ABSTRACT
Borrelia burgdorferi is a tick-borne bacterium responsible for approximately 300,000 annual cases of Lyme disease (LD) in the United States, with increasing incidences in other parts of the world. The debilitating nature of LD is mainly attributed to the ability of B. burgdorferi to persist in patients for many years despite strong anti-Borrelia antibody responses. Antimicrobial treatment of persistent infection is challenging. Similar to infection of humans, B. burgdorferi establishes long-term infection in various experimental animal models except for New Zealand White (NZW) rabbits, which clear the spirochete within 4 to 12 weeks. LD spirochetes have a highly evolved antigenic variation vls system, on the lp28-1 plasmid, where gene conversion results in surface expression of the antigenically variable VlsE protein. VlsE is required for B. burgdorferi to establish persistent infection by continually evading otherwise potent antibodies. Since the clearance of B. burgdorferi is mediated by humoral immunity in NZW rabbits, the previously reported results that LD spirochetes lose lp28-1 during rabbit infection could potentially explain the failure of B. burgdorferi to persist. However, the present study unequivocally disproves that previous finding by demonstrating that LD spirochetes retain the vls system. However, despite the vls system being fully functional, the spirochete fails to evade anti-Borrelia antibodies of NZW rabbits. In addition to being protective against homologous and heterologous challenges, the rabbit antibodies significantly ameliorate LD-induced arthritis in persistently infected mice. Overall, the current data indicate that NZW rabbits develop a protective antibody repertoire, whose specificities, once defined, will identify potential candidates for a much-anticipated LD vaccine.
INTRODUCTION
A variety of pathogenic organisms are equipped with highly evolved antigenic variation mechanisms that constantly allow the microbial invaders to escape an otherwise efficacious antibody response in the infected host (1–11). Borrelia burgdorferi, a causative agent of Lyme disease (LD) (also known as Lyme borreliosis), is no exception. This extracellular bacterium is responsible for 30,000 confirmed cases of human LD each year in the United States alone, although the actual incidence is thought to be 10 times higher (12). Climate change is implicated in the spread of the vector of the LD spirochete, with increasing incidence rates of the disease (13, 14). The debilitating nature of this multisystemic disease is substantially attributable to the ability of B. burgdorferi spirochetes to establish a persistent state of infection. If an early diagnosis is missed, mainly due to transient flu-like symptoms, chronic disease follows, with a variety of symptoms, including fatigue, musculoskeletal pain, arthritis, carditis, peripheral neuropathy, meningitis, encephalitis, cranial neuritis, and/or cognitive dysfunction (15). Unfortunately, antimicrobial treatment of persistent (chronic) infection is challenging, and more importantly, to date, no vaccine for humans is available (16–21).
In the mammalian host, the long-term survival of B. burgdorferi, despite robust antibody responses, is mainly attributed to the variable major protein (VMP)-like sequence (vls) locus (22). The vls locus, which is well characterized in the B. burgdorferi B31 strain, is located near the right telomere end of a 28-kb linear plasmid (lp28-1) and is composed of the vlsE gene and 15 noncoding vls cassettes (474 to 594 bp long). The vlsE gene contains two constant regions that flank one central highly variable region. Because this vlsE central region shares 90.0 to 96.1% nucleotide identity with each silent cassette (5), unprogrammed events of gene conversion take place between each cassette and the vlsE cassette-like region. Importantly, vlsE recombination events are identified in mice by as early as 4 days postinfection, while they are undetectable in vitro or in ticks (22–28). The end product of the vls locus is the expression, on the spirochetal surface, of the highly antigenically variable protein VlsE. This variable VlsE protein is absolutely required for B. burgdorferi to continually evade adaptive antibody responses in order for spirochetes to establish a long-term (lifelong) infection in humans or other mammalian hosts (e.g., mice) (29–38). It has been consistently demonstrated that B. burgdorferi strains lacking the vls locus are rapidly cleared by mouse anti-Borrelia antibodies (36–38).
In contrast to humans (39–44) and numerous animal models (45–60), B. burgdorferi fails to establish a lifelong infection in New Zealand White (NZW) rabbits. NZW rabbits are able to completely clear an active infection by the wild-type B31 strain within 4 to 8 weeks on average (61, 62). The possibility that clearance in NZW rabbits is due to a failure of the vls locus to undergo recombination has been discarded by previous work (62). It has been demonstrated that vlsE recombination could be detected by as early as 2 weeks postinfection and that the average number of vlsE sequence changes in NZW rabbits was comparable to or even higher than those in mice at week 4 postinfection (62). However, that study also showed that 50% of wild-type spirochetes recovered from rabbit skin were devoid of the vls locus-carrying plasmid, suggesting that it was the spontaneous loss of the vls locus that accounted for the failure of B. burgdorferi to establish a long-term infection in NZW rabbits (62). However, the fact that the other 50% of skin isolates retained the plasmid but were still cleared has led us to reexamine this previous finding by directly testing the rate of retention of lp28-1 by skin isolates via colony PCR. The results demonstrated that all the examined rabbit skin isolates of B. burgdorferi uniformly retained the lp28-1 plasmid, indicating that the clearance of B. burgdorferi by NZW rabbits is not due to a loss of lp28-1. With this new finding, we set to define a role of VlsE for B. burgdorferi in NZW rabbits, the subject of this study.
The new data show that despite vlsE upregulation in NZW rabbits, B. burgdorferi establishes only a transient infection. The results demonstrate that host-adapted B. burgdorferi spirochetes, which are otherwise highly immune evasive in the mouse host (37, 63), are susceptible to anti-B. burgdorferi antibodies of NZW rabbits (referred to here as rabbit antibodies). In passively immunized mice, the rabbit antibodies completely abrogate the establishment of infection by the highly immune-evasive (host-adapted) wild type. Additionally, the rabbit antibodies protect mice from an in vitro-grown heterologous strain of B. burgdorferi. The data also show that the protective efficacy of rabbit antibodies is predominantly complement dependent. Finally, the study demonstrates that rabbit antibodies passively transferred to mice with an ongoing infection significantly reduce the pathology of B. burgdorferi-induced arthritis.
RESULTS
VlsE capacitates B. burgdorferi to establish only a transient infection in NZW rabbits.To verify the previous finding that a portion of the B. burgdorferi population spontaneously loses the vls locus-carrying plasmid, lp28-1, and to examine the role of VlsE for rabbit infection (62), groups A and B (4 animals per group) of NZW rabbits were challenged with the in vitro-grown B. burgdorferi B31-A3 strain (referred to here as B31) and B31-A3/lp28-1::kan Δvls (ΔVlsE), respectively (see Table S1 in the supplemental material). At weeks 1, 2, 3, and 4 postinoculation, skin biopsy specimens sampled around the inoculation sites were cultured in liquid Barbour-Stoenner-Kelly medium with 6% rabbit sera (referred to here as BSK-II) and incubated at 35°C under 2% CO2. The culture results revealed that the B31 strain was consistently detected at weeks 1, 2, and 3 in all the challenged rabbits (Table 1). By week 4, however, skin biopsy specimens from 2 out of 4 B31-infected rabbits were culture negative, indicating the beginning of clearance of wild-type infection in NZW rabbits. This finding was consistent with previous data, which demonstrated that by this time point, 3 out of 8 NZW rabbits cleared B31-induced infection (62). Thus, the wild-type B31 strain was able to establish an active infection that was cleared by 50% of the NZW rabbits by week 4 postinoculation.
