Abstract
Neisseria gonorrhoeae is an on-going public health problem due in part to the lack of success with efforts to develop an efficacious vaccine to prevent this sexually transmitted infection. The gonococcal transferrin binding protein B (TbpB) is an attractive candidate vaccine antigen. However, it exhibits high levels of antigenic variability, posing a significant obstacle in evoking a broadly protective immune response. Here, we utilize phylogenetic information to rationally select TbpB variants for inclusion into a gonococcal vaccine and identify two TbpB variants that together elicit a highly cross-reactive antibody response against a diverse panel of TbpB variants and clinically relevant gonococcal strains. This formulation performed well in experimental proxies of real-world usage, including eliciting bactericidal activity against diverse gonococcal strains and decreasing the median duration of colonization after vaginal infection in female mice. These data support the use of a combination of TbpB variants for a broadly protective gonococcal vaccine.
Similar content being viewed by others
Introduction
Neisseria gonorrhoeae (Ngo) causes the sexually transmitted infection gonorrhea, which has been successfully treated with antibiotics for decades. Worryingly, an increasing prevalence of antibiotic-resistant Ngo threatens our ability to successfully treat this infection1, leading to the potential of untreatable disease by multi-drug-resistant strains2,3. This has created a strong impetus to develop new countermeasures to combat gonococcal infections. The modest but promising cross-protection against gonococcal infection elicited by the MeNZB outer membrane vesicle vaccine4 and the multicomponent meningococcal B vaccine 4CMenB (Bexsero)5,6,7 has demonstrated that vaccine-mediated protection against the gonococcus is feasible, and has reinvigorated efforts to develop an efficacious, broadly protective vaccine.
As the gonococcus lacks a polysaccharide capsule, most vaccine efforts have focused on surface-exposed proteins, including gonococcal pili8, which failed to elicit protection in male volunteers. One long-standing candidate is the bacterial transferrin receptor, composed of the integral outer membrane protein transferrin binding protein A (TbpA) and the surface-anchored lipoprotein transferrin binding protein B (TbpB). TbpA is vital for the gonococcus to acquire the essential micronutrient iron from the host protein transferrin (Tf)9, while TbpB specifically binds iron-loaded (holo) human Tf (hTf) to increase the efficiency of iron uptake10. The gonococcal transferrin receptor was shown to be essential for male urethral infection in humans by strains of Ngo that lack a lactoferrin receptor11,12, thus both transferrin binding proteins have been posited to be attractive candidate vaccine targets. While TbpA exhibits far less sequence variation compared to TbpB13, inherent difficulties exist in targeting integral membrane proteins, including protein stability, large-scale protein production, and methods of focusing the immune response to specifically target surface-exposed regions of the protein. In contrast, TbpB is a soluble, stable lipoprotein, making commercial production and formulation of a TbpB-based vaccine more tractable. However, TbpB exhibits substantial antigenic variation14,15, presumably due to its accessibility to the immune system since its lipidated and flexible peptide anchor allows it to extend away from the bacterial surface to capture transferrin.
TbpB from the closely related pathogen Neisseria meningitidis (Nme) has been extensively evaluated for the diversity of variant sequences and cross-reactivity of antibodies elicited after immunization13,16. Meningococcal TbpBs fall into two distinct phylogenetic clusters exhibiting low sequence identity and eliciting minimal cross-reactivity; the smaller isotype I TbpBs (1.8 kb gene, ~68 kDa protein) and the larger and more diverse isotype II TbpBs (2.1 kb gene, ~80–90 kDa protein)17. A crystal structure of a TbpB from each meningococcal cluster has been solved (isotype I, strain B16B6, PDB ID #4QQ113; isotype II, strain M982, PDB ID #3VE218), both of which show a bi-lobed protein containing an N and C lobe, with each lobe consisting of a handle and barrel domain. The N lobe of TbpB is responsible for binding to the C lobe of iron-loaded hTf19,20 and is more variable than the well-conserved C lobe13. Gonococcal TbpB variants all fall within the meningococcal isotype II TbpB cluster, forming two distinct subclusters21. Thus, while there is still antigenic variability to contend with, it is expected that this variability is manageable with respect to the selection of variants for inclusion into a vaccine formulation.
Here, we evaluate the potential of combining multiple gonococcal TbpB variants to elicit broad cross-reactivity. We use a phylogenetically informed selection of variants to generate vaccine formulations consisting of different combinations of TbpBs, quantify the cross-reactive and bactericidal properties of the elicited antibodies, and demonstrate protection against gonococcal challenge in the murine lower genital tract colonization model. Combined, this work reinforces the potential to develop a broadly protective TbpB-based gonococcal vaccine.
Results
Unbiased selection of TbpB antigens based on phylogenetic analysis
Gonococcal TbpBs exhibit antigenic variation that separates into two distinct phylogenetic clusters (Fig. 1A). The larger of these, which we have termed Ngo cluster 1, is most similar to the previously defined Nme cluster 3, which contains the TbpBs from meningococcal strains H44/76 and N224, and the smaller gonococcal cluster which we have termed Ngo cluster 2, which is most similar to TbpBs from Nme cluster 2, containing TbpBs from meningococcal strains MC58 and M982 (Fig. 1B).
A Phylogenetic tree of 325 Ngo TbpB sequences, clustered into five clades. The representative sequence selected for each clade is indicated in green. Gonococcal cluster 1 variants are on the right and encompass the 27230, 48627, 27464, and WHO Y clades, while the gonococcal cluster 2 variants consist of the 41385 clade in the upper left. B Phylogenetic tree of 1485 TbpB sequences from Ngo and Nme, clustered into seven descriptive clades. The two gonococcal clades are found on opposite sides of the tree and are highly dissimilar. The Nme sequences are clustered into five clades, with the Isotype I sequences as the most dissimilar to anything else in the tree. The Nme 3 and Nme 4 clades are well separated and distinct, while the large Nme 1 and Nme 2 clades contain more of a gradient of sequence groups. The five representative Ngo sequences selected for further characterization are indicated in green, and the 24 sequences indicated in gray were utilized in the cross-reactivity panels in Figs. 2–4, and 6.
As a first step towards evaluating potential gonococcal TbpB variants for their ability to elicit broad cross-reactivity, we selected five representative Ngo TbpB sequences using the phylogenetic analysis software Navargator (Fig. 1A)22,23. This program takes a phylogenetic tree as input and selects a designated number of representative sequences such that they are as close as possible to as many other sequences in the phylogenetic cluster they represent. Selected variants are summarized in Table 1 and broken down by which of the two gonococcal clusters they belong to.
With this, we considered what impact the cumulative addition of variants would have on the cross-reactivity of the antibody response against the known diversity of TbpB variants. The selected tbpB genes were codon optimized for expression in E. coli, synthesized, and then recombinantly expressed and purified from E. coli. Individual TbpBs, a bivalent composition consisting of a single representative TbpB from each of the two major gonococcal clusters (Ngo TbpBs 48627 and 41385), and a pentavalent composition that included all five representative TbpBs were formulated with aluminum hydroxide gel or Addavax (research grade MF59) for rabbit immunizations. Details of the formulations given to each rabbit, including antigen amount, can be found in Table 2. These were sub-cutaneously injected into rabbits three times (day 0, 21, and 42), and blood samples were acquired prior to each dose of vaccine and then again as a terminal bleed two weeks after the final dose. These serum samples were subsequently used to assess antiserum cross-reactivity against TbpB variants from both Ngo and Nme.
