Functional characterization of antibodies against Neisseria gonorrhoeae opacity protein loops

Abstract

Background: The development of a gonorrhea vaccine is challenged by the lack of correlates of protection. The antigenically variable neisserial opacity (Opa) proteins are expressed during infection and have a semivariable (SV) and highly conserved (4L) loop that could be targeted in a vaccine. Here we compared antibodies to linear (Ab(linear)) and cyclic (Ab(cyclic)) peptides that correspond to the SV and 4L loops and selected hypervariable (HV(2)) loops for surface-binding and protective activity in vitro and in vivo.

Methods/findings: Ab(SV cyclic) bound a greater number of different Opa variants than Ab(SV linear), including variants that differed by seven amino acids. Antibodies to the 4L peptide did not bind Opa-expressing bacteria. Ab(SV) (cyclic) and Ab(HV2) (cyclic), but not Ab(SV) (linear) or Ab(HV2 linear) agglutinated homologous Opa variants, and Ab(HV2BD) (cyclic) but not Ab(HV2BD) (linear) blocked the association of OpaB variants with human endocervical cells. Only Ab(HV2BD) (linear) were bactericidal against the serum resistant parent strain. Consistent with host restrictions in the complement cascade, the bactericidal activity of Ab(HV2BD) (linear) was increased 8-fold when rabbit complement was used. None of the antibodies was protective when administered vaginally to mice. Antibody duration in the vagina was short-lived, however, with <50% of the antibodies recovered 3 hrs post-administration.

Conclusions: We conclude that an SV loop-specific cyclic peptide can be used to induce antibodies that recognize a broad spectrum of antigenically distinct Opa variants and have agglutination abilities. HV(2) loop-specific cyclic peptides elicited antibodies with agglutination and adherence blocking abilities. The use of human complement when testing the bactericidal activity of vaccine-induced antibodies against serum resistant gonococci is also important.

Figure 1. Conservation of SV and HV₂ loop sequences.

The predicted amino acid sequences of the SV and HV₂ loops of the 8 Opa proteins of strain FA1090 are shown. The sequences of the 36-mer cyclic peptides used to generate SV- and HV_2BD-specific antisera are highlighted in grey. The sequences of the linear peptides used to generate affinity-purified rabbit antibodies against the SV and HV₂ loops are shown in light grey font.

Figure 2. Specificity of antibodies against the conserved Opa loops.

Cell lysates of an Opa-negative variant and each of the 8 Opa–positive variants of strain FA1090 were incubated at RT or 100°C prior to fractionation and incubated with (A and D) Ab_{SV linear} (1∶75,000), (B and E) Ab_{SV cyclic} (1∶6,000), and (C) Ab_{4L linear} (1∶6,000). The broader reactivity of Ab_{SV cyclic} compared to Ab_{SV linear} is clearly shown whether native or denatured Opa proteins are analyzed. Ab_{4L linear} is even more broadly reactive in that it recognizes all denatured Opa proteins well except OpaE. None of the antibodies recognize a protein in the Opa-negative lane (−). The location of a 25 kDa molecular weight marker is indicated, and the denatured Opa proteins migrated at a slightly higher molecular weight than those incubated at RT, which is consistent with well characterized heat modifiable nature of these proteins. The Ab_{4L linear} immunoblot was kindly provided by Dr. Amy Simms.

Figure 3. Detection of antibody surface-binding by the SBI assay.

Opa-specific antibodies or antisera were serially diluted and incubated with defined Opa variants spotted onto nitrocellulose filters and binding was detected as described in the Materials and Methods. Shown are the normalized spot intensities plotted against the concentration of affinity-purified rabbit antibodies (Ab_{HV2 linear}, Ab_{SV linear}, and Ab_{4L linear}) or dilution of antisera (Ab_{SV cyclic} and Ab_{HV2BD cyclic}): (A) Ab_{HV2 linear,} (B) Ab_{SV linear}, (C) Ab_{4L linear} (D) Ab_{SV cyclic,} (E) Ab_{HV2BD cyclic}. All results shown are representative of at least two independent experiments. None of the antibodies bound Opa-negative gonococci. The results for Ab_{HV2BD linear} when tested against Opa-negative variants are shown in panel A as an example.

