INTRODUCTION
Lyme borreliosis (LB) is a major multiorgan infectious disease occurring in the United States, Europe, and Asia and can have serious health consequences if left untreated (
1,
2). LB was the seventh most common notifiable disease in the United States in 2012 (
2,
3). Moreover, recent CDC data estimate that the number of diagnosed cases in the United States is approximately 300,000 annually, which is 10-fold higher than the number of reported cases (
4). There are about 85,000 cases reported annually in Europe; however, this number is largely underestimated, as case reporting is highly inconsistent in Europe, and many LB infections go undiagnosed (
5). LB is caused by a number of bacterial species of the
Borrelia burgdorferi sensu lato complex. Four species,
B. burgdorferi sensu stricto,
Borrelia afzelii,
Borrelia garinii, and
Borrelia bavariensis cause the majority of human disease in Europe, whereas only a single species,
B. burgdorferi sensu stricto, causes LB in the United States (
1,
2).
LB can usually be successfully treated with antibiotics (
1,
2), but patients and health care providers may remain unaware of the infection until the onset of severe disease symptoms (
6). Approximately 60% of the untreated infections in U.S. patients result in Lyme arthritis (
1). Moreover, about 10% of patients in the United States do not respond clinically to antibiotic treatment and are said to have antibiotic-refractory Lyme arthritis (
1,
7).
Not all individuals who are infected with
B. burgdorferi develop symptoms. In the United States, epidemiological studies have reported the frequency of asymptomatic
B. burgdorferi sensu stricto infection to be between 0% and 50% (
8–11). In Europe, where more infections are caused by
B. afzelii,
B. garinii, and
B. bavariensis than by
B. burgdorferi sensu stricto, studies have reported that the majority of seropositive individuals were asymptomatic (
11–14). Importantly, both previously infected asymptomatic individuals and some LB patients who recover after antibiotic treatment remain vulnerable to reinfection (
15,
16); thus, previously infected individuals represent a substantial proportion of those at high risk for LB. Vaccination would be the most effective intervention to prevent LB, but there is no vaccine currently available for human use.
Two monovalent recombinant vaccines based on outer surface protein A (OspA) serotype-1 derived from
B. burgdorferi sensu stricto were developed in the 1990s and demonstrated to be safe and efficacious in placebo-controlled phase III studies in the United States (
17,
18). However, a subsequently disproven hypothesis that a licensed monovalent OspA-1 vaccine might trigger arthritis in some vaccine recipients, mediated by T-cell cross-reactivity with an OspA-1 epitope (
19,
20), was one of a number of factors that contributed to the limited acceptance and subsequent discontinuation of the vaccine in 2002 (
6,
21). Irrespective of these disproven safety concerns, monovalent OspA vaccines were designed for use only in the United States and would not have been effective in preventing LB occurring in Europe or Asia. Because OspA-mediated protection is largely OspA serotype specific and the disease in Europe and Asia is caused by several
Borrelia species that encode antigenically divergent OspA proteins (
1,
2), a globally effective OspA vaccine requires the induction of antibodies against multiple OspA serotypes.
We have developed a novel multivalent OspA vaccine, which comprises three recombinant OspA antigens, each containing protective epitopes from two different OspA serotypes, i.e., OspA serotypes 1 and 2 (
B. burgdorferi sensu stricto and
B. afzelii), 5 and 3 (both
B. garinii), and 6 and 4 (
B. garinii and
B. bavariensis) (
22). The multivalent vaccine is designed to protect against all major disease-causing
Borrelia species in the United States (OspA-1), Europe (OspA-1 to -6), and potentially globally. The hypothetical risk of T-cell cross-reactivity has been eliminated by replacing the putative cross-reactive OspA-1 epitope with the corresponding OspA-2 sequence (
23). In a phase I/II study in healthy adults seronegative for
B. burgdorferi sensu lato infection, the novel multivalent OspA vaccine was demonstrated to be well tolerated and to induce potent antibody responses against all major
Borrelia species after three primary immunizations (
22). In the present study (ClinicalTrials.gov registration no. NCT01504347), we extend our investigation to include a study population seropositive for
B. burgdorferi sensu lato infection. We also investigated the seropersistence of the primary antibody responses up to 12 months after the first immunization, and we evaluated the effectiveness of a booster immunization at 6 or 9 to 12 months after the first immunization.