Formulations with two or more TbpBs elicit broadly cross-reactive antibodies with transferrin-blocking activity
Pre-immune serum, along with serum collected post 1–3 doses of vaccine were assessed from each rabbit using high throughput protein ELISAs24. To represent the natural diversity of neisserial TbpBs, the ELISA panel consisted of 18 recombinantly-expressed gonococcal TbpBs split between Ngo cluster 1 and Ngo cluster 2 (n = 13 representatives for Ngo cluster 1, n = 5 representatives for Ngo cluster 2), and 11 meningococcal TbpB variants (n = 1 isotype I TbpB, n = 10 isotype II TbpBs split across the four isotype II clusters (n = 3 for Nme cluster 1, n = 4 for Nme cluster 2, n = 3 for Nme cluster 3)) (Figs. 1B and 2A). Binding of HRP-labeled hTf (hTf-HRP) was used as an expression control since TbpBs binding ability to hTf-HRP indicates that they were present and properly folded on the ELISA plate (Fig. 2B). As expected, pre-immune serum (0 doses) elicited minimal reactivity to any TbpBs included in the panel, while the breadth and intensity of reactivity generally increased with each dose (Fig. 2A). Serum from rabbits that received a single antigen formulation had reactivity against a subset of the gonococcal TbpBs, with the formulation containing Ngo TbpB 48627, the main representative of the largest TbpB cluster (Ngo cluster 1), having the broadest reactivity after three doses. None of the single TbpB formulations elicited reactivity that extended to broadly recognize the meningococcal TbpBs. In contrast, formulations composed of either the bivalent TbpB formulation (Ngo 48627 + Ngo 41385) or the pentavalent TbpB formulation containing all five TbpBs were broadly reactive against the full gonococcal TbpB panel and, notably, extended the anti-TbpB cross-reactivity to include the meningococcal TbpBs. Despite containing three additional TbpB variants beyond the bivalent formulation, the pentavalent formulation did not substantially alter the breadth of cross-reactivity, particularly when focusing on reactivity with the Ngo TbpBs. Interestingly, there was minimal change in the reactivity of sera from post-two doses as compared to sera from post-three doses, indicating that there may be limited maturation of the antibody repertoire between the second and third dose in the time span evaluated.
A Heat map depicting ELISA-based reactivity of rabbit antiserum against a diverse panel of Neisserial TbpBs. Each row includes a different TbpB protein as a capture antigen, with the phylogenetic cluster indicated to the left. Each column represents either pre-immune or post 1, 2, or 3 vaccine dose serum for a single rabbit immunized with the vaccine formulation indicated below the x-axis. Increasing intensity indicates increasing immunoglobulin G specific for that TbpB variant. Background noise from empty vector control (maltose binding protein, MBP) has been subtracted. B hTf-binding to each TbpB as a plate control to quantification of properly folded (functional) TbpB. C Heat map depicting the ability of post 3 dose immune serum to block hTf binding to the TbpB capture antigen. Values are normalized to the signal obtained from no serum control. Each column represents a different rabbit immunized with the vaccine indicated below, or pooled pre-immune serum as a baseline control. Black represents a lack of hTf binding for the MBP empty vector control. AlOH aluminum hydroxide. Heat maps were generated using GraphPad Prism 10.1.1.
Aside from the significance of inducing broadly cross-reactive antibodies through TbpB immunization, it is postulated that the ability of antibodies to block the interaction between TbpB and hTf may lead to a reduced ability of the gonococcus to acquire iron. Consequently, this could contribute to protection via nutritional immunity, wherein the bacteria are starved of iron25,26. Therefore, we next assessed whether serum antibodies were able to inhibit HRP-conjugated hTf binding to TbpBs on the ELISA plate. Notably, as these immunizations were done in rabbits and rabbit Tf is not able to bind to Ngo TbpBs27, endogenous rabbit Tf is not expected to interfere with antibody development during immunization or to interfere during in vitro analysis. Since the quantity of pre-immune serum was limited, we utilized pooled pre-immune sera as the negative control and observed no reduction in HRP-hTf signal, as expected (Fig. 2C). In contrast, terminal serum led to a substantial reduction in HRP-hTf binding, as denoted by the lighter color on the heat map. Single TbpB formulations elicited variable levels of hTf blocking against the gonococcal TbpBs, with Ngo WHO Y eliciting the broadest hTf blocking among the single formulations, and Ngo 48627 eliciting the next best blocking ability. Among all the single TbpB formulations, blocking was highest against TbpBs that had the highest antibody reactivity, indicating that higher antibody binding leads to increased hTf blocking, as expected. Both the bivalent and the pentavalent TbpB formulations elicited very broad hTf blocking, and this again extended activity into a subset of the meningococcal TbpBs, which was not generally seen with single TbpB formulations. Again, consistent with the TbpB cross-reactivity, there was no substantial difference in TbpB-hTf blocking activity between the bivalent and pentavalent TbpB formulations.
Alum and emulsion-based formulations elicit similar cross-reactivity and transferrin-blocking pattern
As the bivalent composition containing Ngo 48627 and 41385 TbpBs elicited broad coverage against the panel of TbpBs, we next evaluated the effect of adjuvant on the breadth of reactivity on these two antigens as single formulations. While the alum-based Alhydrogel was used for our initial evaluation, we additionally tested the oil-in-water emulsion-based adjuvant Addavax (research grade MF59) to consider whether the coverage we observed remains consistent with a different choice of adjuvant. Serum from rabbits immunized with either Ngo 48627 or 41385 TbpB formulated with Addavax was compared to serum from the rabbits immunized with the same TbpBs formulated with Alhydrogel using the same assays as Fig. 2. As can be seen in Fig. 3A, the choice of adjuvant in the formulation did not substantially impact the breadth or pattern of reactivity, though reactivity after one dose of vaccine was slightly higher in sera from the rabbits that received the formulations containing Addavax. Similarly, the adjuvant did not substantially influence the hTf blocking activity of the antiserum, and still only elicited blocking against gonococcal TbpBs (Fig. 3B). The hTf blocking activity was again higher against the cluster that the TbpB originated from, as expected, and was primarily seen after two and three doses of vaccine.
A Heat map depicting the reactivity of rabbit antiserum against a broad ELISA panel of Neisserial TbpBs. Each row includes a different TbpB capture antigen, with the phylogenetic cluster indicated to the left. Each column represents either pre-immune or post 1, 2, or 3 dose serum for a single rabbit immunized with the vaccine indicated below the x-axis. Increasing intensity indicates increasing immunoglobulin G specific for that TbpB variant. Background noise from the empty vector (MBP) has been subtracted. B Heat map depicting the ability of post 1, 2, and 3 dose immune serum to block hTf binding by the TbpB capture antigen. Values are normalized to the signal obtained from no serum control. Each column represents a different rabbit immunized with the vaccine formulation indicated below. Black represents a lack of hTf binding for the MBP empty vector control. AlOH aluminum hydroxide. Heat maps were generated using GraphPad Prism 10.1.1.
TbpB-based compositions provide broad recognition of clinically relevant gonococcal isolates
To consider the ability of a TbpB-based vaccine to elicit antibodies that recognize the gonococcal surface, a whole-cell ELISA panel was developed to assess serum reactivity against 36 gonococcal strains, including three commonly used lab strains (MS11, FA1090, and FA19) and the WHO reference panel strains28, which contain some of the TbpB variants assessed in Figs. 2 and 3, along with disease isolates acquired from Public Health Ontario (Canada) and a collection of gonococcal strains acquired from endocervical swabs from female sex workers in Nairobi, Kenya (Kenyan Ngo Collection)29. Additional information regarding the disease isolates can be found in Supplementary Table 2. Included in the Public Health Ontario strains are recent isolates that lead to mucosal infections as well as disseminated gonococcal infections. Strains were grown under iron restricted conditions via chelation by deferoxamine to induce expression of the transferrin receptor and then heat killed prior to adherence to ELISA plates via drying. Binding of HRP-hTf was used as a control for receptor expression (Fig. 4B), although this control cannot differentiate between TbpA and TbpB expression based upon hTf binding alone. More variability in hTf binding is apparent here compared to what was seen with the recombinantly expressed TbpB proteins in Figs. 2 and 3. This may reflect inherent strain-to-strain differences in expression and/or different levels of surface transferrin receptor expression in the culture conditions employed to grow the bacteria used to coat the ELISA wells. Regardless, all strains showed some level of hTf binding (Fig. 4B), and all showed some reactivity with the TbpB-specific antisera (Fig. 4A).
A Heat map depicting the reactivity of rabbit antiserum against a broad ELISA panel consisting of heat-inactivated whole bacteria. Each row is a different gonococcal isolate, with the originating collection indicated to the left. Each column represents either pre-immune, or post 1, 2, or 3 dose serum for a single rabbit immunized with the vaccine indicated below the x-axis. AlOH, aluminum hydroxide. B hTf-binding to each isolate is used as a readout for surface expression of the TbpAB receptor. Heat maps were generated using GraphPad Prism 10.1.1. Arrows indicate strains that were used for serum bactericidal assays shown in Fig. 5.