Culture results for skin biopsy specimens sampled weekly from NZW rabbits challenged with in vitro-grown B31 or ΔVlsE
As opposed to B31, by week 1, skin biopsy specimens of 2 rabbits (animals B3 and B4) out of 4 rabbits were already culture negative for ΔVlsE spirochetes, indicating very early clearance by 50% of the ΔVlsE-inoculated rabbits. Thus, the presence of ΔVlsE infection at week 1 was directly confirmed, via culture, only in rabbits B1 and B2. However, despite being culture negative at week 1, rabbits B3 and B4 were still able to develop pronounced erythema migrans (as did rabbits B1 and B2) around each inoculation site at week 1 postchallenge, indicating that the ΔVlsE clone had actively infected these two rabbits as well (61, 64) (Fig. S1). A previous study showed that when NZW rabbits were inoculated with the avirulent B. burgdorferi B31 strain, no erythema migrans developed (61). Importantly, the infectivity of the ΔVlsE inoculum was verified using five C3SnSmn.CB17-Prkdcscid/J SCID (referred to here as SCID) mice, whose blood and ear skin tissues collected at days 7 and 21 postchallenge, respectively, were uniformly culture positive for the ΔVlsE clone (data not shown). Since rabbit B4 was added to the group later and, therefore, the inoculation was performed separately, successful infection of this rabbit with the ΔVlsE clone was also verified by culture-positive results for skin biopsy specimens taken at day 3 postinoculation (data not shown). All the subsequent skin biopsy specimens sampled at weeks 2, 3, and 4 from rabbits B1, B2, and B3 (rabbit B4 was sacrificed at day 7) were culture negative, which demonstrated an earlier clearance of the VlsE-deficient clone than of its isogenic wild-type counterpart (P < 0.05) (Table 1). Together, the results indicated that in NZW rabbits, the lack of the vls locus resulted in a very early clearance of B. burgdorferi.
To verify the previously reported loss of lp28-1 by the wild type, a total of 200 B31 isolates were individually PCR screened for the presence of lp28-1. The results showed that, in contrast to the previous finding, where 25 out of 50 isolates recovered from skin of B31-infected rabbits were PCR negative for lp28-1 (62), every single isolate of B31 contained the plasmid (Table 2). The discrepancy between the previous and current findings could well be accounted for by the different approaches taken. In the previous work, prior to the PCR screen, the 50 B31 isolates were first expanded in BSK-II in order to obtain a sufficient amount of DNA (62). However, it was consistently shown that during in vitro propagation, B. burgdorferi rapidly lost various plasmids, including lp28-1 (22, 65–69). Thus, in the present study, this propagation step was purposely avoided by directly screening all the B31 isolates for the presence of lp28-1 via colony PCR. Together, the present data unequivocally demonstrated that B. burgdorferi does not lose the vls-carrying lp28-1 plasmid during rabbit infection.
Rates of retention of the lp28-1 plasmid by the B31 clone isolated from infected NZW rabbits
The rabbit antibodies are potent in preventing infection by VlsE-expressing B. burgdorferi.Our results showed that, despite plasmid retention, lp28-1-carrying spirochetes were still cleared in NZW rabbits. It was possible that the failure of wild-type B. burgdorferi to establish a long-term infection in NZW rabbits was due to the lack of VlsE expression. To examine this, quantitative reverse transcription-PCR (qRT-PCR) was used to detect vlsE transcription in B31-infected rabbits at days 3, 7, 14, 21, and 28 postchallenge. The failure to detect any vlsE transcripts at days 3 and 28 (data not shown) was suggestive of very low numbers of spirochetes in the respective skin biopsy specimens. vlsE transcription was consistently detected at days 7, 14, and 21, indicating that vlsE was upregulated in NZW rabbits (Fig. 1).
vlsE transcription in rabbit skin-residing B. burgdorferi. vlsE transcript levels were upregulated, as determined by qRT-PCR, in the skin of B31-infected NZW rabbits at days 7, 14, and 21 postinfection (pi). The flaB gene of B. burgdorferi B31 served as the endogenous control to normalize the vlsE expression level.
It has been consistently demonstrated that, mainly due to the VlsE system, B. burgdorferi establishes a long-term infection in mice despite their strong antibody responses (22, 24, 29–38, 70). Thus, to examine the potency of the rabbit antibodies against B. burgdorferi in the mouse host, five C3H mice (group IA) were passively immunized with anti-B31 immune sera previously collected and pooled from three B31-infected NZW rabbits at day 28 postinoculation. As a control, C3H mice of groups IB and IC (3 animals per group) received saline and preimmune sera, respectively. Immediately after the respective treatments, all the mice were challenged with the wild-type B31 clone. Since the VlsE expression level was previously shown to be significantly higher in the spirochetes that resided in mouse tissues (e.g., skin) than in in vitro-grown B. burgdorferi (71, 72), host-adapted B31 (ha-B31) (ear skin from infected SCID mice) was used for the challenge. Thus, it was anticipated that immediately upon challenge, spirochetes were armored with abundant VlsE on their surface to defend themselves against the passively transferred antibodies. It was previously shown that only when in a host (mouse skin)-adapted state did wild-type B. burgdorferi gain the capacity to resist anti-B31 mouse antibodies, which were otherwise sufficiently potent to prevent in vitro-grown B31 or the ha-ΔVlsE clone (37).
At days 7 and 21 postchallenge, blood and other tissues, specifically, bladder, heart, ear skin, and tibiotarsal joint (referred to here as other tissues), respectively, were harvested and cultured in BSK-II to evaluate the outcome of challenge. The results showed that 5 out of 5 mice that had received the rabbit antibodies were completely protected from the VlsE-expressing wild type. In contrast, all the control mice showed culture-detectable spirochetemia at day 7 and disseminated infection at day 21 postchallenge, confirming the full infectivity of the ha-B31 clone. Together, the data demonstrated that the rabbit antibodies had the capacity to efficaciously clear spirochetes despite their fully functional vls system in the mouse (natural) host.
Protective efficacy of the rabbit antibodies against B. burgdorferi is mainly complement dependent.To test whether the observed protective efficacy of the rabbit antibodies against B. burgdorferi depends on the complement, 12 C3H mice (group IIA) were treated with heat-inactivated immune sera and then immediately challenged with the ha-B31 clone. These sera were previously collected from B31-infected rabbits at day 28 postinfection. Blood and the other tissues harvested at days 7 and 21 postchallenge, respectively, were cultured in BSK-II. The results demonstrated that the heat-inactivated immune sera prevented culture-detectable spirochetemia in 12 out of 12 mice but failed to block disseminated infection in 9 out of these 12 animals (Table 3). Importantly, the infectivity of host-adapted wild-type spirochetes was verified with group IIB mice: 3 out of 3 control mice exhibited spirochetemia and disseminated infection at days 7 and 21 postchallenge, respectively. Thus, based on the results for groups IA and IIA, complement inactivation significantly reduced the protective efficacy of the rabbit anti-B31 antibodies (5/5 versus 3/12; P = 0.0090) (Table 3), indicating that the borreliacidal function of the rabbit antibodies is predominantly complement dependent.
Protective efficacy of anti-Borrelia antibodies of NZW rabbits against homologous B. burgdorferi
To further explore a role of the rabbit complement for the potency of the rabbit anti-B31 antibodies, 9 C3H mice (group IIIA) were treated with rabbit heat-inactivated anti-B31 sera and immediately challenged with the in vitro-grown B31 clone. As a control, 3 C3H mice (group IIIB) and 5 C3H mice (group IIIC) were treated with saline and preimmune sera, respectively, prior to the challenge. All the control animals exhibited culture-detectable spirochetemia and disseminated infection at days 7 and 21 postchallenge, respectively. Thus, the control results verified the full infectivity of the B31 inoculum and, importantly, demonstrated the inability of rabbit preimmune sera (intact complement, which has not been activated by infection) to block the needle B. burgdorferi (Table 3). In contrast, the culture results for blood and the other tissues collected from group IIIA at days 7 and 21 postchallenge, respectively, showed that 9 out of 9 mice were completely protected from challenge with in vitro-grown B31 (P < 0.0001) (Table 3). Together, the results indicated that the rabbit anti-B31 antibodies were fully protective against in vitro-grown B31 spirochetes even in the absence of complement.
