N. gonorrhoeae strains F62 and FA1090 A25 were used in mouse intravaginal challenge, assays of serum bactericidal and opsonophagocytic activity assays, and to generate OMV as indicated. Gonococcal strains were cultured on GC base medium (Difco) supplemented with Kellogg’s supplement I under 5% CO2 at 37°C. GC-VNCTS agar [GC agar with vancomycin, colistin, nystatin, trimethoprim (VCNTS supplement; Difco) and 100 μg/ml streptomycin (Sm)] and heart infusion agar (HIA) were used to isolate N. gonorrhoeae and facultatively anaerobic commensal flora, respectively, from murine vaginal swabs.

All animal experiments were conducted at the Uniformed Services University according to the guidelines of the Association for the Assessment and Accreditation of Laboratory Animal Care using a protocol approved by University’s Institutional Animal Care and Use Committee. Female BALB/c mice [4-5 weeks, Charles River Laboratories (Wilmington, MA)] were housed with autoclaved food, water and bedding and allowed to acclimate to the animal facility for 10 days. Mice were immunized three times at 3-week intervals with 250 µL of 4CMenB (Bexsero™, Glaxo Smith Kline) administered subcutaneously, or with a 1:1 mixture of 12.5 µg MC58 ΔABR (24) and Alhydrogel (InvivoGen) adjuvant administered intraperitoneally [prepared as previously described (23)]; Alhydrogel (0.75 mg Aluminum) diluted to 250 µL in Phosphate Buffered Saline administered subcutaneously was administered as a control. Three weeks after the last immunizations, mice in the diestrus stage of the estrous cycle were identified by vaginal smear and intravaginally challenged with 1x10⁶ CFU of N. gonorrhoeae F62 as previously described (22). Vaginal swabs were quantitatively cultured for N. gonorrhoeae on days 1, 3, 5, and 7 post-inoculation and culture results were expressed as CFU/ml of vaginal swab suspension (limit of detection: 20 CFU). The immunization/challenge was performed twice. In each experiment, fifteen mice were immunized in each vaccine group and between 12 and 15 mice in each group were in the correct estrus stage to undergo N. gonorrhoeae challenge. The percentage of mice with positive cultures at each time point was plotted using a Kaplan Meier curve and analyzed by the Log Rank test. N. gonorrhoeae bacterial recovery and vaginal PMN counting area under the curve (AUC) differences between multiple groups were performed by ordinary one-way ANOVA with Tukey’s multiple comparisons. Correlation analyses between bacterial burden and PMN levels used Pearson Correlation Coefficient. Analyses were performed by GraphPad Prism software (GraphPad Software, Version 9, La Jolla, Calif)

Microtiter plates (384-well) were coated with 80 ng/well of N. gonorrhoeae strain F62 OMV, and total IgG, IgG1, IgG2a, IgG3 and IgA antibodies from serum and vaginal lavage were measured by standard ELISA (25) using a BioMek i7 liquid handler and isotype-specific detection antibodies from Southern Biotech. Results were reported as endpoint titer or EC50 of a 4PL curve fit using GraphPad Prism software (GraphPad Software, v9 or higher).

Crude outer membranes from N. gonorrhoeae strain F62 were prepared with deoxycholate and separated by electrophoresis using single well gels (Invitrogen Cat# EC6026BOX). The separated proteins were transferred to nitrocellulose membranes and immunoblotting performed using a slot blot apparatus to allow the single lane of proteins to be probed with each individual mouse serum specimen. Immunoblots were developed by incubation with horseradish peroxidase–conjugated anti-mouse IgG secondary antibodies (Bio-Rad), followed by Pierce ECL Western Blotting Substrate (ThermoFisher). The blots were imaged and band intensities determined using the FlourChem E instrument (Protein Simple). Antibody titers or levels, and immunoblot intensity level differences among multiple groups were compared by ordinary one-way ANOVA with Tukey’s multiple comparisons using GraphPad Prism software (GraphPad Software, Version 9, La Jolla, Calif). To assess the change in antibody levels post-gonococcal challenge relative to pre-challenge levels, fold changes were calculated using the formula post/pre and expressed on the log₂ scale. The mean change in bacterial burden (AUC of log₁₀ CFU) with respect to antibody levels and immunoblot intensity was assessed with simple linear regression and 95% confidence intervals using STATA18.