Whole-cell reactivity was variable based on the immunizing TbpB, with serum raised against Ngo WHO Y TbpB showing the broadest reactivity of the single protein formulations, while reactivity with either the bivalent or pentavalent TbpB formulations showed broad reactivity after only two doses of vaccine (Fig. 4A). Interestingly, there was high reactivity by the bivalent formulation against the clinical isolates (Public Health Ontario and Kenyan Collections) that was not improved by the inclusion of the additional three TbpBs present in the pentavalent formulation. The specificity of immune serum to TbpB was confirmed by performing Western blots on whole cell lysates from a subset of gonococcal strains using post-three dose serum against the bivalent TbpB formulation. Pre-immune serum from the same rabbit and isogenic knockouts of TbpB and TbpA/TbpB in FA19 Ngo were additionally included as baseline controls (Supplementary Fig. 1A). Notable considering the potential for lipooligosaccharide (LOS) sialylation to mask some surface antigens, growth of N. gonorrhoeae in CMP-NANA did not influence immune sera recognition of TbpB on the bacterial surface (Supplementary Fig. 2).
TbpB-based compositions elicit bactericidal antibodies
Eight gonococcal strains, consisting of three common laboratory strains (MS11, FA1090, and FA19) and five WHO strains (WHO G, P, W, X, and Y), were chosen to assess the bactericidal activity of immune rabbit serum (Fig. 5). These strains were chosen to include representatives from each gonococcal TbpB cluster, porin type (which may influence gonococcal susceptibility to the bactericidal activity of human serum30), as well as to account for the varying degree of antibody reactivity and hTf binding (as denoted by arrows in Fig. 4). Normal human serum (NHS) without antibody depletion was utilized as the source of complement in the bactericidal assays. Prior to conducting these assays, NHS was titrated (0.5–15%) against individual strains to determine the highest percentage of NHS that did not impede Ngo recovery in the absence of an exogenous antibody source. Thus, bactericidal assays were conducted on strains that are highly resistant (FA1090 and WHO G, 15% NHS), intermediate resistant (WHO P, 5% NHS), and sensitive (MS11, FA19, WHO W, WHO X, and WHO Y, 2.5% NHS) to NHS. Strains were grown with a high concentration of an iron chelator to mimic in vivo conditions and ensure optimal TbpB expression.
A–H Different N. gonorrhoeae strains were tested for bactericidal killing with the porin type, TbpB cluster, and percentage of normal human serum (NHS) used in the assay indicated above each panel. Bars represent serum bactericidal activity (SBA) Log2 titer (50% killing compared to no antibody control) for pooled pre-immune serum (black), post-dose 3 serum from rabbits immunized with a single (gray), bivalent (pink), or pentavalent (teal) TbpB variant composition. Dotted line indicates the highest dilution tested, where bars that exceed that line indicate that <50% killing was not obtained at any serum dilution tested.
High levels of bactericidal activity were seen in sera from rabbits immunized with any one of the single TbpB formulations when tested against Ngo FA1090, WHO G, WHO W, and WHO Y (Fig. 5B, D, F, and H). The remaining strains (MS11, FA19, WHO P, and WHO X) were killed by some single TbpB-based formulations but not others (Fig. 5A, C, E, and G). Notably, sera from bivalent and pentavalent TbpB formulations elicited bactericidal activity against all eight strains. The bivalent TbpB formulation elicited bactericidal activity one dilution lower than the pentavalent formulation for four strains tested (MS11, WHO G, WHO W, and WHO Y), equally to the pentavalent formulation for two strains tested (FA1090 and FA19), and better compared to the pentavalent formulation for the final two strains tested (WHO P and WHO X). These data indicate that a bivalent TbpB formulation is sufficient to elicit broad bactericidal antibody coverage against diverse gonococcal strains. To ensure that the bactericidal activity was TbpB specific, we utilized a TbpB knockout strain, which was resistant to killing by post-dose 3 bivalent serum antibodies while the parental FA19 strain was susceptible (Supplementary Fig. 1B). Also notable is that strains WHO W and WHO X had minimal binding to rabbit serum in the whole cell ELISAs but were still effectively killed in the bactericidal assays, indicating that a combination of reactivity and functional assays are needed to fully understand potential effector functions of antibodies elicited by different vaccine formulations.
Bivalent TbpB composition accelerates clearance of heterologous gonococcal strains from the lower genital tract of female mice
As the bivalent TbpB formulation was sufficient to elicit broad cross-reactivity against the full panel of TbpBs, against the panel of gonococcal strains, and at eliciting bactericidal antibodies that recognized and killed all eight gonococcal strains tested, we selected this composition to evaluate protection in the female genital tract colonization model in mice. Ngo 48627 and Ngo 41385 TbpBs were formulated together with Addavax, the emulsion-based adjuvant. Female C57Bl/6 mice were immunized three times intra-peritoneally with the bivalent formulation or with Addavax alone and serum samples were collected two weeks after each dose. Immunogenicity for each mouse was tested against the two-component TbpBs, Ngo 41385 (Fig. 6A) and Ngo 48627 (Fig. 6B). A dose-dependent increase in TbpB-specific antibody was observed, with slightly higher reactivity obtained against 41385 TbpB compared to 48627 TbpB. Despite mouse-to-mouse variability, reactivity was observed across the full panel of gonococcal TbpBs (Fig. 6C). However, in contrast to the reactivity seen by the immune rabbit serum raised to the bivalent TbpB formulation, minimal reactivity was seen against the meningococcal TbpBs. Consistent with the response in rabbits, the bivalent TbpB formulation elicited broadly effective hTf-blocking antibodies in mice compared to the adjuvant control (Fig. 6D) when tested against a panel of 15 gonococcal TbpBs.
A, B Detection of anti-TbpB IgG in serum of C57BL/6 mice against each immunizing antigen after 1–3 doses of the bivalent (Ngo 48627 + Ngo 41385) TbpB vaccine. Separate ELISAs were performed using 41385 TbpB (A) and 48627 TbpB (B) protein-coated plates to assess the IgG response against each antigen. N = 23 for TbpB vaccinated cohort, N = 6 for adjuvant controls. Bars represent mean, error bars represent standard error. C Cross-reactivity of terminal serum from bivalent TbpB immunized mice against a broad panel of Neisserial TbpB proteins belonging to different clusters. Optical density (OD) values were normalized with Ngo 41385 TbpB and MBP (empty vector) set at 1 and 0, respectively, with any negative values (signal below empty vector) plotted as 0. N = 10 serum samples from mice immunized with bivalent TbpB formulation evaluated. Bars represent mean, error bars represent standard error. Inset, ELISA plate capture control using hTf-HRP binding as a readout to detect folded TbpB. Mean ± standard deviation of three technical replicates. D Human transferrin blocking ability of immune mouse serum. Serum samples from N = 15 mice/group were added to wells coated with 15 different gonococcal TbpBs. hTf-HRP was added to the probe for hTf blocking. OD was normalized to the no serum control for each TbpB-hTf signal. Each dot represents serum from one mouse immunized with either the bivalent TbpB formulation (pink bars, squares) or Addavax only (white bars, circles). ** represents p < 0.01, *** represents p < 0.001, **** represents p < 0.0001. Non-significant differences in panels A and B are not shown. Serum immunogenicity was compared between doses by 2 way ANOVA with Sidak’s multiple comparisons test comparing matched samples. hTf blocking compared with two-way ANOVA with Sidak’s multiple comparisons test comparing normalized OD from sera from bivalent immunized mice versus Addavax immunized mice against each captured TbpB.
Mice that received the bivalent TbpB formulation were randomized into two groups and infected vaginally with one of two heterologous strains of N. gonorrhoeae: either Ngo WHO L, which contains a TbpB 94.0% identical to Ngo 48627, or Ngo WHO F, which contains a TbpB 86.7% identical to Ngo 41385. Group size distribution for each challenge arm of the study can be found in Table 3. These strains were selected as being the closest match to the selected immunogens from the WHO panel of strains which was already utilized for murine infections by our group. Mice were followed daily to establish the bacterial burden and duration of colonization. Mice that received the bivalent TbpB composition decreased the median duration of colonization from 12 (Ngo WHO L, Fig. 7A, B) or 10 (Ngo WHO F, Fig. 7F, G) days for mice that received adjuvant only down to 4 days for both strains, though the difference in clearance was not statistically significant. The recovered CFU from Ngo WHO F was lower in mice that received the bivalent composition compared to mice that received adjuvant only (Fig. 7J), however no difference in overall recovered bacteria was seen for Ngo WHO L (Fig. 7E). While minimal pre-challenge serum was available, terminal sera from uninfected mice that did not stage in (did not enter the requisite estrus phase) for the infection studies for each of the bivalent formulation and Addavax only control were tested and found to be bactericidal against the two challenge strains, WHO L and WHO F (Supplementary Fig. 4). Growth in CMP-NANA did not alter the bactericidal titer elicited by pooled, post-challenge mouse serum from animals that received the bivalent formulation for either Ngo WHO L or WHO F, while pooled sera from adjuvant immunized animals lacked bactericidal activity under either growth condition (Supplementary Fig. 5). Although a robust polymorphonuclear leukocyte (PMN) influx has traditionally not been observed in the C57BL/6 breed of mice31, which were used in these protection studies, we have recently shown that opsonophagocytosis is the main effector mechanism that confers protection against nasal meningococcal colonization after repeat exposure32. Preliminary opsonophagocytosis studies utilizing bivalent TbpB-rabbit serum and a murine macrophage cell line (RAW 264.7) suggest both WHO L and WHO F are susceptible to opsonophagocytic killing (Supplementary Fig. 3). Taken together, our various in vitro assays using immune serum from mice, rabbits, or both species support multiple potential modes of action by the bivalent TbpB formulation, including complement-mediated lysis, opsonophagocytic killing, and iron starvation may be contributing to protection.