Potential contribution of VlsE to B. burgdorferi evasion of complement-independent antibody killing in NZW rabbits.The finding that, in contrast to the needle B. burgdorferi, host-adapted spirochetes were more resistant to complement-independent antibody killing prompted us to examine a potential role of VlsE in this evasion. The fact that the VlsE expression level was shown to be significantly higher in the mouse host than in vitro (72) could explain the higher susceptibility of the needle wild type to complement-independent antibody killing. To test this, 5 and 3 SCID mice were treated with heat-inactivated anti-ΔVlsE sera (group IVA) and saline (group IVB [control]), respectively, and then challenged with the ha-ΔVlsE clone. The anti-ΔVlsE sera were previously collected from ΔVlsE-infected rabbits at day 28 postinfection. Blood and the other tissues harvested at days 7 and 21 postchallenge, respectively, were cultured in BSK-II. As opposed to the control mice, the culture results for the experimental group showed that anti-ΔVlsE antibodies completely prevented, in a complement-independent manner, both spirochetemia and disseminated infection by host-adapted VlsE-deficient B. burgdorferi (P = 0.0179) (Table 3). This is in contrast to the culture results for group IIA, where heat-inactivated anti-B31 sera were protective against the host-adapted VlsE-expressing wild type for only 3 out of 12 mice (P = 0.0090). Logically, the observed difference could potentially be accounted for by lower titers of antibodies to surface antigens other than VlsE in the anti-B31 sera than in the anti-ΔVlsE sera. However, the levels of total IgG were not significantly different between the two types of immune sera (P > 0.05) (Fig. S2). Moreover, Western blot analysis of anti-B31 and anti-ΔVlsE immune sera collected from B31-infected and ΔVlsE-infected rabbits, respectively, at days 7, 14, 21, and 28 postchallenge revealed no noticeable difference in their reactivity to whole-cell lysates of the B31 clone (Fig. S3). Together, the data suggested that in NZW rabbits, VlsE contributed to the evasion of B. burgdorferi of complement-independent antibody killing. These results could potentially explain why, in the NZW rabbits, the VlsE-deficient clone was cleared so early compared to VlsE-competent B. burgdorferi B31.
The rabbit antibodies are cross-protective against heterologous B. burgdorferi.To investigate whether the rabbit antibodies were cross-protective against heterologous B. burgdorferi, C3H mice were treated with anti-B31 sera and then challenged with the host-adapted (group AI) or in vitro-grown (group B1) 297 strain (3 mice per group). The infectivity of ha-297 and in vitro-grown 297 was confirmed with the respective saline-treated control groups (groups AII and BII), whose tissues were consistently culture positive for the 297 strain (3 animals per group) (Table 4). The culture results for blood and the other tissues of group AI demonstrated that the rabbit anti-B31 sera did not prevent infection by ha-297 in any of the three mice. In contrast, both culture-detectable spirochetemia and disseminated infection by needle 297 were successfully blocked in 5 out of 5 mice (P < 0.05). Importantly, the other 5 out of 5 mice treated with preimmune sera (group BIII) were successfully infected with in vitro-grown 297, demonstrating the inability of the rabbit complement, which has not been activated by infection, to block heterologous needle B. burgdorferi. Together, the data showed that the rabbit anti-B31 antibodies were cross-protective against heterologous in vitro-grown B. burgdorferi, which is consistent with a previous study that demonstrated cross-protection by the rabbit antibodies against in vitro-grown spirochetes in passively immunized hamsters (73).
Protective efficacy of anti-Borrelia antibodies of NZW rabbits against heterologous B. burgdorferi
To examine whether the observed cross-protection by the anti-B31 antibodies was dependent on the rabbit complement, 5 C3H mice (group CI) were treated with heat-inactivated anti-B31 sera and then challenged with in vitro-grown 297. The infectivity of the 297 inoculum was fully confirmed by culture-positive tissues of 3 control animals (group CII) (Table 4). The culture results for blood and the other tissues from group CI mice showed that upon complement depletion via heat inactivation, the rabbit anti-B31 antibodies were still capable of preventing heterologous infection in 3 out of 5 mice (Table 4). Thus, the data indicated that antibody-mediated killing of heterologous needle B. burgdorferi was partly complement dependent.
Finally, the fact that the ha-297 clone was resistant to the rabbit anti-B31 antibodies brought up the question of whether, in NZW rabbits, this heterologous strain of B. burgdorferi had the capacity to persist longer than the B31 clone. To test this, three NZW rabbits were challenged with in vitro-grown 297 at a dose and via a route identical to those for the B31 challenge. The three rabbits all became infected, as consistently determined by culture-positive skin biopsy specimens from weeks 1, 2, 3, and 4 postinoculation (Table S2). By week 5, however, 2 out of 3 rabbits cleared strain 297 infection. Skin biopsy specimens from later time points (weeks 6, 7, and 8) were culture negative for the three rabbits, demonstrating the complete clearance of B. burgdorferi infection. Thus, the results were consistent with findings obtained from the B31-infected rabbits in this and previous studies (61, 62) and once again confirmed that in NZW rabbits, B. burgdorferi failed to persist long-term.
The rabbit antibodies significantly reduce the pathology of B. burgdorferi-induced arthritis.To examine whether the rabbit antibodies were sufficiently potent to exert any therapeutic effect, the experimental design involved nine C3H mice that were infected with in vitro-grown B31. Infection of each mouse was confirmed with culture-positive blood sampled at day 7 postinoculation (Table S3). Next, at days 14, 18, 22, and 26, 4 animals were retro-orbitally treated with the rabbit antibodies (anti-B31 sera), and the other 5 mice were left untreated (control group). At day 45 postchallenge, bladder, ear skin, heart, and tibiotarsal joint tissues were harvested from each animal and cultured in BSK-II. The results showed that most of the tissues from all 9 mice were culture positive for the B31 clone, indicating that the rabbit antibodies did not have the capacity to abrogate ongoing infection by B. burgdorferi (Table S3).
In order to compare infection-induced pathological changes between the two groups, heart and two tibiotarsal joints of each C3H mouse were subjected to histopathological analysis. To ensure rigor and reproducibility, the assessment was performed in a blind manner and independently by two board-certified pathologists. The results demonstrated that there was a noticeable but statistically insignificant difference in the histopathological scores of heart tissues between the treated animals (score, 1.00 ± 1.35) and control animals (1.80 ± 1.04) (P = 0.1162) (Fig. S4). However, for the joints, a statistically significant difference was detected for synovial hyperplasia and overall inflammation scores between the treated and control animals (P < 0.05) (Table 5; Fig. S5). Interestingly, there were two outliers, one joint for each group, whose removal from the analysis also resulted in statistically significant differences for the exudate within the joint and overall scores between the two groups (P < 0.05 by multiple t tests) (data not shown).
Therapeutic effect of anti-Borrelia antibodies of NZW rabbits on B. burgdorferi-infected mice as determined by histopathology of tibiotarsal joints
Despite the reduced pathology detected in the C3H mice, it was possible that the potency of the rabbit antibodies was reduced by mouse anti-rabbit immune responses in these immunocompetent animals. To overcome this potential caveat, a follow-up experiment was performed. This time, the experimental design involved 12 immunodeficient (SCID) mice that lacked adaptive immunity (74–77). All the SCID mice were first infected with the in vitro-grown B31 clone. Infection was confirmed with culture-positive blood sampled from each mouse at day 7 postchallenge. At days 14, 18, 22, and 26, six SCID mice were retro-orbitally treated with the rabbit antibodies (anti-B31 immune sera), and the other six animals were left untreated (control group). Hearts and right tibiotarsal joints from the 12 SCID mice were then harvested at day 45 postchallenge and subjected to histopathological analysis. In addition, left tibiotarsal joints and ear skin from all 12 SCID mice were analyzed by quantitative PCR (qPCR) in order to compare spirochetal loads between the two groups.
