mM IPTG (isopropyl-β-d-thiogalactopyranoside). rBta112 was purified by nickel chelate chromatography.

Rabbit anti-rBta112 was produced by Cocalico Biologicals, Inc. Preimmunization serum samples were collected from two rabbits, and the animals were immunized intraperitoneally with 50 μg of rBta112 using complete Freund adjuvant. The animals were immunized three subsequent times at 2-week intervals using Freund incomplete adjuvant. Serum samples were collected and evaluated for specificity to rBta112 and the native protein by immunoblotting.

Surface localization assays, immunoblotting, and densitometry analysis.

To determine the surface localization of Bta112, proteinase K assays and immunoblotting were performed as previously described (50, 65). Moreover, for all immunoblotting assays B. turicatae was grown at 35°C. For proteinase K assays, spirochetes were grown to a density of >5 × 10⁷ cells/ml, pelleted at 1,000 × g for 10 min at room temperature, washed in PBS plus MgCl₂, pelleted again, and resuspended in PBS plus MgCl₂. Spirochetes were incubated with increasing concentrations (5, 50, and 200 μg/ml) of proteinase K (Promega, Madison, WI) for 15 min at room temperature. PBS plus MgCl₂ was used as a vehicle control. Proteinase K was inactivated by boiling samples at 100°C for 10 min. SDS-PAGE and immunoblotting were performed as previously described using Any kD Mini-Protean TGX stain-free precast gels (Bio-Rad, Hercules, CA) (50). Then, 1 μg of recombinant protein or 1 × 10⁷ spirochetes were electrophoresed per lane, and a Trans-Blot cell (Bio-Rad) was used to transfer proteins onto polyvinylidene fluoride membranes. Rabbit, murine, chicken, and NHP serum samples were used to probe immunoblots at a concentration of 1:200, and antibody binding was detected with the appropriate secondary antibody and the ECL Western blotting reagent (VWR, Atlanta, GA). ImageLab (6.0.1) was used to quantify the relative density of FlaB of spirochetes incubated with 5, 50, and 200 μg/ml proteinase K to spirochetes that did not undergo proteinase K treatment.

ELISA.

Immulon 2HB flat-bottom microtiter polystyrene plates (Thermo Fisher, Waltham, MA) were coated with 1 μg/ml of rBipA or rBta112 using 1× coating solution (KPL, Gaithersburg, MD). The plates were washed three times with wash buffer (1× PBS and 0.05% Tween 20) and blocked with diluent (1× PBS, 5% horse serum, 0.05% Tween 20, 0.001% dextran sulfate) overnight at 4°C. Plates were washed again and probed with the NHP serum samples at a 1:100 dilution in diluent, followed by incubation for 1 h at room temperature. Plates were washed and incubated for 1 h at room temperature with peroxidase-labeled goat anti-monkey IgG (KPL) at a 1:4,000 dilution. Plates were washed and incubated with ABTS peroxidase substrate (KPL) for 30 min and read at 405 nm on an Epoch microplate spectrophotometer (BioTek, Winooski, VT). Samples were considered statistically significant if their mean optical density was more than three times the standard deviation above the mean of the pre-tick-challenge sera (P ≤ 0.003).