Figure 4. Bactericidal and agglutination activities.

The bactericidal activity of SV and HV₂ loop-specific antibodies was measured against (A) OpaB variants of the serum sensitive strain FA1090_{F62por 5–8} in the presence of 1% NHS and (B) the constitutively OpaB-expressing strain of FA1090 (SR) incubated with Ab_{HV2BD linear} in the presence of NHS or BBS. Results are expressed as the percent of bacteria recovered compared to wells with no antibody present. Only Ab_{HV2BD linear} was bactericidal against the OpaB-expressing serum sensitive strain. The use of a non-human complement source (BBS) resulted in ∼8-fold greater bactericidal activity for Ab_{HV2BD linear} against serum resistant (wild type) OpaB-expressing variants. Ab_{HV2BD linear} was not bactericidal when HI-NHS and HI-BBS were used (data not shown). Agglutination activity was measured as the average number of bacterial aggregates following incubation of (C) OpaB variants with increasing dilutions of Ab_{HVBD cyclic}, Ab_{SV cyclic}, or NMS and (D) OpaA or OpaK variants with Ab_{SV cyclic} or NMS. Antibodies raised to the cyclic SV peptides showed agglutinating activity except when tested against OpaK variants.

Figure 5. Antibody-mediated inhibition of gonococcal association with tissue culture cells.

Pre-incubation with Ab_{HV2BD cyclic} but not Ab_{SV cyclic} decreased the total number of ME180 cell-associated OpaB-expressing gonococci when an antiserum dilution of 1∶30 was used. Shown is the average percent of cell-associated bacteria that were preincubated with Ab_{HV2BD cyclic} (test) or NMS (control) based on combined data from three independent assays. A two-tailed, unpaired Student's t-test was used to assess statistical differences between test and control wells.

Figure 6. Duration of topically applied antibodies in the vagina.

A fifty percent reduction in the concentration of topically applied rabbit antibodies was detected in vaginal washes collected 1 hr post-inoculation compared to a 2 min time point. Antibody levels decreased further over time. The limit of detection was 7.8 ng/ml. (A) Concentration of rabbit IgG (µg/ml) in mouse vaginal washes was determined in duplicate by capture ELISA. (B) The percent of antibody remaining relative to the average 2 minute value is shown for each time point. Results are combined from two experiments and each animal was used for a single time point (n = 2–3 mice per time point).

Figure 7. Passive protection experiments.

Mice were inoculated with OpaA or OpaB variants that were preincubated with antibodies against linear or cyclic Opa SV or HV₂-loop sequences. (A) OpaA variants with Ab_{SV linear}, Ab_{HV2A linear}, or Ab_{HV2BD linear}. There was no difference in the number of gonococci recovered from mice in the Ab_{SV linear} and Ab_{HV2A linear} test groups compared to the Ab_{HV2B linear} control group on days 1 or 2 post-inoculation (p≤0.14). (B) OpaB variants with Ab_{HV2BD cyclic}, Ab_{SV cyclic} or NMS. There was no difference in the number of gonococci recovered from mice treated with Ab_{SV cyclic} compared to the control (NMS) group (p≤0.20) on days 1 or 2 post-inoculation. (C) OpaB variants with Ab_{HV2BD linear} and Ab_{HV2A linear}. Shown are combined data from three independent experiments that showed no difference in the number of bacteria recovered from each group on days 1 and 2 post-inoculation at the level of p = 0.07 and p = 0.08, respectively (n = 7–15 mice/group in each experiment; total number = 35 mice/group). Symbols indicate a single animal and horizontal bars indicate the group mean with the SEM shown. Statistical differences were analyzed by a two-tailed Student's t test.