The histopathological analysis, again performed in a blind manner and independently by the two pathologists, demonstrated noticeable but statistically insignificant differences for heart pathology between the treated (score, 0.33 ± 0.41) and control (0.92 ± 0.38) groups (P = 0.0583) (Fig. S6). In contrast, joint pathology was significantly reduced in all treated mice for the following histopathological changes: synovial hyperplasia, superficial inflammation of bone, and overall inflammation (P < 0.05) (Table 5; Fig. 2). The qPCR results showed that there were noticeable but statistically insignificant differences in spirochetal loads of the joint tissues between the two groups (P > 0.05) (Fig. 3). In contrast, a significant reduction in spirochetal numbers was observed between the ear skin tissues of treated animals and those of control animals (P = 0.0170). Together, the results of two independent experiments demonstrated that the rabbit antibodies had the capacity to significantly reduce the pathology of B. burgdorferi-induced arthritis.
Treatment of actively B. burgdorferi-infected SCID mice with rabbit anti-Borrelia antibodies results in significant reduction of infection-induced joint pathology. Shown are representative histological pictures of tibiotarsal joints from treated (A, C, E, and G) and control (B, D, F, and H) groups. (A) Only a few inflammatory cells are observed. (B) Severe inflammation is indicated with the white arrow. (C) No synovial changes are observed. (D) Moderate synovial hyperplasia is present. (E) No bone inflammation or resorption is observed. (F) Moderate to severe superficial inflammation and resorption of bone are indicated by the black arrow. (G) Inflammation (asterisks) is present in only one area. (H) Multifocal inflammation is observed. Shown is H&E staining. Bar, 100 μm. The histopathological scores of synovial hyperplasia, superficial inflammation and/or resorption of bone, and overall inflammation, including the overall scores, were significantly different between the two groups (P < 0.05) (Table 5).
Spirochete burdens in tibiotarsal joints (A) and ear skin tissues (B) of B. burgdorferi-infected SCID mice that were treated with anti-Borrelia antibodies of NZW rabbits or remained untreated. A total of 12 SCID mice were first inoculated with the B31 clone, and next, at days 14, 18, 22, and 26 postinfection, 6 SCID mice were treated with anti-B31 immune sera and the other 6 control animals remained untreated. To compare the effects of the rabbit antibodies on spirochetal loads between the two groups, the left tibiotarsal joint and ear skin tissues were collected from each mouse at day 45 postinfection. DNA was extracted from each sample and subjected to a quantitative PCR (qPCR) assay to quantify the actB DNA copies relative to the recA gene. All standard dilutions and DNA samples were amplified in triplicate. The number of spirochetes in each DNA sample was calculated as the ratio of recA DNA copies per copy of the actB gene. A significant reduction in spirochetal numbers was observed only between the ear skin tissues of treated and control animals (P = 0.0170).
Comparison of developed anti-VlsE antibody repertoires in NZW rabbits and mice.In an attempt to identify any difference in anti-VlsE antibody repertoires between NZW rabbits and mice, an approach that involved random peptide phage display (Ph.D.) libraries coupled with next-generation sequencing (RPPDL/NGS), followed by our previously developed computational algorithms, was used (78). As a result of RPPDL/NGS application, global mimotope repertoires of the rabbit day 14 (nonprotective) and day 28 (protective) and mouse day 28 (nonprotective) sera were identified. Next, all the repertoires were mapped, via BLASTP, to the primary structure of VlsE of the B. burgdorferi B31 strain (referred to here as B31-VlsE). Two pooled preimmune sera taken from the same three C3H mice and three rabbits prior to their challenge, respectively, were used as two background controls. The overall results showed that there were no major differences detected between the reactivities of rabbit and mouse sera to the primary structure of VlsE, with a few noteworthy exceptions (Fig. 4). First, there was a pronounced reactivity of 3 rabbit day 28 and 2 out of 3 day 14 sera to the VlsE24-31 epitope of the N-terminal invariable domain. In contrast, the three mouse day 28 sera were not reactive to this epitope. Interestingly, the recognition of VlsE24-31 by the rabbit antibodies was congruent with previous data derived from human sera (79). The analyses of microarrays, which contained synthesized overlapping peptides of B31-VlsE, demonstrated that the VlsE21-31 epitope was targeted by most LD patient sera (79). Second, in contrast to the rabbit day 28 antibodies, 2 out of 3 mouse sera intensely reacted with two epitopes: VlsE199-204 of invariant region 3 (IR3) and VlsE350-356 of the C-terminal invariant domain (Fig. 4). The latter reactivity was fully supported by the previously observed reactivity of day 28 and day 70 antibodies from C3H mice that were persistently infected with heterologous B. burgdorferi strain 297 (78). Moreover, the overall reactivity to the C-terminal invariant domain was also common for LD patient sera (79). Thus, the current data fully support the previous conclusion that the C-terminal invariable domain of VlsE is highly immunodominant (80). Interestingly, the previously observed reactivity of antibodies from 297-infected C3H mice to 297-VlsE IR1 (78) was not shared by any of the anti-B31 mouse or rabbit sera. The lack of reactivity to IR1 could be explained by overall low degrees of identity (46%) and similarity (53%) between B31-VlsE and 297-VlsE (5). Importantly, the anti-B31 rabbit and mouse antibodies were also expectedly reactive to IR6, the invariant region that was uniformly demonstrated to be highly immunogenic in humans, monkeys, and mice (81–84). Consistent with our previous study (78), the present data demonstrated no and little reactivity to the conserved VlsE regions IR2 and IR4, respectively. This is in contrast to other work where B. burgdorferi-infected C3H mice induced strong antibodies to IR2 and IR4 (83). In sum, the present study did not identify any anti-VlsE antibody reactivity that is unique to the protective rabbit sera, suggesting that any potentially protective invariant linear epitopes of VlsE are not accessible to the rabbit antibodies. This is consistent with previous findings where the well-conserved regions of VlsE (e.g., IR6) were shown to be inaccessible to antibodies on intact spirochetes (82, 84, 85). Overall, the present data suggest that the rabbit anti-VlsE antibodies do not play a role in the sterilizing immunity of NZW rabbits to the LD pathogen.
Epitope mapping of B31-VlsE. The VlsE primary structure of the B. burgdorferi B31 strain is composed of two invariable domains and a central variable domain demarcated by two direct repeats (light green). The variable domain has six invariable (gray) and six variable (pink) regions (109). Heat maps were generated from the predicted reactivity of mouse and rabbit anti-B31 antibodies to the primary structure of B31-VlsE. Immune sera were harvested from three B31-infected C3H mice at day 28 postinfection (M1_28, M2_28, and M3_28) and three B31-infected rabbits at day 14 (R1_14, R2_14, and R3_14) and day 28 (R1_28, R2_28, and R3_28) postinfection. Mouse and rabbit preimmune sera were collected and pooled in equal amounts from the three mice (PreImmMouse) and three rabbits (PreImmRabbit), respectively, prior to B31 challenge.
DISCUSSION
An intricate role of VlsE in NZW rabbits.In humans with chronic LD, B. burgdorferi persists for years (39–44) despite a strong immune response to the infection (70, 86–88). Numerous experimental models of various animals, such as hamster (45, 46), dog (47–51), gerbil (52), guinea pig (53), monkey (54, 55), mouse (56, 57), pony (58, 59), and rat (60), show that B. burgdorferi establishes persistent (chronic) infection in these mammals. The exception is NZW rabbits, which are the only known animal model where B. burgdorferi fails to sustain a persistent infection (61, 89, 90).