RESULTS
Study population.
Table 1 shows the demographic and clinical characteristics of the participants at baseline. The dose groups were balanced with respect to all demographic characteristics. Of the 151 participants who were seropositive for
B. burgdorferi sensu lato antibodies as assessed by C6-ELISA (range of Lyme index, 1.10 to 11.47 for seropositive subjects and 0.03 to 0.90 for seronegative subjects), 89 (58.9%) reported previously having LB symptoms, indicating that a substantial proportion of seropositive subjects had had asymptomatic infection with
B. burgdorferi sensu lato.
Figure 1 shows the trial profile. A total of 199 seronegative and 151 seropositive subjects were randomized to receive three priming immunizations and a booster immunization with either 30 μg or 60 μg of OspA antigen. The safety and immunogenicity data sets contain all subjects who were vaccinated at least once and had baseline and at least one postvaccination titer measurement.
Safety and tolerability.
Figure 2 shows the rates of systemic and injection site reactions within 7 days of each immunization. The majority of the AEs occurred within 24 h of immunization, were predominantly mild in severity, and resolved spontaneously within 72 h. Successive vaccinations were generally associated with a decrease in systemic reaction rates. There were no statistically significant differences in the rates of systemic (relative risk [RR], 1.09 to 1.13;
P > 0.5862) or injection site (RR, 1.02 to 1.16;
P > 0.2261) reactions in the seronegative and seropositive subjects.
The rates of individual solicited systemic and injection site reactions after the first immunization are shown in
Table 2. The most common injection site reactions in both the seronegative and seropositive populations were pain (36.7 to 48.7%) and tenderness (30.3 to 49.5%). Most solicited systemic reactions occurred very infrequently, irrespective of dose or serological status, and the vast majority were rated as mild; only headache (4.0 to 11.9%), myalgia (5.3 to 13.3%), and fatigue (5.1 to 7.9%) occurred at frequencies of >6%. The frequencies of local and systemic reactions in the seronegative and seropositive subjects receiving the 30-μg or 60-μg dose were very similar (RR, 0.88 to 1.05;
P = 0.3370 to 0.9511). The risk of moderate or severe systemic reactions for the 30-μg and 60-μg dose was very similar in both populations (RR, 0.97 to 1.11;
P = 0.9054 to 0.9651).
No deaths or vaccine-related serious AEs were reported during the entire study, and no symptoms reminiscent of LB or chronic arthritis were reported. Three cases of transient arthritis were reported in three subjects immunized with the 60-μg vaccine dose. In one case, a seronegative subject reported an inflamed left toe joint, which was mild and resolved within 4 days. This subject had a history of joint pain with the same localization prior to the study. In the second case, a seropositive subject with a history of LB presented with acute arthritis of the proximal interphalangeal finger joints of two fingers of the right hand, of moderate severity; however, an X-ray examination gave no indication of arthritis or arthrosis, and although the subject experienced tenderness, swelling, and pain, all rheumatologic and laboratory assessments (e.g., rheumatoid factor, anti-citrullinated protein antibody, and C-reactive protein) were negative. In the third case, a seronegative subject reported arthritis in both hands with an onset of 36 days after the third vaccination, which resolved within 11 days. Ten severe adverse effects (SAEs) were reported in 8 subjects, all between 6 months after the first immunization and the booster immunization, i.e., during a period when no immunizations took place; all were considered to be unrelated to vaccination.
Total IgG antibody response after primary immunizations.
Figure 3A shows the total IgG ELISA antibody responses to each OspA antigen after the primary immunization schedule. At baseline, the geometric mean titers (GMTs) against all six OspA antigens included in the vaccine were very low, and there was no difference in the OspA antibody titers between the C6-seronegative and -seropositive populations. Substantial GMTs against all of the six OspA antigens were induced after the primary immunization schedule in the seronegative and seropositive populations (range, 3,799 to 8,543 and 2,413 to 9,435, respectively, for the two different formulations).