A, F Percentage of vaccinated and adjuvant controls that remain colonized over a period of 15 days after vaginal infection with a Ngo cluster 1 strain (WHO L) and cluster 2 strain (WHO F), respectively. N = 10–12 animals/group. Differences were not significant using Log-rank (Mantel–Cox) statistical test. B, G Graph depicting the number of days taken for 50% of animals to clear infection. Error bars depict the interquartile range (IQR). C, D, H, I Daily CFU recovery from the lower genital tract of infected mice. Thin gray lines depict individual animals, while thicker lines represent the average CFU recovered per day. E, J Box plot showing total bacterial burden in each animal. Bioburden was estimated using the area under the curve of Log10 transformed CFU data. *p < 0.05 obtained using two-tailed t-test, ns denotes not significant. Statistical analysis was done using GraphPad Prism 10.1.1.
Discussion
The rational selection of immunogens for inclusion into a vaccine formulation is a fundamental step toward developing a broadly cross-protective vaccine. The use of surface-exposed protein antigens as targets for vaccine development allows for the targeting of virulence factors and other proteins essential for the survival of the pathogen. However, due to the accessibility of these targets to the immune system, these proteins often evolve to exhibit high levels of antigenic diversity.
In this study, we demonstrate that a bivalent TbpB formulation is sufficient to elicit broad cross-reactivity against a panel of gonococcal TbpBs that encompasses the natural diversity of this protein, and against an equally diverse panel of Ngo strains. Expanding the vaccine formulation to include five gonococcal TbpBs did not further increase the breadth of cross-reaction, indicating that one rationally selected protein variant is sufficient to elicit a response against each of the two Ngo TbpB clusters. This supports the utility of our custom software Navargator to rationally select variants to include in a vaccine formulation. With this, we were able to leverage sequence diversity information by performing clustering using the phylogenetic distance between the variants. This analysis allows us to deterministically define clades, and to select the ‘central’ variant from each.
In rabbits, the bivalent and pentavalent formulations each elicited a response that also recognized a subset of meningococcal TbpBs. However, this inter-bacterial species response was not apparent in mice. Regardless, our results suggest that the addition of rationally selected TbpBs originating from Nme strains would likely be more effective to expand this coverage for pan-pathogenic Neisseria protection. The bivalent TbpB formulation also elicited broad bactericidal activity against eight representative gonococcal strains, elicited antibodies able to block human transferrin from binding to gonococcal TbpBs in both mice and rabbits, and decreased the median duration of colonization by two heterologous gonococcal strains in the murine lower genital tract colonization model, supporting the protective capacity of this response.
Inclusion of multiple variants of a single immunogen is a strategy that has been successfully utilized in previous vaccine formulations, most notably the bivalent factor H binding protein (fHbp) vaccine (Bivalent rLP2086, also known as ‘Trumenba’, Pfizer), wherein one fHbp variant from each of subfamily A and subfamily B were included as lipidated recombinant proteins33. Inclusion of multiple proteins in a vaccine formulation increases the complexity of each phase of development, including process development and process characterization. Therefore, identifying the minimum number of variants in a formulation needed to elicit broad coverage becomes vital in developing a vaccine formulation that is feasible, scalable, and cost-effective. As shown here, there was minimal expansion in cross-reactivity, transferrin blocking ability, or bactericidal activity when expanding our two variant protein composition up to five, indicating that there is value in rational selection of immunogens and pre-clinical evaluation to reduce the burden of production.
One limitation to the use of small animal models in pre-clinical vaccine research for host-restricted pathogens is that these animals lack host-specific factors that are vital during natural infection. The gonococcus is exquisitely host-restricted and able to interact with an array of human factors, including those involved in adhesion (human CEACAMs34), immune evasion (human Factor H,35 human C4bp36, human IgA137), and nutrient acquisition (human transferrin38, human lactoferrin39, human calprotectin40, human psoriasin41). These factors are not present in small animal models, including mice or rabbits, which limits the extent that we are able to accurately model gonococcal infections and the response to vaccination with bacterial proteins able to specifically bind to these human factors. Previous studies using either transgenic mice expressing the human factor of interest (e.g. meningococcal fHbp immunization into mice expressing human Factor H42) or studies evaluating homologous receptors in the native host (immunization with Glaesserella parasuis TbpB into pigs25) have demonstrated that the use of non-binding mutant proteins are superior immunogens compared to wild type versions that maintain their ability to bind to the host protein. This is presumably because the native proteins will complex with their host-derived target when administered as a vaccine component. While beyond the scope of this study, our ongoing efforts are focussed on developing non-binding versions of the TbpBs included in the bivalent formulation that will potentially enhance immunogenicity and safety for use in humans.
The use of rabbits in this study allowed us to gather comparatively large volumes of serum from each animal. However, the trade-off to that was using only one rabbit per formulation for the initial evaluation due to constraints in space and cost. Animal-to-animal variability is always a concern as individual differences may skew the interpretation of the resulting data. To strengthen the preliminary results from rabbits, we utilized a follow-up experiment in mice with the selected bivalent formulation wherein groups of 24 mice received either the bivalent TbpB formulation or adjuvant (50% v/v Addavax) only, which were then randomized into two challenge groups. The bivalent TbpB formulation elicited a robust antibody repertoire in both mice and rabbits. This antibody response had the ability to broadly recognize and block human transferrin binding by a comprehensive panel of gonococcal TbpB variants, as well as initiate complement-dependent bactericidal lysis of serum-sensitive and highly serum-resistant strains. Leveraging immunization in two species allowed us to increase the confidence in the results, and lower genital tract infections in mice with two challenge strains has allowed us to interrogate not just the activity of immune sera against the proteins or strains of interest but to evaluate potential protection beyond antibody-mediated bacterial clearance. It is pertinent to note that our analyses in the current study have focused on humoral responses, including immunogenicity, bactericidal activity, and transferrin blocking ability of immune sera. It remains possible that individual TbpB variants might influence the cellular responses to infection since, for example, N. gonorrhoeae is adept at suppressing Th1 and Th2 adaptive responses during infection43, so this should be considered in future TbpB-based protection studies. Indeed, as the immune correlate of protection for gonococcal clearance remains elusive, the use of small animal challenge models remains a valuable tool for understanding vaccine efficacy prior to moving into human clinical trials.
Selection of adjuvants to elicit protection is a vital component of vaccine development. While a broad evaluation of adjuvants was beyond the scope of this study, the use of both aluminum hydroxide gel (Alhydrogel) and Addavax (research grade MF59) in rabbits with the two major immunogens of interest demonstrated that the overall breadth of response was not substantially different between the two formulations. Previous studies with TbpB have shown that formulation with alum-based adjuvants elicited higher levels of anti-TbpB serum IgG versus formulation with Addavax, however strength of the antibody response has not correlated with levels of protection44. While Addavax was utilized for the murine protection studies presented here, recent insights by our group44 underscore the necessity for additional optimization of the adjuvant to attain robust protection.
Together, the studies shown here demonstrate that broad gonococcal cross-protection can be obtained with a bivalent TbpB formulation selected rationally based on phylogenetic analysis of the diversity of this protein.
Methods
Bioinformatic analysis and selection of the TbpBs of interest
To generate the large Neisserial phylogenetic tree, we collected all TbpB sequences from the PubMLST database that were annotated as originating from either N. meningitidis or N. gonorrhoeae. We removed all sequences shorter than 500 or longer than 750 residues, those containing internal stop codons, and one sequence (ID 21071) that was radically different from all others and is likely a misannotation. We then added several sequences of prototype strains commonly used in various laboratories (MS11, FA1090, and FA19) as well as the WHO reference panel28. The signal peptides were removed, and duplicate sequences were filtered out, resulting in 1485 unique total sequences: 1178 from N. meningitidis and 307 from N. gonorrhoeae. 304 of the 307 sequences from N. gonorrhoeae were found in two monophyletic groups, and these were used to build the N. gonorrhoeae phylogenetic tree; the three outlier sequences were widely spread throughout the large tree and were not included as they were assumed to be misannotated or the result of horizontal exchange from N. meningitidis. An additional 21 N. gonorrhoeae TbpB sequences were collected from a more recent version of the PubMLST database, and these 325 sequences were used to build the N. gonorrhoeae phylogenetic tree.