In the mouse model, for B. burgdorferi to constantly avoid a mortal effect of a robust antibody response, LD spirochetes must express antigenically variable VlsE (30–38). Since in NZW rabbits, clearance is mediated by humoral immunity (89, 90), the previously observed spontaneous loss of the vlsE locus-carrying plasmid by B. burgdorferi during rabbit infection (62) could potentially account for the inability of B. burgdorferi to persist in NZW rabbits. However, our current data unequivocally disprove this previous finding by demonstrating that LD spirochetes uniformly retain the vls system and yet are cleared by NZW rabbits. Thus, is the clearance of B. burgdorferi a result of the (partially) impaired vls system in NZW rabbits or a highly efficacious rabbit antibody response?
Numerous pieces of evidence obtained by the previous and present studies indicate that the vls system in rabbit-residing B. burgdorferi is fully functional. First, in NZW rabbit skin, vlsE is highly upregulated and expressed. Our study shows that vlsE is upregulated during rabbit infection (Fig. 1). Consistently, by immunoblot analysis, the previous study revealed that VlsE is the most prominent antigen expressed between day 7 and day 21 postinfection (91). Second, in NZW rabbits, vlsE recombination occurs, and levels of vlsE gene conversion are comparable to or higher than those in mice (62). Third, the present data demonstrate that the vls system, which is highly efficacious for B. burgdorferi to constantly evade antibodies in the mouse (30–38), fails its VlsE-mediated immune evasion function only when immunocompetent mice receive the rabbit antibodies. Thus, as opposed to mice (and humans), whose antibodies are ineffective against VlsE-expressing B. burgdorferi during natural infection (30–38, 92), anti-B. burgdorferi antibodies of NZW rabbits clear the LD pathogen despite its vls system.
Although VlsE does not ensure a long-term infection in NZW rabbits, the present study also suggests that the presence of the vls locus is still required for LD spirochetes to temporarily survive by presumably evading a very early immune response of NZW rabbits. VlsE-deficient B. burgdorferi, which is transiently infectious in NZW rabbits, is cleared much faster than its parental VlsE-competent wild-type strain. The reason for this early clearance is not apparent. Despite our attempts to detect vlsE upregulation at day 3 being unsuccessful, likely due to low numbers of spirochetes, it is still possible that antigenically variable VlsE is expressed much earlier than day 7 (91). If expressed early, it can be speculated that VlsE helps B. burgdorferi avoid T-cell-independent (TI) antibodies, the early responders whose appearance does not require T-cell help (87). In mice, TI antibody responses are critical for the clearance of spirochetemia by relapsing fever Borrelia (93, 94). It is also plausible that VlsE is required for an early stage of spirochetal adaptation to the rabbit host independent of antibody-mediated evasion. Despite the fact that further experiments are needed to more closely examine the exact function of VlsE for early survival in NZW rabbits, the present study concludes that in these animals, VlsE is required for B. burgdorferi to establish only a very transient infection.
It is also possible that the inability of the B. burgdorferi B31 and 297 strains to persist in NZW rabbits is due to their lack of adhesins or other bacterial factors. Previous studies showed that some wild-caught cottontail rabbits were either serologically or culture positive for B. burgdorferi, suggesting that cottontail rabbits may sustain a long-term infection of rabbit-adapted B. burgdorferi strains and therefore play a role in the enzootic cycle of the LD pathogen (95, 96). Regardless, in NZW rabbits, it has been consistently demonstrated that anti-B. burgdorferi immune sera are ultimately responsible for the clearance of B. burgdorferi B31 (89, 90). NZW rabbits completely clear spirochetes from visceral organs by 8 weeks postinfection (61). Similarly, dermal infection by the wild-type B31 strain is cleared on average within 6.7 weeks postinoculation (61).
Potency of rabbit antibodies.The present study demonstrates that rabbit antibodies are remarkably potent against highly immune-evasive B. burgdorferi. Mice treated passively with the rabbit anti-B31 antibodies are completely protected not only from needle but also from highly immune-evasive mouse-adapted B. burgdorferi B31. These results are supported by previous studies which showed that NZW rabbits with infection-derived immunity are fully protected against the wild type introduced in the form of rabbit skin biopsy specimens (rabbit-adapted spirochetes) (97, 98). It should be reemphasized that host adaption prior to challenge allows wild-type spirochetes to be supremely immune evasive, which is mainly due to the abundant expression of VlsE by tissue-residing spirochetes (37, 72). In passively treated SCID mice, mouse anti-B31 antibodies are avoided by wild-type B. burgdorferi only when spirochetes are host adapted (37). Neither the in vitro-grown and tick-derived wild type, whose VlsE expression level is low (23, 25, 71, 72), nor a host-adapted VlsE-deficient isogenic clone has the capacity to resist mouse anti-B31 antibodies (37, 38).
In addition to protection against homologous challenge, rabbit antibodies can be cross-protective against heterologous infection. The present study demonstrates that anti-B31 antibodies protect mice from needle B. burgdorferi 297, the strain that represents one of the three major clades of B. burgdorferi, rRNA gene intergenic spacer type II (RST2) (99, 100). However, the same rabbit antibodies, when passively transferred to mice, fail to block heterologous B. burgdorferi when the 297 strain is host adapted. The overall results are consistent with a previous study which showed that B31 infection-immune NZW rabbits are susceptible to heterologous challenge with host-adapted B. burgdorferi N40 (RST3), Sh-2-82, or 297 (98). Together, the data indicate that upon host adaption, protective epitopes shared between the needle B. burgdorferi B31 and 297 strains disappear, become subdominant, and/or become inaccessible to the rabbit antibodies.
Although numerous previous studies showed that antibody-mediated killing of B. burgdorferi is complement dependent (54, 82, 101–105), none has examined and compared a role of the rabbit complement for protection mediated by infection-induced rabbit antibodies against host-adapted and in vitro-grown B. burgdorferi. As shown by the present findings, the protective efficacy of rabbit antibodies predominantly depends on the presence of the rabbit complement. When the complement is destroyed, the potency of rabbit antibodies against the host-adapted homologous B31 strain is significantly reduced (Table 3). Surprisingly, the efficacy of rabbit antibodies against much less immune-invasive (in vitro-grown) spirochetes does not depend on the complement. All the mice treated with heat-inactivated rabbit anti-B31 immune sera are uniformly protected from the homologous needle B31 strain (Table 3). Interestingly, antibody-mediated killing of heterologous in vitro-grown B. burgdorferi 297 is only partially complement dependent (Table 4).
Finally, in the present study, we demonstrate that in mice with an established B. burgdorferi infection, the rabbit antibodies have the capacity to significantly reduce the pathology of LD arthritis (P < 0.05). Amelioration of arthritis is associated with a noticeable yet statistically insignificant reduction of spirochetal loads in joints of the treated mice compared to the controls. This is in contrast to the ear skin tissues, where the numbers of spirochetes are significantly reduced in the treated animals. The lack of statistical significance for spirochetal reduction in the joint tissues may be partially accounted for by a protracted time line between the final antibody treatment and tissue harvest (19 days), the short half-life of rabbit antibodies (106), and/or the limited amount of applied sera. In addition, the joints may simply be a better protective niche for B. burgdorferi against the rabbit antibodies (107).