SBA was performed as described (26). In brief, twenty microliters of 5 x 10⁴ CFU/mL bacteria were suspended in 2% bovine serum albumin in Hank’s Balanced Salt Solution containing calcium and magnesium (HBSS; Gibco, Catalog # 14025-092), and incubated with 20µL of heat-inactivated sera pooled from 4–6 immunized mice or control antibodies for 15 minutes at 37°C with 5% CO₂. IgG/IgM-depleted pooled normal human serum (Igd-NHS, PelFreez Catalog #34010, Lot #28341) used as active complement source was added at 4% final concentration. Samples were mixed and incubated for another 30 minutes. Aliquots were plated on GCB agar and incubated overnight. CFUs of test sera were enumerated after overnight culture at 37°C with 5% CO_2. Differences in bactericidal titers were assessed by unpaired t and Mann Whitney tests. Analyses were performed by GraphPad Prism software (GraphPad Software, Version 9, La Jolla, Calif)

OPK assays were performed using fresh human neutrophils as effector phagocytes and C6-depleted pooled normal human serum as a source of complement as previously described (26). Neutrophils were purified from the venous blood of healthy human subjects who provided written consent, following a protocol approved by the University of Virginia Institutional Review Board for Health Sciences Research (#13909). In brief, strain FA1090 bacteria (1000 CFU/well) in HBSS were incubated with 20µL of heat-inactivated sera pooled from 4–6 immunized mice for 30 minutes at 37°C with 5% CO_2. Sixty µL of C6-depleted pooled NHS (5% final concentration, Complement Technology, Catalog # A323) and 100 µL of 2 x 10⁶/mL neutrophils were added to each well, mixed well and incubated for 2 hours at 37°C with 5% CO₂. Aliquots of 20 µL of each suspension were plated onto GCB agar. CFU were enumerated and presented as the percentage of bacteria killed. Differences in opsonophagocytic killing were assessed by unpaired t and Mann Whitney tests using GraphPad Prism software (GraphPad Software, Version 9, La Jolla, Calif)

Mice vaccinated and inoculated with N. gonorrhoeae F62 (26-29 mice/group), were euthanized at the end of the experiment using CO₂ overdose. Spleens were harvested, and a single cell suspension of splenocytes was prepared By crushing and passing the suspension through a 70µm filter in T cell Media (RPMI 1640, 10% FBS, 1% HEPES, 1% Pen/Strep, 0.1% 2-ME, 1% 1N NaOH, 1% sodium pyruvate, 1% MEM non-essential amino acids, 2% MEM amino acids). Cells were centrifuged at 400 x g for 10 minutes and resuspended in ACK lysing buffer (Thermo Fisher Scientific) then washed twice and resuspended in T cells media at the concentration of 1x10⁷ cells/mL. Splenocytes were re-stimulated ex vivo by mixing 100µl of cells with 100 µl of [1 µg/ml] Ng-OMV or media as control in 96-well plate. The plates were incubated at 37°C 5% CO₂ for 48 hours and supernatants were harvested and stored at -80°C. Thawed samples were analyzed using the Invitrogen Th1/Th2/Th9/Th17/Th22/regulatory T cell (Treg) cytokine 17-plex mouse ProcartaPlex panel (Thermo Fisher Scientific) to determine levels of cytokines/chemokines. Assays were read on a Luminex 200 or FlexMAP 3D reader, and the background-corrected mean fluorescence intensities were compared to standard curves to calculate concentrations of each protein using BioPlex Manager v6.2 (Bio-Rad). Results below the lower limit of quantitation (LLOQ) were imputed as half the LLOQ. Differences between groups were assessed with one-way ANOVA followed by no paring Šidák’s multiple comparison test, with a single polled variance. using GraphPad Prism software (GraphPad Software, Version 9, La Jolla, Calif). The effect of each measured cytokine on the mean change in bacterial burden (AUC of log₁₀ CFU) was assessed with simple linear regression with 95% confidence intervals using STAT18.