Sequence alignments were performed using the E-INS-i algorithm with default parameters from MAFFT v7.47545. The phylogeny was reconstructed with RAxML v8.2.1246 using the PROTCATLG model of evolution.
To choose representative variants from N. gonorrhoeae, we started first by selecting representative variants from two clusters, Ngo 48627 as the large cluster representative and Ngo 41385 as the small cluster representative. We further used Navargator22,23 to identify several potential clustering patterns on the current tree and settled on five clusters as the smallest number that partitioned the tree effectively. This yielded our final list of five representative N. gonorrhoeae TbpB sequences: 27464, 48627, WHO Y, 27230, and 41385.
Cloning, expression, and purification of selected TbpBs
The five representative TbpB variants were codon optimized and synthesized as DNA fragments (GeneArt String, ThermoFisher) with 27 base pairs of complementary sequences matching our custom expression vector (codon optimized sequences available in Supplementary Table 1). The inserts were subcloned by overlap extension PCR cloning47 downstream and in-frame of an N-terminally hexa-histidine tagged maltose binding protein (MBP) and a cleavable TEV linker sequence. Escherichia coli MM294 was used to amplify the vectors and sequence verification was done by in-house Sanger sequencing (The Center for Applied Genomics, The Hospital for Sick Children). Protein expression was carried out in E. coli OverexpressionTM C43 DE3 cells (Sigma Aldrich), and freshly heat-shocked transformants were recovered in Luria-Bertani (LB) broth for 1 hour, shaking at 175 RPM and 37 °C, before selection by additional LB broth (3 mL) supplemented with ampicillin (100 μg/mL) and grown for a further 3 h. The starter culture was inoculated into 6 L of ZYP-5052 autoinduction media (ZY media supplemented with NPS and 5052) with ampicillin (50 μg/mL)48 at a 1:1000 dilution with the culture flasks shaken at 175 RPM at 37 °C for 18 h and then at 20 °C for additional 12–16 h.
Cells were harvested by centrifugation at 5000×g for 30 min at 4 °C, resuspended in binding buffer (50 mM Tris pH 8.0, 300 mM NaCl) supplemented with DNase, lysozyme, and protease inhibitors (cOmpleteTM Mini, Roche) and either lysed by sonication (S-450 Sonifier, Branson) or homogenization (EmulsiFlex-C3, Avestin). The resulting lysate was clarified by centrifugation at 16,000×g for 90 min at 4 °C, followed by 0.22 μm syringe filtration (Millipore). Clarified lysate was then recirculated and captured by nickel Sepharose affinity chromatography (HisTrap excel, Cytiva). The affinity column was then mounted onto an ÄKTA Purifier 100 (GE Healthcare), washed, and eluted with imidazole-containing buffer (50 mM Tris pH 8.0, 300 mM NaCl, 300 mM imidazole pH 7.4). Eluted fraction was dialyzed overnight into exchange buffer (50 mM Tris pH 8.0, 150 mM NaCl), and sample purity was assessed by 12% SDS–PAGE analysis and concentration by NanoDrop absorbance at 280 nm (ThermoFisher); TEV protease, purified in-house, was added to separate the MBP fusion partner from TbpB. The samples were then exchanged into a low salt buffer (50 mM Tris pH 8.0, 10 mM NaCl) and loaded onto an anion exchange column (HiTrapQ FF, Cytiva) and the fraction containing TbpB concentrated and passed through a size exclusion column (HiPrep 26/60 Sephacryl S-200 HR, Cytiva) to further remove any contaminants and exchanged into the final sample buffer (50 mM HEPES, 50 mM NaCl). A final polishing step with a strong anion exchange column (MonoQ 5/50 GL, Cytiva) was used to remove lipopolysaccharide, and the final sample was concentrated to 5 mg/mL in aliquots, flash-frozen by liquid nitrogen, and stored at −80 °C until use.
Rabbit immunizations
Rabbits were provided food and water ad libitum and handled humanely under animal use protocol 20012577, approved by the Animal Care Committee at the University of Toronto. New Zealand female rabbits (Charles River) were obtained at approximately three months of age and acclimated to the facility for one week. Rabbits 1–5, 8, and 9 (Table 2) received an initial dose of 100 μg of protein formulated with the respective adjuvant (38.5 μL of 2% Alhydrogel per 500 μL dose, Sigma Aldrich, cat. A8222 or 50% (v/v) Addavax, InvivoGen, cat. vax-adx-10) in 500 μL total volume, immunized subcutaneously. Two booster doses were given containing half the amount of protein as the initial vaccine at 3 and 6 weeks post-first dose. Rabbit 6 received the bivalent TbpB composition which was composed of 100 μg of each TbpB (dose 1) or 50 μg of each TbpB (doses 2 and 3) for formulations containing 200 and 100 μg total protein respectively, while Rabbit 7 received the pentavalent formulation composed of 50 μg of each TbpB (dose 1) or 25 μg of each TbpB (doses 2 and 3) for formulations containing 250 and 125 μg total protein, respectively (Table 2). Bleeds were collected 2.5 weeks after each dose and at euthanasia (two weeks after the third dose). Euthanasia was performed under isoflurane anesthesia where rabbits were exsanguinated, followed by intra-venous injection of T-61.
Protein ELISAs
Protein ELISAs were performed as previously described24. Briefly, TbpBs of interest were cloned into a custom His-Bio-MBP-Tev vector and transformed into E. coli strain ER2566. E. coli strains were grown overnight in 10 mL cultures of ZYP-5052 autoinduction media containing 100 μg/mL ampicillin and grown overnight at 37 °C with shaking. Lysates were obtained by centrifuging the E. coli cultures (3000×g for 5 min) and resuspending the pellets in buffer (50 mM NaHP2PO4, 300 mM NaCl, 10 mM imidazole, pH 8.0). 1 mm glass beads were added to the suspension and cells were disrupted by shaking in a cell disruptor. Homogenate was centrifuged for 20 min at 4 °C at 14,000×g. Supernatant was diluted 1:5 in PBS and added to 384 well ELISA plates (VWR, cat. CA62409-064) that had been previously coated with 1 μg/mL NeutrAvidin (Thermo Fisher, cat. A2666, coated overnight at 4 °C) and blocked with 5% bovine serum albumin (BSA, BioShop Canada, cat. ALB001). Protein lysate was coated for 2 h at room temperature, washed, and then plates were blocked again with 5% BSA for 1 h. Serum was added overnight at 4 °C at a 1:8000 dilution for rabbit sera and at a 1:10,000 dilution for mouse sera. The following day, plates were washed, the relevant secondary antibody was added (1:10,000 dilution of goat anti-mouse IgG H&L (HRP) (Abcam, cat. ab6789) or 1:10,000 dilution of Peroxidase AffiniPure goat anti-rabbit IgG (H&L) (Jackson ImmunoResearch, cat. 111-035-144)) and incubated at room temperature for two hours. Plates were washed and developed with KPL SureBlue TMB Microwell Peroxidase Substrate (SeraCare, cat. 5120-0077). Reactions were quenched with 2 N H2SO4 and read at 450/570 nm.
Gonococcal strains
N. gonorrhoeae strains used for these studies include laboratory strains FA19, FA19ΔtbpB, FA19ΔtbpAB (obtained from Dr. Cynthia Nau Cornelissen), MS11, and FA1090, the WHO reference strains (obtained from Public Health England culture collection)28, clinical isolates obtained from Public Health Ontario which consists of both mucosal and disseminated strains (described in Supplementary Table 2), and a collection of gonococcal strains acquired from endocervical swabs from female sex workers in Nairobi, Kenya (Kenyan Ngo Collection, described in Supplementary Table 2)29,49.