Taken together, we propose a model that may explain the key difference between anti-B. burgdorferi antibody responses that developed in mice as an animal model extensively used to study VlsE antigenic variation (or LD patients), whose antibodies fail to fight off infection, and NZW rabbits, whose antibodies efficiently clear LD spirochetes (Fig. 5). Despite the fact that upon infection, mice and LD patients develop robust humoral responses against B. burgdorferi (70, 86–88), their antibodies cannot clear wild-type spirochetes due to antigenically variable VlsE (29–38). It has been repeatedly shown that in infected mice and LD patients, antibodies to various dominant protective (and nonprotective) epitopes of VlsE and other surface antigens are developed (30, 70, 78, 79, 81–84, 86–88, 108–112). During abundant VlsE expression, any dominant protective epitopes of surface antigens become inaccessible, likely via putative VlsE-mediated shielding, to mouse (human) anti-B. burgdorferi antibodies (63, 84, 111). At the same time, as previously proposed (109), protective epitopes of VlsE lateral antigenically invariant surfaces are protected from mouse (human) antibodies by the dense packing of VlsE molecules with each other and/or other surface proteins. As a result, only changeable protective epitopes of VlsE variable regions become a predominant target of borreliacidal anti-B. burgdorferi antibodies (30). However, because of ongoing VlsE-mediated antigenic variation, any newly appearing spirochetes with previously unseen VlsE variants constantly escape otherwise protective mouse (human) anti-VlsE antibodies (29–38). Thus, B. burgdorferi establishes lifelong infection in these mammalian species. In contrast, in NZW rabbits, anti-B. burgdorferi antibodies are (additionally) developed against protective epitopes of surface antigens whose antibody access is not obscured by VlsE (Fig. 5). Thus, as opposed to LD patients and any known LD experimental mammalian model, in NZW rabbits, the development of this protective antibody repertoire results in the complete clearance of LD spirochetes. Currently, as part of a larger study, we are identifying protective epitopes that are specifically recognized by the rabbit antibodies and testing them as potential targets for the development of a long-awaited LD subunit vaccine.
Simplified model that proposes the key difference in protective anti-B. burgdorferi antibodies between mice and NZW rabbits. (A) In infected mice, protective antibodies to various (other than VlsE) surface antigens develop. When VlsE is abundantly expressed in a host tissue, protective epitopes of non-VlsE surface antigens become sterically inaccessible for the respective mouse antibodies. In contrast, protective epitopes of VlsE variable regions become a predominant target of mouse anti-VlsE antibodies. However, due to antigenic variation of VlsE variable regions, newly appearing spirochetes with previously unrecognized VlsE variants will evade any already developed and otherwise protective anti-VlsE antibodies. As a result, in mice, B. burgdorferi establishes lifelong infection. (B) NZW rabbits develop (additionally) antibodies to protective epitopes of invariant surface antigens, whose antibody access is not hindered by VlsE. This results in complete clearance of LD spirochetes in NZW rabbits. Thus, the unique repertoire of NZW rabbit antibodies specifically targets protective epitopes of B. burgdorferi that are not recognized by the mouse humoral immune system.
The results of the passive-transfer experiment, where fully immunocompetent mice gain the capacity to prevent host-adapted B. burgdorferi only when they receive the rabbit antibodies, indicate that in the mouse host, protective epitopes of invariant (non-VlsE) surface antigens are still accessible to the rabbit antibodies in the presence of VlsE (hence, exposed protective epitopes). This in turn means that in NZW rabbits, the profile of exposed protective epitopes is not (drastically) different from those of the respective protective epitopes exposed by B. burgdorferi in the mouse environment. Thus, the main difference between the mouse and rabbit antibodies may lie in the ability of the rabbit antibodies to specifically target those protective epitopes that, despite abundant VlsE expression, remain exposed on the surface of B. burgdorferi, be it in the mouse or rabbit environment. It can be further speculated that the protective epitopes that are recognized by the rabbit antibodies are therefore dominant and subdominant for NZW rabbit and mouse immune systems, respectively.
In summary, the present study provides unequivocal evidence that spirochetes retain the fully functional vls system during rabbit infection. However, despite VlsE-mediated antigenic variation, B. burgdorferi fails to establish persistent infection in NZW rabbits. The data show that the vls locus is required for spirochetal survival during a very early stage of rabbit infection. Later, however, a fully developed anti-Borrelia antibody response clears B. burgdorferi despite its antigenically variable VlsE. The rabbit antibodies are very potent against highly immune-evasive (mouse-adapted) B. burgdorferi. In addition to homologous protection, the rabbit antibodies are also cross-protective against heterologous yet less immune-evasive (in vitro-grown) B. burgdorferi. Finally, the rabbit antibodies have the capacity to significantly reduce arthritis pathology in mice with an established LD infection. Overall, the indication that NZW rabbits develop a unique repertoire of highly protective antibodies against the LD pathogen needs to be exploited to delineate protective targets for the development of the long-overdue LD vaccine.
MATERIALS AND METHODS
Ethics statement.The animals were maintained at Texas A&M University in an animal facility accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care International (AAALAC). The experimental practices involving animals were approved by the Institutional Animal Care and Use Committee (IACUC) of Texas A&M University (approval number 2017-0390) and were carried out in accordance with Public Health Service Policy on Humane Care and Use of Laboratory Animals (113), the Guide for the Care and Use of Agricultural Animals in Research and Teaching (114), and the Guide for the Care and Use of Laboratory Animals (115).
B. burgdorferi clones and culture conditions.Clones of B. burgdorferi, B31-A3 (B31), B31-A3/lp28-1::kan Δvls (ΔVlsE), and 297 were generous gifts from Troy Bankhead (36, 116) (see Table S1 in the supplemental material). The B. burgdorferi clones were cultivated in liquid Barbour-Stoenner-Kelly medium with 6% rabbit sera (Gemini Bio-Products, USA) (referred to here as BSK-II) and incubated at 35°C under 2% CO2. The ΔVlsE clone was grown in BSK-II supplemented with 200 μg/ml kanamycin or 50 μg/ml streptomycin, respectively.
Infection of rabbits with B. burgdorferi.A total of 11 female New Zealand White (NZW) rabbits 12 to 14 weeks of age (Charles River Laboratories, USA) were inoculated intradermally at six sites along the spine with B31, ΔVlsE, or 297 at 106 spirochetes per site, as described previously (62). The B31- and ΔVlsE-inoculated rabbits were bled via the marginal ear vein, and skin biopsy specimens (3 mm in diameter) (Integra; Miltex, USA) were taken around each inoculation site at weeks 1, 2, 3, and 4 postchallenge. Similarly, blood and skin biopsy specimens were sampled weekly from the 297-inoculated NZW rabbits from week 1 through week 8. The blood was kept at 4°C overnight and then centrifuged at 5,000 × g for 10 min to collect sera. The samples were then stored at −80°C until use. Skin biopsy specimens that represented the six inoculation sites of each rabbit were cultured in 5 ml of BSK-II that contained 0.02 mg ml−1 phosphomycin, 0.05 mg ml−1 rifampin, and 2.5 mg ml−1 amphotericin B (referred to here as the antibiotic cocktail) at 35°C under 2% CO2 and also preserved in Invitrogen RNAlater stabilization solution (Thermo Fisher Scientific, USA) at −80°C until use. All the skin cultures were examined weekly via dark-field microscopy for up to 6 weeks for the presence of spirochetes. The B31- and ΔVlsE-inoculated rabbits were humanely sacrificed at week 4 postchallenge; 297-challenged rabbits were sacrificed at week 8. Blood was obtained from each rabbit via cardiac puncture at the time of sacrifice.