In the seronegative participants, the ELISA antibody responses induced against each OspA serotype were comparable for the 30-μg (range, 3,799 to 6,937) and 60-μg (range, 4,575 to 8,543) doses (P ≥ 0.062). However, in the seropositive subjects, the 60-μg dose (range, 4,895 to 9,435) resulted in significantly higher GMTs than those for the 30-μg dose (range, 2,413 to 4,371) (P < 0.0001). There was a trend toward lower antibody titers in the seropositive population compared to those of the seronegative population for the 30-μg dose; however, this was not statistically significant. The antibody titers against all six serotypes were significantly affected by time in both the seronegative and seropositive subjects (P, <0.0001 and <0.0001, respectively), logarithmic titer at baseline (P, ≤0.0001 and ≤0.0045, respectively), and age (P, ≤0.0067 and ≤0.0536, respectively).
Total IgG antibody seropersistence and booster response.
Six months after the first immunization, the ELISA antibody titers were maintained at levels substantially and significantly above baseline in the seronegative (
P < 0.0001) and seropositive (
P < 0.0001) subjects (
Fig. 3B). Analogous to the situation 1 month after the third immunization, no differences were observed in the seronegative subjects (GMT range, 933 to 1,641 and 1,213 to 2,036 for the 30-μg and 60-μg doses, respectively), but the seropositive subjects receiving the 60-μg dose maintained higher antibody titers than those receiving the 30-μg dose (range, 1,342 to 2,306 and 680 to 1,103, respectively).
In both the seronegative and seropositive subjects receiving a booster 6 months after the first immunization, the ELISA antibody responses against all six OspA serotypes contained in the vaccine were substantially higher than those measured after the primary immunization schedule (
Fig. 3B). The postbooster antibody GMTs were comparable for the subjects administered the 30-μg (seronegative range, 9,927 to 14,591; seropositive range, 10,419 to 15,896) and 60-μg (seronegative range, 11,545 to 18,102; seropositive range, 8,064 to 11,167) doses, with no statistically significant dose effect.
Nine to 12 months after the first immunization, the antibody titers were maintained at levels substantially and significantly above baseline in the seronegative (
P < 0.0001) (GMT range, 339 to 558 and 590 to 1,054 for the 30-μg and 60-μg doses, respectively) and seropositive (
P ≤ 0.0141) (GMT range, 300 to 613 and 531 to 870 for the 30-μg and 60-μg doses, respectively) subjects (
Fig. 3C). In the subjects receiving a booster 9 to 12 months after the first vaccination, increases in the antibody GMTs against all six OspA serotypes were observed for both vaccine doses in both populations (
Fig. 3C), which were higher than those observed after the 6-month booster. In the seronegative subjects, the postbooster antibody GMTs were comparable for subjects across the two doses (range, 23,799 to 41,735), with no statistically significant differences across all OspA serotypes. However, in the seropositive subjects, the postbooster antibody GMTs were significantly higher against 5 of the 6 OspA serotypes (
P ≤ 0.0359) in the subjects administered the 60-μg adjuvanted dose (range, 28,735 to 42,381) than those in the subjects who received the 30-μg dose (range, 12,653 to 17,485). Reverse cumulative distributions of the ELISA antibody titers induced by the multivalent OspA vaccine are shown in
Fig. 4. After the third primary immunization, ≥93% of the seronegative and ≥94% of the seropositive participants had ELISA antibody titers of ≥1,000 against the individual OspA serotypes; ≥57% and ≥61%, respectively, achieved titers of ≥5,000. After the booster immunization, ≥95% of the seronegative and ≥97% of the seropositive participants had ELISA antibody titers of ≥5,000; ≥87% and ≥79%, respectively, achieved titers of ≥10,000.
Functional antibody responses.
In addition to the high titers of the total IgG OspA antibodies induced by the multivalent OspA vaccine, potent functional antibody responses were generated that bound to the surface of
B. burgdorferi sensu stricto (OspA-1),
B. afzelii, (OspA-2),
B. bavariensis (OspA-4), and
B. garinii (OspA-3, -5, and -6) (
Fig. 5). It should be noted that the differences in surface-binding antibody titers measured in the different OspA types reflect differences in the expression levels of OspA on the different
Borrelia strains used rather than differences in the potencies of the functional antibody responses. Thus, the comparisons are valid within the same dose and population across time points but not between different OspA serotypes.