Whole-cell ELISAs
Whole-cell ELISAs were performed with a panel of representative strains of N. meningitidis and N. gonorrhoeae. Strains were grown overnight on GC agar supplemented with Isovitalex (BD, cat. B11876) and 10 μM deferoxamine mesylate salt (Sigma, cat. D9533) at 37 °C with 5% CO2 and then restreaked onto several fresh GC agar supplemented with Isovitalex and 10 μM deferoxamine mesylate salt for another overnight growth. Bacteria were resuspended in PBS and heat-killed at 56 °C for a minimum of 60 min. To evaluate a potential effect of sialylation on TbpB accessibility, strains were grown overnight on GC agar supplemented with Isovitalex, sub-cultured in RPMI 1640 medium (Wisent, cat. 350-000-CL) at starting OD550 of ~0.25 and allowed to grow at 37 °C with shaking at 160 rpm. After 1.5 h, the culture was split into two—with or without the addition of CMP-NANA to a final concentration of 50 µg/mL (Sigma, cat. 8271) and allowed to grow for an additional 1.5 h. Bacteria were pelleted (3000×g for 10 min), and pellets were resuspended in PBS for heat inactivation as above. 384-well plates were coated with 20 µL of heat-killed bacteria diluted in PBS at an OD550 of 0.2 and dried in a biosafety cabinet until inactivation of bacteria was confirmed. Plates were blocked with 5% BSA and ELISAs were performed as described above.
Transferrin blocking assays
Transferrin blocking assays were done using the protein-based ELISA protocol described above with mouse or rabbit serum added in a 1:10 dilution; and instead of probing with goat-anti-mouse or anti-rabbit IgG, horse radish peroxidase conjugated hTf (HRP-hTf, Rockland, cat. 009-0334) was added at a dilution of 1:1000 and incubated at room temperature for 30 min. Plates were developed as above, and a decrease in signal indicated the blocking of transferrin binding. Signal was compared to transferrin binding to the protein in question with no serum present.
Serum bactericidal assays
Serum bactericidal assays (SBA) were performed with rabbit and mouse serum against strains of N. gonorrhoeae with varying concentrations of pooled normal human serum (NHS) (Sigma, cat. H4522, Lot #SLCD1946), each using the highest concentration of NHS (range 2.5–15%) that the strain was intrinsically resistant to in the absence of any antibody source. Strains were grown overnight on GC agar supplemented with Isovitalex at 37 °C with 5% CO2. The next day strains were restreaked onto GC agar supplemented with Isovitalex and 100 µM deferoxamine mesylate salt (Sigma, cat. D9533) for 4 h. (Note: this concentration is 10× higher than what was used for preparing ELISA plates, allowing for optimal TbpB expression albeit less robust growth.) Bacteria were collected off the plates using Dacron swabs, resuspended in RPMI 1640 media and OD550 was measured. Assay was set up in 40 µL total volume using ~200–1000 colony-forming units (CFU) of bacteria suspended in RPMI 1640 and allowed to incubate with 2-fold serial dilutions of heat-inactivated rabbit/mouse serum for 5 min, followed by the addition of the indicated NHS percentage for another 60 min at 37 °C with 5% CO2. 10 µL was plated on either GC-Isovitalex or GC-Kelloggs to enumerate the number of viable bacteria. The highest dilution at which 50% killing was observed relative to no antibody control (NHS only) was reported as the SBA titer. SBAs using uninfected mouse serum against WHO L and WHO F were performed using the same protocol as above, with varying percentages of NHS (between 2.5% and 25%) shown in Supplementary Fig. 4. For SBAs using pooled terminal (post-infection) mouse serum to determine the impact of sialylation in Supplementary Fig. 5, the same protocol was used as above using NHS as the complement source, but with Ngo grown under a different condition. For the sialylation SBAs, strains were first grown overnight on GC agar supplemented with Isovitalex, sub-cultured in RPMI 1640 medium (Wisent, cat. 350-000-CL) at starting OD550 of ~0.3 and allowed to grow at 37 °C with shaking at 160 rpm. After 2.5 h, the culture was split into two - with or without the addition of CMP-NANA to a final concentration of 50 µg/mL (Sigma, cat. 8271) and allowed to grow for an additional 1.5 h. For a confirmatory SBA with mouse serum using IgG/IgM depleted NHS, Ngo was grown overnight on GC agar supplemented with Kellogg’s at 37 °C with 5% CO2, followed by overnight growth in RPMI 1640 at 37 °C with shaking at 160 rpm. The next day, bacteria were harvested (3000×g for 10 min) and resuspended in RPMI 1640. To allow opsonization, ~500 CFUs were incubated with various concentrations (1:20, 1:40, 1:80, 1:200, 1:400) of serum for 20 min, followed by the addition of 12% pooled human complement depleted of IgG/IgM (Pel Freez, cat. 34010, Lot 22344) and incubated for 60 min at 37 °C with 5% CO2. The highest dilution at which 50% killing was observed relative to no antibody control in 60 min was reported as the SBA titer. All mouse/rabbit test sera used in SBAs had been heat-inactivated at 56 °C for 30–45 min to remove endogenous complement activity. For each assay, bacteria in RPMI 1640 alone (no antibody and no complement) were plated at the start of the assay and after 60 min to ensure there was no loss in viability in assay media alone or in the complement source alone.
Opsonophagocytosis assays
Opsonophagocytic assays were performed with RAW Dual cells (RAW 264.7 murine macrophage cell line stably transfected with two reporter genes). Briefly, cells grown in DMEM + 10% FBS (fetal bovine serum) + GlutaMAX supplement (Life Technologies, cat. 35050061) were seeded at a high density of 105 cells/well in a 96 well tissue culture treated plate (Fisher Scientific, cat. 08-772-2C) the day before. Ngo WHO L and WHO F were grown overnight on GC agar supplemented with Isovitalex at 37 °C with 5% CO2. On the day of the assay, Ngo was restreaked onto GC agar supplemented with Isovitalex and 100 µM deferoxamine mesylate salt for 4 h to induce TbpB expression. Bacteria was swabbed off the plates using a Dacron swab, resuspended in RPMI 1640 media, and OD550 was measured and diluted to a concentration of 107 CFU/mL. 100 µL aliquots of bacterial suspension were incubated for 20 min at 37 °C with 5% CO2 with heat-inactivated post-dose 3 bivalent TbpB or pre-immune rabbit serum from the same animal. 10 µL of opsonized bacterial suspension or untreated control (multiplicity of infection (MOI) of 1) was added per well to the 96-well plate containing RAW 264.7 cells for which media had been replaced with 90 µL RPMI 1640 after 2× washes. Plates were centrifuged at 500×g for 5 min to synchronize infection, and incubated for 60 min at 37 °C with 5% CO2. 100 µL of 2% saponin (Sigma, cat. S7900) dissolved in RPMI 1640 was added to lyse cells for an additional 10 minutes at 37 °C with 5% CO2, after which the bacterial/cellular suspension was pipetted up and down several times and serially diluted and plated to enumerate viable bacteria. CFU counts of serum opsonized Ngo were normalized to non-opsonized untreated Ngo and reported as a percentage. Untreated and opsonized bacteria (prior to adding to cells) were also enumerated to ensure consistency in the infectious dose across treatments.
Western blots
Gonococcal strains were grown overnight on GC agar supplemented with Isovitalex at 37 °C with 5% CO2, followed by an additional subculture on GC agar supplemented with Isovitalex and 100 µM deferoxamine mesylate salt for 4 hours to induce TbpB expression. Bacteria were collected off the plate with a Dacron swab, resuspended in PBS and then centrifuged at 3000×g for 10 min. Pellets were resuspended in 2× sodium dodecyl sulfate (SDS) buffer (0.125 M Tris, 4% (w/v) SDS, 20% (v/v) Glycerol and 0.01% (w/v) bromophenol blue) and boiled for 10 min and then stored at −20 °C until use. Samples were resolved on 10% SDS–PAGE and transferred to PVDF membranes. Membranes were blocked in 5% skim milk in Tris-buffered saline with 0.05% Tween (TBST) for 1 h, followed by overnight incubation at 4 °C with gentle agitation in 1:5000 dilution of either pre-immune or post-dose 3 bivalent TbpB rabbit serum in 1% skim milk. After 3 TBST washes, peroxidase-conjugated goat-anti-rabbit IgG H + L (Jackson ImmunoResearch, cat. 111-035-144) was used at 1:10,000 dilution in 1% skim milk for 2 h at room temperature, washed 3 times with TBST and developed with the Novex™ ECL Chemiluminescent Substrate Reagent Kit (Invitrogen).