Principal Coordinates Analysis (PCoA) was conducted in R to assess sample similarity patterns with respect to antibody and antigen-specific cellular responses. Data were preprocessed using the dplyr package for filtering and transforming, and standardized with clusterSim. The pcoa function in the ape package function computed principal coordinates based on a distance matrix, providing a visual representation of sample similarities in reduced dimensional space. The package ggfortify was used for visualization and graphing purposes. The following variables were considered for the generation of the distance matrix: the pre-gonococcal and post-gonococcal serum immunoglobulin, post-gonococcal challenge and vaginal wash immunoglobulin, and the magnitude of the antigen-specific cellular immune responses (as defined by the log₂ fold change in secreted cytokines measured by multiplex Luminex when the splenocytes of immunized and challenged mice were co-cultured with gonococcal OMV).

Prior studies have demonstrated that immunization with 4CMenB, administered subcutaneously, or MC58 ΔABR OMV, administered intraperitoneally, leads to a shortened duration of infection and reduced recovery of viable N. gonorrhoeae bacteria in a murine genital tract infection model (22, 23). In order to determine whether there were shared mechanisms of enhanced clearance between these two OMV-containing vaccines administered by different routes, BALB/C mice were vaccinated with 3 doses of 4CMenB (administered subcutaneously), MC58 ΔABR OMV (administered intraperitoneally), or Alum (administered subcutaneously) and subsequently intravaginally challenged with gonococcal strain F62 ( Figure 1 ). 4CMenB- and MC58 ΔABR OMV- vaccinated mice both demonstrated a significant increase in bacterial clearance compared to Alum controls ( Figure 2A ; Supplementary Figures 1A , 2A ). At day 7, 34.8% of 4CMenB immunized mice and 64.2% of MC58 ΔABR OMV-immunized mice remained infected with N. gonorrhoeae while 88.5% of the Alum-injected control animals continued to have recoverable N. gonorrhoeae. The total bacterial burden, measured as the area under the curve for recovered N. gonorrhoeae CFU plotted against time for each individual mouse, was lower in animals with clearance of N. gonorrhoeae ( Figure 2B ; Supplementary Figures 1B , 2B ). ΔABR-immunized mice had a slight delay in neutrophil influx with significantly lower vaginal neutrophils on day 5, when compared to Alum and 4CMenB ( Figure 2C ; Supplementary Figures 1C , 2C ). However, there was not a statistically significant difference in the total neutrophil influx during the infection measured as the area under the curve of vaginal neutrophil count plotted against time ( Supplementary Figure 3). For mice across all groups, the total neutrophil influx was inversely correlated to the total recovered N. gonorrhoeae CFU at the individual mouse level ( Supplementary Figure 4 ). However, when looking groups of animals based on vaccination status only, 4CMenB-immunized animals showed significant inverse correlation between neutrophil influx and total recovered bacterial burden ( Figure 2D ; Supplementary Figures 1D , 2D ). Because 4CMenB immunization demonstrated a similar inflammatory response but was associated with a significant inverse correlation between bacterial burden and neutrophil response, these finding suggest the possibility that 4CMenB immunization leads to an immune response that enhances neutrophil based clearance of N. gonorrhoeae in these mice.

Antibodies can play important roles in protection against bacterial infection through complement activation, neutralization, adherence blocking, and antibody-dependent cell-mediated cytotoxicity. To compare cross-species directed serological immunity elicited from meningococcal OMV containing vaccines, we collected serum from immunized mice both before and after the N. gonorrhoeae challenge and used an ELISA with OMV from N. gonorrhoeae strain F62 as a capture antigen to determine the differences in the anti-N. gonorrhoeae antibodies between immunized groups. Serum was collected 52 days after the first immunization (pre-challenge serum) and around 10 days after intravaginal challenge with N. gonorrhoeae (post-challenge serum). The titers of anti-N. gonorrhoeae IgG and IgG1 in sera of mice that received either vaccine were dramatically higher than those found in alum-injected controls, in which almost all animals exhibited titers below the level of detection for the assay ( Figures 3A, B ). 4CMenB immunized mice had significantly higher anti-N. gonorrhoeae IgG, IgG1 and IgG2a levels than MC58 ΔABR OMV immunized mice. In pre-challenge sera, only 4CMenB-immunized mice demonstrated significantly elevated titers of anti-N. gonorrhoeae IgG2a when compared to alum injected animals, though MC58 ΔABR OMV immunized mice did have a trend towards increased N. gonorrhoeae-directed IgG2a. ( Figure 3A ; Supplementary Figures 5A , 6A ). Post challenge, both 4CMenB and MC58 ΔABR OMV immunized mice demonstrated increased anti-N. gonorrhoeae IgG2a titers when compared to alum immunized animals ( Figure 3B ; Supplementary Figures 5B , 6B ). Interestingly, after vaginal N. gonorrhoeae infection, both groups of immunized mice demonstrated a significant rise of anti-N. gonorrhoeae specific IgG2A titer suggesting a recall response to N. gonorrhoeae exposure after immunization with N. meningitidis OMV. ( Figure 3C ; Supplementary Figures 5C , 6C ). Overall, as previously shown, immunization with these N. meningitidis OMV-containing vaccines elicit anti-N. gonorrhoeae serologic responses. However, we now show that the level and IgG subtype responses to immunization differ between vaccines as do the responses to antigen exposure during vaginal infection.