At baseline, the seropositive population had significantly higher surface-binding antibody titers against 5 of the 6 Borrelia strains used in the surface-binding assay (P ≤ 0.0486). After the primary immunization schedule, the surface-binding antibody GMTs among the different OspA serotypes ranged from 10.3 to 1,092.7 in the seronegative subjects and from 9.4 to 1,193.8 in the seropositive subjects. After the 9- to 12-month booster immunization, the GMTs among the different serotypes ranged from 108.4 to 3,744.8 in the seronegative subjects and from 79.3 to 4,093.6 in the seropositive subjects.
The antibody titers measured by the surface-binding assay were highly significantly correlated with those measured by ELISA (Spearman correlation coefficient, 0.786 to 0.943 for the six different OspA serotypes). In agreement with the ELISA data, there were no differences in the GMTs induced by the 30-μg and 60-μg doses in the seronegative subjects, but there was a dose response in the seropositive subjects, with higher GMTs induced by the 60-μg dose than by the 30-μg dose. Also in agreement with the ELISA data, higher surface-binding antibody titers were induced by the 9- to 12-month booster than those by the 6-month booster (
Fig. 5).
DISCUSSION
A novel multivalent OspA vaccine is safe and immunogenic in healthy seronegative and seropositive adults and induces a significant increase in total ELISA and functional surface-binding antibody titers against all six OspA serotypes after vaccination with either a 30-μg or 60-μg OspA antigen dose adjuvanted with alum. The antibody responses induced after a three-dose primary immunization schedule declined up to 12 months after the first immunization but remained significantly above baseline and were effectively boosted by a fourth dose either 6 months or 9 to 12 months after the first immunization. The booster response was higher when the booster immunization was administered 9 to 12 months than at 6 months after the first immunization. In the seronegative subjects, the antibody responses induced by the 30-μg and 60-μg doses were similar, in agreement with a previous study of the multivalent OspA vaccine in a seronegative population (
22). However, in the seropositive subjects, the 60-μg dose induced significantly higher antibody titers than did the 30-μg dose. The vaccine was well tolerated, with similar tolerability profiles across the two doses in both populations.
As expected, because OspA is downregulated upon transmission to the infected host (
25), such that OspA antibodies are not usually detectable in infected individuals, there was no difference in the baseline OspA antibody titers in the seropositive and seronegative populations. In contrast, because the surface-binding assay detects all antibodies that bind to
Borrelia, the baseline surface-binding antibody titers were significantly higher against 5 of the 6
Borrelia strains used in the surface-binding assay in the seropositive population than those in the seronegative population.
It is not clear why a dose response was observed in the seropositive but not the seronegative population, such that the 30-μg dose induced lower antibody titers in the seropositive population than those in the seronegative population. It has been suggested that prior infection with
B. burgdorferi sensu lato might blunt the immune response, resulting in a lower ability of seropositive subjects to mount an immune response to vaccination with OspA (
26,
27). Alternatively, because of the lipidated nature of the OspA vaccine and the high expression of lipoproteins by
B. burgdorferi sensu lato, it is possible that the anti-lipoprotein antibodies induced in seropositive subjects by previous infection interfere with the induction of antibody responses against the OspA vaccine, such that higher antigen doses are required in the seropositive population.
One previous study of a monovalent OspA-1 vaccine investigated the antibody responses to vaccination in seropositive subjects, but this study did not include a head-to-head seronegative cohort (
27). Because of the lack of a direct comparator group in this previous study and the use of a different assay format than those in studies on the same vaccine in seronegative populations (
28–33), it is not clear whether the previous monovalent OspA-1 vaccine induced similar antibody titers in seronegative and seropositive populations.