Mouse immunizations and genital tract challenge
Mouse studies were performed under the animal use protocol 20011775, approved by the Animal Care Committee at the University of Toronto. Female C57BL/6 mice (4–5 weeks of age at time of purchase, Charles River) were acclimated to the animal facility for one week, where they received water and rodent chow ad libitum and were housed in specific pathogen-free conditions. Mice were randomly assigned to receive either the bivalent TbpB formulation (12.5 µg 48627 TbpB, 12.5 µg 41385 TbpB in sterile PBS, formulated with 50% v/v Addavax (InvivoGen) in 100 µL total volume) or adjuvant alone (50% v/v Addavax in PBS). Mice were immunized three times at three-week intervals via intraperitoneal injection. Vaccines were well tolerated, and no adverse reactions were noted. Saphenous bleeds were performed two to three weeks after each dose and serum was stored at −20 °C until analysis.
Two weeks after the final dose, mice were challenged via the lower genital tract challenge model50. Briefly, the stage of the estrus cycle was tracked via evaluation of cell morphology of vaginal lavage samples under light microscopy. When mice entered diestrus (denoted as day −2 relative to infection), they were given β-estradiol (0.5 mg in 200 µL sub-cutaneous injection, Sigma-Aldrich, cat. E4389) to lock them in estrus and antibiotics (intraperitoneal 0.6 mg vancomycin (BioShop Canada, cat. VAN990.5) and 2.4 mg streptomycin (BioShop Canada, cat. STP101.5), as well as 0.04 g/100 mL trimethoprim (Sigma-Aldrich, cat. T7883) in the drinking water ad libitum) to control overgrowth of commensal flora. Mice received two doses of injectable antibiotics the following day (day −1 relative to infection). On day 0, mice were infected intra-vaginally with approximately 107 CFU of either N. gonorrhoeae WHO L or N. gonorrhoeae WHO F (obtained from Public Health England28, streptomycin resistance induced for compatibility with this model). Mice received two additional doses of β-estradiol (one on the day of infection and another on day +2 relative to infection) and daily antibiotic injections for the duration of the study. Vaginal lavage samples (15 µL PBS++, containing calcium and magnesium, diluted in 60 µL) were collected daily and dilutions were plated on GC agar supplemented with Kellogg’s and VCNT (vancomycin, colistin, nystatin, and trimethoprim) inhibitor (BD BBLTM VCNT Inhibitor (Thayer and Martin), Fisher Scientific, cat. B12408) for selection of N. gonorrhoeae. Mice were considered no longer colonized when a minimum of three consecutive lavages had no recovery of N. gonorrhoeae. Group sizes for each vaccine group and challenge strain can be found in Table 3.
Upon completion of the infection study, mice were humanely euthanized by CO2 overdose followed by exsanguination. Terminal serum was collected by cardiac puncture and stored at −20 °C until analysis. The mouse challenge study was performed as a single study with two challenge arms, where mice were infected on different days in three to four cohorts, depending on the challenge strain.
Graphing and statistical analyses
All graphing and statistical analyses were performed in Prism 10.1.1. Phylogenetic tree figures were created using Navargator22.
Data availability
Sequence data utilized in this study were primarily taken from the PubMLST database (https://pubmlst.org/), supplemented with 52 sequences from commonly used strains of N. meningitidis and 15 from N. gonorrhoeae, as detailed in the methods section. The relevant sequence and tree files are available from the corresponding author upon reasonable request. The data that support the rabbit and mouse serology and challenge studies are available from the corresponding author upon reasonable request.
References
Unemo, M., Golparian, D. & Eyre, D. W. Antimicrobial resistance in Neisseria gonorrhoeae and treatment of gonorrhea. In Methods in Molecular Biology 37–58 (Humana, New York, NY, 2019).
Eyre, D. W. et al. Gonorrhoea treatment failure caused by a Neisseria gonorrhoeae strain with combined ceftriaxone and high-level azithromycin resistance, England, February 2018. Eurosurveillance 23, 1800323 (2018).
Pleininger, S. et al. Extensively drug-resistant (XDR) Neisseria gonorrhoeae causing possible gonorrhoea treatment failure with ceftriaxone plus azithromycin in Austria, April 2022. Eurosurveillance 27, 2200455 (2022).
Petousis-Harris, H. et al. Effectiveness of a group B outer membrane vesicle meningococcal vaccine against gonorrhoea in New Zealand: a retrospective case-control study. Lancet 390, 1603–1610 (2017).
Leduc, I. et al. The serogroup B meningococcal outer membrane vesicle-based vaccine 4CMenB induces cross-species protection against Neisseria gonorrhoeae. PLoS Pathog. 16, e1008602 (2020).
Wang, B. et al. 4CMenB sustained vaccine effectiveness against invasive meningococcal B disease and gonorrhoea at three years post programme implementation. J. Infect. 87, 95–102 (2023).
Wang, B. et al. Effectiveness and impact of the 4CMenB vaccine against invasive serogroup B meningococcal disease and gonorrhoea in an infant, child, and adolescent programme: an observational cohort and case-control study. Lancet Infect. Dis. 22, 1011–1020 (2022). 1474-4457 (Electronic).
Boslego, J. W. et al. Efficacy trial of a parenteral gonococcal pilus vaccine in men. Vaccine 9, 154–162 (1991).
Anderson, J. E., Sparling, P. F. & Cornelissen, C. N. Gonococcal transferrin-binding protein 2 facilitates but is not essential for transferrin utilization. J. Bacteriol. 176, 3162–3170 (1994).
Cornelissen, C. N. & Sparling, P. F. Binding and surface exposure characteristics of the gonococcal transferrin receptor are dependent on both transferrin binding proteins. J. Bacteriol. 178, 1437–1444 (1996).
Anderson, J. E., Hobbs, M. M., Biswas, G. D. & Sparling, P. F. Opposing selective forces for expression of the gonococcal lactoferrin receptor. Mol. Microbiol. 48, 1325–1337 (2003).
Cornelissen, C. N. et al. The transferrin receptor expressed by gonococcal strain FA1090 is required for the experimental infection of human male volunteers. Mol. Microbiol. 27, 611–616 (1998).
Adamiak, P., Calmettes, C., Moraes, T. F. & Schryvers, A. B. Patterns of structural and sequence variation within isotype lineages of the Neisseria meningitidis transferrin receptor system. Microbiol. Open 4, 1–14 (2015).
Baarda, B. I., Zielke, R. A., Holm, A. K. & Sikora, A. E. Comprehensive bioinformatic assessments of the variability of Neisseria gonorrhoeae vaccine candidates. mSphere 6, e00977–00920 (2021).
Harrison, O. B., Maiden, M. C. & Rokbi, B. Distribution of transferrin binding protein B gene (tbpB) variants among Neisseria species. BMC Microbiol. 8, 66 (2008).
Rokbi, B. et al. Evaluation of recombinant transferrin binding protein B variants from Neisseria meningitidis for their ability of induce cross reactive and bactericidal antibodies against a genetically diverse collection of serogroup B strains. Infect. Immun. 65, 55–63 (1997).
Rokbi, B. et al. Allelic diversity of the two transferrin binding protein B gene isotypes among a collection of Neisseria meningitidis strains representative of serogroup B disease: implication for the composition of a recombinant TbpB-based vaccine. Infect. Immun. 68, 4938–4947 (2000).
Calmettes, C., Alcantara, J., Schryvers, A. B. & Moraes, T. F. The structural basis of transferrin iron sequestration by transferrin binding protein B. Nat. Struct. Mol. Biol. 19, 358–360 (2012).
Ling, J. M., Shima, C. H., Schriemer, D. C. & Schryvers, A. B. Delineating the regions of human transferrin involved in interactions with transferrin binding protein B from Neisseria meningitidis. Mol. Microbiol. 77, 1301–1314 (2010).
Retzer, M. D., Yu, R.-H., Zhang, Y., Gonzalez, G. C. & Schryvers, A. B. Discrimination between apo and iron-loaded forms of transferrin by transferrin binding protein B and its N-terminal subfragment. Microb. Pathog. 25, 175–180 (1998).
Fegan, J. E. et al. Utility of hybrid transferrin binding protein antigens for protection against pathogenic Neisseria species. Front. Immunol. 10: eCollection (2019).
Curran, D. Navargator (accessed 2 March 2023); https://www.compsysbio.org/navargator.
Curran, D. M. et al. Phylogeny-based selection of representative variants with Navargator: proof of principle using humoral cross-reactivity data from two immunization studies. Preprint at bioRxiv https://doi.org/10.1101/2024.09.08.611914 (2024).
Fegan, J. E., Yu, R. H., Islam, E. A. & Schryvers, A. B. Development of a non-biased, high-throughput ELISA for the rapid evaluation of immunogenicity and cross-reactivity. J. Immunol. Methods 493, 113037 (2021).