Because vaginal washing prior to N. gonorrhoeae inoculation might impact infection dynamics, we only measured anti-gonococcal vaginal antibody titers at the completion of the experiment, though prior studies have demonstrated anti- N. gonorrhoeae immunoglobulins are present in vaginal lavage from animals immunized with these vaccines. Both vaccine immunization groups demonstrated detectable anti-N. gonorrhoeae immunoglobulin levels in their vaginal wash fluid while anti-N. gonorrhoeae immunoglobulins were not detected in fluid from alum-immunized mice ( Figure 3D ; Supplementary Figures 5D , 6D ). Because all animals had received N. gonorrhoeae vaginal inoculations, the lack of anti-gonococcal immunoglobulin in the vaginal washes of alum control mice suggests that while immunization or the combination of immunization and inoculation lead to vaginal anti-N. gonorrhoeae antibodies, N. gonorrhoeae infection alone was insufficient to induced detectable vaginal anti-gonococcal immunoglobulins.

The relationship between anti-N. gonorrhoeae OMV immunoglobulin levels measured by ELISA and the level of N. gonorrhoeae recovered from vaginal swabs after intravaginal challenge was examined. In pre-challenge serum, anti-N. gonorrhoeae IgG, IgA or specific IgG subtype (IgG1 and IgG2a) titer was not significantly associated with total recovered N. gonorrhoeae after challenge in MC58 ΔABR-immunized animals ( Figure 4A ; Supplementary Figure 7 ). A significant correlation between anti-N. gonorrhoeae OMV IgG2a and lower recovered N. gonorrhoeae was observed in serum from these animals collected after challenge ( Figure 4B ; Supplementary Figure 7 ). Similarly, the magnitude of change in IgG2a levels after challenge was also correlated with reduced bacterial recovery after MC58 ΔABR immunization ( Figure 4C ; Supplementary Figure 7 ). Despite having higher titers of anti-N. gonorrhoeae immunoglobulins than those observed after MC58 ΔABR immunization ( Figure 3 ), 4CMenB-immunized animals demonstrated no association between these titers and the level of recovered bacteria after challenge ( Figures 4A–C ). Unexpectedly, anti-N. gonorrhoeae immunoglobulins present in vaginal wash fluid after challenge were correlated with increased N. gonorrhoeae burden after challenge in MC58 ΔABR immunized animals ( Figure 4D ).

We also sought to assess whether vaccination with 4CMenB or MC58 ΔABR elicited antibodies that induced bactericidal antibody activity or promoted human neutrophil opsonophagocytic killing activity against N. gonorrhoeae, and whether these functional responses correlated with protection. Sera from mice were pooled to provide adequate volume of specimen to conduct the killing assays. Mice were grouped into pools by the N. gonorrhoeae bacterial burden to allow comparisons in these assays between serum from mice that rapidly cleared infection and those in which N. gonorrhoeae infection persisted ( Supplementary Figure 8A ). While both immunizations were associated with increased serum bactericidal activity and opsonophagocytic killing activity when compared to serum from alum immunized animals, the point estimate for SBA titer was higher in MC58 ΔABR immune sera ( Supplementary Figures 8B, C ) and the point estimate for OPK activity was higher in 4CMenB immune sera ( Supplementary Figures 8E, F ). Though the number of pools was insufficient for rigorous statistical testing, there was no association between SBA titer ( Supplementary Figure 8D ) or OPK activity ( Supplementary Figure 8G ) and the recovered bacterial burdens during infection.