Identification of rates of retention of the lp28-1 plasmid for rabbit skin isolates of B. burgdorferi.Individual B. burgdorferi clones from positive cultures of skin biopsy specimens from B31-inoculated rabbits were isolated by limiting dilution. A total of 200 B31 isolates were PCR screened for the presence of the lp28-1 plasmid. Each B31 isolate was directly screened via colony PCR upon limiting dilution to avoid any loss of lp28-1. For all the isolates, a PCR mixture was prepared in 50 μl of the following reaction mixture: 10 μM each primer (1 μl each), 200 μM each deoxynucleoside triphosphate (dNTP) (New England BioLabs, USA), 1.25 U of Taq DNA polymerase (New England BioLabs) with 10× standard Taq polymerase buffer (New England BioLabs), and 50 to 200 ng of DNA or 2 μl of culture. B31 isolates were individually screened by using the following previously reported primers: 5′-ACACCGCACTAACATCGGGTTC-3′ and 5′- GATACACCTCCTAGTTTGGGTCCTC-3′ (117). The PCR program was carried out as follows: a denaturation step at 95°C for 10 min followed by 25 cycles of 30 s at 95°C, 30 s at 54.4°C, and 45 s at 72°C, with a final extension step for 6 min at 72°C.
qRT-PCR.To examine levels of vlsE transcripts, quantitative reverse transcription-PCR (qRT-PCR) was performed. Skin biopsy specimens from four B31-infected rabbits were collected weekly in RNAlater stabilization solution (Thermo Fisher Scientific) for 4 weeks and stored at −80°C. An Aurum total RNA fatty and fibrous tissue kit (Bio-Rad Laboratories, USA) was utilized to extract RNA from skin tissues collected at days 0, 3, 7, 14, 21, and 28 (two rabbits only) postinfection. DNA samples from day 0 postinfection (prior to infection) were used as a negative control. Next, 1 μg of RNA from the sample from each time point was used to prepare cDNA via an Invitrogen SuperScript II reverse transcriptase kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. To normalize the vlsE expression level, the flaB gene of B. burgdorferi B31 was used. qRT-PCR was performed by using our newly developed primers and probe for vlsE (primers 5′-CGA TCT TAA TAG TTT GCC TAA GGA-3′ and 5′-TCA ACG GCA GTT CCA ACA-3′ and probe 5′-CCC GTC GTA CTA CTT ATA TCG CTT-3′), previously developed primers and probe for flaB (primers 5′-TGT TGC AAA TCT TTT CTC TGG TGA-3′ and 5′-CCT TCC TGT TGA ACA CCC TCT T-3′ and probe 5′-TCA AAC TGC TCA GGC TGC ACC GG-3′) (91), and SsoAdvanced universal probe supermix (Bio-Rad) on a CFX 96 Touch real-time PCR detection system (Bio-Rad). cDNA samples and no-template controls were analyzed in triplicate. The amplification program included (i) heating at 95°C for 3 min for polymerase activation and DNA denaturation and (ii) amplification for 39 cycles with denaturation at 95°C for 15 s and extension and annealing at 60°C for 45 s. The comparative threshold cycle (CT) method (2−ΔΔCT) was used to assess relative gene expression levels (118).
Generation of host-adapted B. burgdorferi clones.Host-adapted B. burgdorferi spirochetes were generated as detailed previously (37, 63, 119, 120). In short, male C3SnSmn.CB17-Prkdcscid/J SCID (SCID) mice (The Jackson Laboratory, USA) at 4 to 6 weeks of age were subcutaneously inoculated with in vitro-grown B31, ΔVlsE, or 297 at 104 cells per mouse. Infection of each animal was confirmed by positive culture of blood samples (50 μl) and ear skin biopsy specimens (∼3 mm in diameter) taken at days 7 and 21 postinoculation and immediately cultivated in 3 and 1 ml of BSK-II with the antibiotic cocktail at 35°C under 2% CO2, respectively. Ear skin tissues harvested at day 21 were stored at −80°C until use.
Passive immunization of mice at the time of challenge.Passive-immunization assays involved male C3H/HeN (C3H) or SCID mice at 5 to 6 weeks of age (The Jackson Laboratory, USA). In the experiments involving heat-inactivated sera, C3H or SCID mice were also used. Although these mice produce their own complement, its contribution to any potential rabbit antibody-mediated killing is insignificant for the following reasons. First, the classical complement pathway was consistently demonstrated to have near-zero activity (121, 122). Second, mouse sera are known to contain an inhibitor(s) of classical complement activation (122). Finally, in addition to very low levels of complement proteins in mouse sera, the complement C4 protein lacks classical pathway C5 convertase subunit activity (123).
Mice were challenged with in vitro-grown B. burgdorferi via subcutaneous inoculation of 104 spirochetes per mouse or host-adapted B. burgdorferi at day 0. The ear skin tissues from infected SCID mice (∼3 mm in diameter) were transplanted into the dorsal lumbar region (one piece per mouse). The animals were retro-orbitally injected with 150 or 200 μl of preimmune or immune sera obtained from B. burgdorferi B31- or ΔVlsE-infected rabbits at day 28 postchallenge. Heat inactivation of complement was performed at 56°C for 45 min (124). Additionally, some mice were individually retro-orbitally inoculated with 150 μl of immune sera at days 2, 4, 6, and 8 postchallenge. Control mice were similarly challenged with in vitro-grown or host-adapted B. burgdorferi and treated with the respective amounts of saline or preimmune sera to ensure that B. burgdorferi clones were infectious at the time of challenge. At day 7, 50 μl of blood from each mouse was cultured in 3 ml of BSK-II with the antibiotic cocktail. At day 14 or 21, bladder, ear pinnae, heart, and tibiotarsal joint tissues harvested from each animal were individually cultured in 1 ml (heart tissue in 3 ml) of BSK-II with the antibiotic cocktail. Dark-field microscopy was used to examine cultures for the presence of viable spirochetes.
Western blotting.Western blotting was performed as described previously (78). In short, B31 spirochetes were grown in BSK-II, harvested by centrifugation, and then washed two times with ice-cold phosphate-buffered saline (PBS), followed by resuspension with sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) sample buffer. About 1 × 106 spirochetes were loaded into each lane of a 15% acrylamide gel (SDS-PAGE). Resolved proteins were transferred to a polyvinylidene difluoride (PVDF) membrane with a pore size of 0.45 μm (Bio-Rad). After blocking with 5% nonfat dry milk, the membrane was incubated in milk containing 1:500-diluted rabbit anti-B31 or -ΔVlsE immune or preimmune sera overnight at 4°C. To generate anti-B31 or -ΔVlsE immune sera, NZW rabbits (3 animals per group) were infected with B31 or ΔVlsE as described above. Blood was collected at days 7, 14, 21, and 28 postinfection, and equal amounts of immune sera from each animal of the same group per time point were pooled. Preimmune sera were used as a negative control. The membrane was washed with Tris-buffered saline (TBS) containing Tween 20 (TBST), and the primary antibodies were detected by using goat anti-rabbit horseradish peroxidase (HRP)-conjugated secondary antibodies (Bio-Rad). The blots were developed via enhanced chemiluminescence.
ELISA.To quantify total levels of IgG in the B31- and ΔVlsE-infected rabbits, sera obtained from the respective rabbits at days 7, 14, 21, and 28 postinfection were individually analyzed by an enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s instructions (Life Diagnostics, USA). The samples were diluted to 1:100,000 or 1:1,000,000 and analyzed in duplicate. Preimmune sera were used as a negative control.
Passive immunization of mice with established B. burgdorferi infection.A total of 9 C3H and 12 SCID mice at 5 to 6 weeks of age (The Jackson Laboratory, USA) were subcutaneously inoculated in the scapular region with in vitro-grown B31 at 104 cells per mouse. Infection was confirmed via culture of blood drawn from each mouse at day 7 postchallenge. At days 14, 18, 22, and 26, individually, 4 C3H and 6 SCID mice were retro-orbitally injected with 200 μl of nontreated (intact complement) immune sera. The pooled sera originated from three B31-infected rabbits at day 28 postinfection. The other 5 C3H and 6 SCID mice were left untreated (controls). At day 45 postchallenge, the mice were sacrificed. Bladder, ear skin, heart, and tibiotarsal joint tissues were harvested and cultured in BSK-II with the antibiotic cocktail. Tibiotarsal joints and heart tissues were subjected to histopathological analysis.