Frandoloso, R. et al. Nonbinding site-directed mutants of transferrin binding protein B enhances their immunogenicity and protective capabilities. Infect. Immun. 83, 1030–1038 (2015).
Guizzo, J. A. et al. The amino acid selected for generating mutant TbpB antigens defective in binding transferrin can compromise the in vivo protective capacity. Sci. Rep. 8, 7372 (2018).
Lee, B. C. & Schryvers, A. B. Specificity of the lactoferrin and transferrin receptors in Neisseria gonorrhoeae. Mol. Microbiol. 2, 827–829 (1988).
Unemo, M. et al. The novel 2016 WHO Neisseria gonorrhoeae reference strains for global quality assurance of laboratory investigations: phenotypic, genetic and reference genome characterization. J. Antimicrob. Chemother. 71, 3096–3108 (2016).
Fudyk, T. C. et al. Genetic diversity and mosaicism at the por locus of Neisseria gonorrhoeae. J. Bacteriol. 181, 5591–5599 (1999).
Lewis, L. A. & Ram, S. Complement interactions with the pathogenic Neisseriae: clinical features, deficiency states, and evasion mechanisms. FEBS Lett. 594, 2670–2694 (2020).
Packiam, M., Veit, S. J., Anderson, D. J., Ingalls, R. R. & Jerse, A. E. Mouse strain-dependent differences in susceptibility to Neisseria gonorrhoeae infection and induction of innate immune responses. Infect. Immun. 78, 433–440 (2010).
Currie, E. G. et al. Protection against Neisseria meningitidis nasopharyngeal colonization relies on antibody opsonization and phagocytosis by neutrophils. Preprint at bioRxiv https://doi.org/10.1101/2024.09.11.612551 (2024).
Perez, J. L. et al. From research to licensure and beyond: clinical development of MenB-FHbp, a broadly protective meningococcal B vaccine. Expert Rev. Vaccines 17, 461–477 (2018).
Billker, O., Popp, A., Gray-Owen, S. D. & Meyer, T. F. The structural basis of CEACAM-receptor targeting by neisserial Opa proteins. Trends Microbiol. 8, 258–260 (2000).
Shaughnessy, J. et al. Molecular characterization of the interaction between sialylated Neisseria gonorrhoeae and factor H. J. Biol. Chem. 286, 22235–22242 (2011).
Ram, S. et al. C4bp binding to porin mediates stable serum resistance of Neisseria gonorrhoeae. Int. Immunopharmacol. 1, 423–432 (2001).
Pohlner, J., Halter, R. & Meyer, T. F. IgA protease of Neisseria gonorrhoeae: isolation and characterization of the gene and its extracellular product. EMBO J. 3, 1595–1601 (1984).
DeRocco, A. J. & Cornelissen, C. N. Identification of transferrin-binding domains in TbpB expressed by Neisseria gonorrhoeae. Infect. Immun. 75, 3220–3232 (2007).
Noinaj, N., Cornelissen, C. N. & Buchanan, S. K. Structural insight into the lactoferrin receptors from pathogenic Neisseria. J. Struct. Biol. 184, 83–92 (2013).
Kammerman, M. T. et al. Molecular insight into TdfH-Mediated Zinc piracy from human calprotectin by Neisseria gonorrhoeae. mBio 11, e00949–00920 (2020).
Maurakis, S. et al. The novel interaction between Neisseria gonorrhoeae TdfJ and human S100A7 allows gonococci to subvert host zinc restriction. PLoS Pathog. 15, e1007937 (2019).
Beernink, P. T. et al. A meningococcal factor H binding protein mutant that eliminates factor H binding enhances protective antibody responses to vaccination. J. Immunol. 186, 3606–3614 (2011).
Liu, Y. & Russell, M. W. Diversion of the immune response to Neisseria gonorrhoeae from Th17 to Th1/Th2 by treatment with anti-transforming growth factor β antibody generates immunological memory and protective immunity. mBio 2, e00095–00011 (2011).
Islam E. A. et al. Adjuvant-dependent impacts on vaccine-induced humoral correlates of protection in preclinical models of nasal and genital colonization by the pathogenic Neisseria. Preprint at bioRxiv https://doi.org/10.1101/2024.09.07.611809 (2024).
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).
Stamatakis, A. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313 (2014).
Bryksin, A. V. & Matsumura, I. Overlap extension PCR cloning: a simple and reliable way to create recombinant plasmids. Biotechniques 48, 463–465 (2010).
Studier, F. W. Protein production by auto-induction in high density shaking cultures. Protein Expr. Purif. 41, 207–234 (2005).
Sintsova, A. et al. Selection for a CEACAM receptor-specific binding phenotype during Neisseria gonorrhoeae infection of the human genital tract. Infect. Immun. 83, 1372–1383 (2015).
Jerse, A. E. Experimental gonococcal genital tract infection and opacity protein expression in estradiol-treated mice. Infect. Immun. 67, 5699–5708 (1999).
Acknowledgements
The authors would like to acknowledge Dr. Vanessa Allen (Public Health Ontario) for generously sharing recent gonococcal isolates collected in Ontario, Canada, from between 2015 and 2018, Dr. Cynthia Nau Cornelissen for providing the FA19deltaTbpB and FA19deltaTbpAB strains used in this study, Dr. Gursonika Binepal for helping prioritize which Public Health Ontario strains to include in the present study, Ethan Gray-Owen for technical support during cloning and ELISA optimization, and Dr. Nelly Leung for general laboratory support. The authors would also like to thank the animal support staff at the Division of Comparative Medicine at the University of Toronto for technical and welfare support for both the rabbit and mouse studies presented here, including performing rabbit immunizations and blood sampling. The authors appreciate the financial support received for the research, authorship, and/or publication of this article. This research was supported by National Institute of Health funding #R01-AI125421-01A1 and #1RO1AI141229. S.D.G. and T.F.M. are supported by the Canada Research Chair Program.
Author information
Authors and Affiliations
Contributions
J.E.F. conceptualized the rabbit immunization study; conceptualized, performed, analyzed, and interpreted the murine infection study and related serological analysis; and wrote the original manuscript. E.A.I. conceptualized, performed, analyzed, and interpreted rabbit and mouse cross-reactivity ELISAs and serum bactericidal assays; performed and interpreted the murine infection study and related serology; contributed to writing and editing the manuscript. D.M.C. performed the bioinformatics analysis and contributed to writing and editing the manuscript. D.N. designed constructs and performed the protein purification for the rabbit immunizations; contributed to writing and editing the manuscript. N.A. performed the protein purification for the murine immunizations, provided technical support for the murine challenge study, and edited the manuscript. E.G.C. performed the murine infection study and related serology, and edited the manuscript; J.Z. and J.L. performed the murine infection study, and edited the manuscript. A.B.S., T.F.M., and S.D.G. provided conceptualization support, supervision, and edited the manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare the following competing interests: J.E.F., E.A.I., D.M.C., D.N., N.A., A.B.S., T.F.M., and S.D.G. are named inventors on the below patent and patent application which include intellectual property regarding the use of transferrin binding protein B as a vaccine antigen for protection against N. gonorrhea: 1. Schryvers, A.B., Moraes, T., and Gray-Owen, S.D. Immunogenic compositions and vaccines derived from bacterial surface receptor proteins. International Patent Application PCT/CA2014/051146. Filed December 1, 2014. PCT National Phase Filings in: US, Canada, Europe, China, Australia, Japan, India, Brazil, New Zealand, Eurasia, South Korea. US Patent #10149900B2 issued December 11, 2018. Australian Patent 2014360590 was granted on August 28, 2019. Eurasian Patent 201691140 was granted on April 4, 2021. New Zealand Patent 721121 was granted on December 1, 2021. European Patent #3077512 was granted on February 9, 2022. South Korea patent #10-2507993 was granted on March 6, 2023. India Patent #424510 was granted on March 9, 2023. 2. Fegan, J., Islam, E., Curran, D., Ng, D., Au, N., Schryvers, A.B., Moraes, T.F. and Gray-Owen, S.D. Vaccines and methods for the treatment of Neisseria gonorrhoeae infections. US Patent and Trademark Office #63/691,519, filed September 6, 2024. All other authors (E.G.C., J.J.Z., and J.L.) declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Fegan, J.E., Islam, E.A., Curran, D.M. et al. Rational selection of TbpB variants yields a bivalent vaccine with broad coverage against Neisseria gonorrhoeae. npj Vaccines 10, 10 (2025). https://doi.org/10.1038/s41541-024-01054-0
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41541-024-01054-0