The higher antibody responses to the 9- to 12-month booster than to the 6-month booster is a known phenomenon reported in other studies of inactivated vaccines, in which later booster vaccinations were found to induce higher antibody titers (
34). Residual antibodies present after priming can result in the formation and subsequent rapid clearance of antigen-antibody complexes (
35,
36). Antibody secreting B or plasma cells may also negatively regulate cognate T-helper-cell functions essential for antigen-specific secondary B-cell responses (
37,
38). In a previous study of a monovalent OspA-1 vaccine that compared the antibody responses induced by a two-dose primary immunization regimen followed by either a 6-month or 12-month booster, higher antibody responses were also induced after the later booster than after the earlier booster, although it was concluded that the two vaccination schedules elicited equivalent immune responses (
33).
A considerable drop in antibody titers was observed prior to the booster immunizations. Despite this, the antibody levels for all OspA serotypes were substantially and significantly above baseline at these time points. Although a correlate of protection can be determined only in phase III efficacy studies, it is likely that repeat booster immunizations will be necessary to maintain high levels of circulating antibodies required for protection.
The primary ELISA-based immunogenicity data are supported by the demonstration that vaccine-induced antibodies are functionally capable of binding to
B. burgdorferi,
B. afzelii,
B. bavariensis, and
B. garinii species expressing OspA serotypes 1 to 6, which are representative of all major human-pathogenic species. In preclinical studies in mice, OspA and surface-binding antibody titers have been shown to correlate with protection from
Borrelia challenge (
39). Importantly, the antibody titers measured by the surface-binding assay were highly correlated with those measured by ELISA, and they were also in agreement with the ELISA data with respect to the differences in dose response in the seronegative and seropositive subjects, as well as the higher titers induced by the 9- to 12-month booster than those of the 6-month booster.
Taken together, the study data suggest that the novel multivalent OspA vaccine is equally safe and well tolerated in seronegative and seropositive subjects, and it induces substantial antibody responses against all six OspA serotypes included in the vaccine. A substantial booster response was induced in both populations by a booster vaccination either 6 months or 9 to 12 months after the first immunization. The study data also demonstrate that a 60-μg dose is the preferred dosage for entry into phase III trials, as this dose induces significantly higher titers than does a 30-μg dose in the seropositive population, with no significant difference in tolerability. This is the first study reporting the direct comparison of an OspA vaccine in B. burgdorferi sensu lato-seronegative and -seropositive populations. The potential limitations of this study include the fact that the screening C6-ELISA may not discriminate 100% between subjects previously exposed to LB and seronegative subjects. Additionally, our study did not include patients who had preexisting antibody titers to OspA, as is sometimes found in patients with late-stage Lyme disease. We also did not systematically capture a detailed medical history on previous LB manifestations in the seropositive individuals. However, as seropositive individuals remain at high risk for subsequent infection with B. burgdorferi sensu lato, it is reassuring that the study data suggest that the novel multivalent OspA vaccine should be effective in both populations.
ACKNOWLEDGMENTS
We thank the following members of the Baxter research and development team for their role in this study: Michael Krammer, Sabine Geyer, Claudia Rieder, Sandra Foco, and Sandra Zinke. We thank the members of the independent data monitoring committee: Gerold Stanek, Frank V. Sonnenburg, and Andreas Krause. We also thank Marika Gaile at the University of Tübingen Institute of Tropical Medicine.
The study was designed and funded by Baxter. Baxter employees were responsible for the study design, data collection, analysis and interpretation, writing of the report, and the decision to submit for publication.
N.W., P.N.B., E.-M.P., M.O., D.P., M.G.S., B.A.C., I.L., T.D., and G.A. are employed by Baxter. N.W., P.N.B., E.-M.P., M.O., M.G.S., B.A.C., I.L., T.D., and G.A. have stock and share options in Baxter. N.W., P.N.B., M.O., M.G.S., B.A.C., I.L., and G.A. report planned, pending, or issued patents on Lyme vaccines. B.S., M.Z., M.E., P.G.K., and R.A. report receiving a grant from Baxter for performing the study. T.J. and R.W. report receiving reimbursement for performing the study. H.K and I.F.K.N. report no conflicts of interest. We have all completed the ICMJE Form for Disclosure of Potential Conflicts of Interest.