Histopathology.Tibiotarsal joints and hearts were fixed in 10% neutral buffered formalin, processed, and stained with hematoxylin and eosin (H&E). Scoring was performed in a blind manner and independently by two pathologists. The following criteria were used to evaluate the joints: synovial hyperplasia (0, no change; 1, mild; 2, moderate; 3, severe; 4, severe with papilliform growth), exudate within the joint and/or tendon sheath (0, no change; 1, <10 inflammatory cells [neutrophils, macrophages, and lymphocytes]; 2, 10 to 49 inflammatory cells; 3, 50 to 100 inflammatory cells; 4, >100 inflammatory cells), superficial inflammation/resorption of bone (0, no change; 1, mild; 2, moderate; 3, moderate to severe; 4, severe), and overall inflammation (0, no change; 1, 1 to 24% inflammatory cells [neutrophils, macrophages, and lymphocytes]; 2, 25 to 40%; 3, 41 to 60%; 4, 61 to 100%). The overall scores were defined as follows: 0 for no change (no synovial hyperplasia, no exudate within joints and/or tendon sheaths, no superficial bone inflammation/resorption, and no overall inflammation), 1 for mild changes (mild synovial hyperplasia, <10 inflammatory cells within the exudate in joints and/or tendon sheaths, mild superficial bone inflammation/resorption, and 1 to 24% overall inflammation), 2 for moderate changes (moderate synovial hyperplasia, moderate superficial bone inflammation/resorption, >10 to <50 inflammatory cells within the exudate in joints and/or tendon sheaths, and 25 to 40% overall inflammation), 3 for moderate to severe changes (severe synovial hyperplasia, moderate to severe superficial bone inflammation/resorption, >50 to <100 cells within the exudate in joints and/or tendon sheaths, and 41 to 60% overall inflammation), and 4 for severe changes (severe synovial hyperplasia with papilliform growth, severe superficial bone inflammation/resorption, >100 cells within the exudate in joints and/or tendon sheaths, and 61 to 100% overall inflammation). The histological sections of heart tissues were scored as follows: 0 for no changes (no inflammation), 1 for minimal changes (minimal inflammation of lymphocytes and neutrophils [single inflammatory cells] and edema), 2 for mild changes (mild inflammation of lymphocytes and neutrophils [clusters of inflammatory cells] and edema), 3 for mild to moderate changes (mild clusters of lymphocytes and neutrophils and edema with arteritis), and 4 for moderate to severe changes (moderate to severe inflammation of lymphocytes and neutrophils [larger foci of inflammation] and edema).
qPCR analysis.Left tibiotarsal joint and ear skin tissues were collected from each of the 12 SCID mice sacrificed at day 45 postchallenge and stored at −80°C. DNA was extracted via a DNeasy blood and tissue kit (Qiagen) according to the manufacturer’s instructions. Quantitative PCR (qPCR) assays were performed with the CFX 96 Touch real-time PCR detection system by utilizing SsoAdvanced universal SYBR green supermix (Bio-Rad). The following forward and reverse primers were used for qPCR: 5′-AGA GGG AAA TCG TGC GTG AC-3′ and 5′-CAA TAG TGA TGA CCT GGC CGT-3′, respectively, for the actB gene (125) and 5′-GTG GAT CTA TTG TAT TAG ATG AGG CTC TCG-3′ and 5′-GCC AAA GTT CTG CAA CAT TAA CAC CTA AAG-3′, respectively, for the recA gene (126). Amplification was performed in a 20-μl reaction mixture with 300 nM each primer and 100 ng of DNA. The cycling parameters, (i) 95°C for 10 s and (ii) 39 cycles of 95°C for 30 s and 60°C for 1 min, were optimized according to the SsoAdvanced supermix manufacturer’s protocol. To generate absolute standards, 138-bp and 222-bp DNA fragments of actB and recA were amplified from C3H mouse and B31 DNAs, respectively. Amplification was performed in a 50-μl reaction mixture that contained 0.2 μM each primer, 200 μM dNTP, 1.25 U of Taq DNA polymerase (New England BioLabs), and 100 ng of DNA. The following PCR program was used: a denaturation step at 95°C for 10 min followed by 25 cycles of 30 s at 95°C, 30 s at 60°C (50°C for recA), and 45 s at 72°C, with a final extension step for 6 min at 72°C. The PCR products were purified via a PCR purification kit (Qiagen). All standard dilutions and DNA samples were amplified in triplicate. The number of spirochetes was calculated as the ratio of recA DNA copies per copy of actB gene.
Phage display library.A mix of 50 μl or 20 μl of each mouse or rabbit serum sample, respectively, and 10 μl of random peptide library Ph.D.-7 (New England BioLabs, USA) was incubated at 25°C for 18 h. Antibody-bound phages were then isolated and eluted as previously detailed (78, 127). After amplification, the phages were subjected to two rounds of biopanning and then isolated via protein G-agarose beads. DNA was extracted by phenol-chloroform extraction and ethanol precipitation. The multiplex PCR-amplified DNA library was generated as described previously (78) and sequenced via an Illumina HiSeq 2500 platform (University at Buffalo Genomics and Bioinformatics Core, New York State Center of Excellence in Bioinformatics and Life Sciences, Buffalo, NY). The sequencing generated approximately 1.79 × 108 DNA reads, which were demultiplexed based on unique bar codes, as detailed previously (78). The obtained data were analyzed via Python (https://www.python.org).
Prediction of anti-B31 antibody reactivity to VlsE.For each serum sample, all peptides were mapped to VlsE of the B. burgdorferi B31 strain (here B31-VlsE) (GenBank accession number AAC45733.1), utilizing BLASTP with an identity threshold of 4: alignments with 4 or more exact amino acid matches were considered. For each VlsE position X, a peptide with the amino acid matched to position X and with K different VlsE matches contributed its frequency divided by K to the coverage of X. The coverage of X, C(X), was computed as the sum of contributions of all peptides matched to X.
Statistical analysis.Two-tailed Fisher’s exact test was used for comparison of mouse groups. One-way analysis of variance (ANOVA) (GraphPad Prism software) was applied to compare the levels of vlsE transcription. Histopathological scores were analyzed by a paired t test or an unpaired t test corrected for multiple comparisons using the Holm-Sidak method. Differences were considered significant at a P value of <0.05.
ACKNOWLEDGMENTS
We thank Troy Bankhead for providing the B31 and ΔVlsE clones. We are also thankful to Sankar P. Chaki, Jianhua Guo, Raquel R. Rech, Kelsey L. Johnson, and Nicholas Wolff for technical assistance.
The work at Texas A&M University was partially supported by NIH grant R03AI135159-02, the Department of Veterinary Pathobiology, Texas A&M College of Veterinary Medicine and Biomedical Sciences, and Texas A&M AgriLife. The work at the Roswell Park Comprehensive Cancer Center was partially supported by the Phillip Hubbell Family Fund. The work at Georgia State University was partially supported by NSF grant DBI-1564899, NSF grant CCF-16119110, and NIH grant 1R01EB025022-01.
FOOTNOTES
- Received 27 February 2019.
- Returned for modification 20 March 2019.
- Accepted 11 April 2019.
- Accepted manuscript posted online 15 April 2019.
Supplemental material for this article may be found at https://doi.org/10.1128/IAI.00164-19.
- Copyright © 2019 American Society for Microbiology.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.↵
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