SDS-PAGE and immunoblot analysis were used to qualitatively and semi-quantitatively assess the identify of N. gonorrhoeae proteins that cross-react with immune sera collected from immunized mice after intravaginal challenge with N. gonorrhoeae. Sera from 4CMenB and MC58 ΔABR immunized mice were used individually to probe a membrane carrying proteins from OMV derived from N. gonorrhoeae strain F62. Serum from alum immunized animals did not recognize N. gonorrhoeae proteins at the dilution used for these experiments ( Figure 5A ). Serum from each immunized mouse recognized a similar but not completely overlapping set of N. gonorrhoeae OMV proteins within each immunization group though the intensity of immunoreactivity against each band had some variation between individual mice ( Figure 5A ). Antibodies in the sera of MC58 ΔABR immunized mice sera consistently bound to antigens with apparent molecular weights of approximately 85, 66, 55, and 42 kDa, while antibodies in the sera of 4CMenB immunized mice consistently recognized antigens with a molecular weight of approximately 66, 55, 37, and 15 kDa ( Figure 5A ). The signal intensity for the entire lane and for each of the consistently identified bands were determined using densitometry for each serum specimen used in these immunoblots. The relationship between immunoblot reactivity of each serum specimen and the quantity of N. gonorrhoeae recovered during vaginal challenge was explored using linear regression. For ΔABR OMV-immunized mice, the total level of immunoreactivity as well as intensity of the 85 and 66 kDa bands were associated with increased recovered N. gonorrhoeae ( Figure 5B ; Supplementary Figure 9 ). For 4CMenB-immunized animals, the immunoreactivity against the 37 kDa band was correlated with lower N. gonorrhoeae recovery ( Figure 5B ; Supplementary Figure 9 ) suggesting an association of immunoglobulins directed against this 37 kDa protein with protection from infection. Overall, these serologic studies demonstrate that immunoreactivity against N. gonorrhoeae antigens measured by ELISA and immunoblot likely detect antibodies with different properties and ability to provide protection from gonococcal infection.

To determine whether 4CMenB and MC58 ΔABR vaccinations activated anti-Ng-OMV specific cellular immune responses, spleens of immunized and N. gonorrhoeae challenged mice were harvested, and single-cell suspensions were prepared and cultured in the presence or absence of Ng- OMV as a source of gonococcal antigens. After two days of restimulation, cell supernatants were collected and used to measure production of 12 different cytokines using bead based multianalyte arrays. Baseline splenocyte secretion of all measured cytokines from MC58 ΔABR-immunized and N. gonorrhoeae challenged mice was significantly lower than that of Alum control and 4CMen immunized and N. gonorrhoeae challenged mice ( Supplementary Figures 10 - 12 ). To assess the adaptive immune responses specific to N. gonorrhoeae antigens that were specific to immunization, we compared the fold-increase between baseline and stimulation with Ng-OMV ( Figure 6A ; Supplementary Figures 13A , 14A ) or 4CMenB ( Figure 6B , Supplementary Figures 13B , 14B ) for each secreted cytokine in splenocytes harvested from alum, MC58 ΔABR, and 4CMenB immunized mice after N. gonorrhoeae challenge was completed. After restimulation with Ng-OMV antigens, splenocytes from 4CMenB and MC58 ΔABR immunized mice exhibited increased IL-2, IL-4 and IL-5 secretion compared to the controls ( Figure 6A ; Supplementary Figures 13A , 14A ). Splenocytes from MC58 ΔABR immunized animals also exhibited enhanced N. gonorrhoeae OMV-stimulated IFN-γ, TNF-α, IL-10, and IL-17 when compared to splenocytes from control mice. While splenocytes from MC58 ΔABR immunized animals exhibited stimulation of a broader variety of cytokines than 4CMenB immunized animals in response to N. gonorrhoeae OMV, splenocytes from 4CMenB-immunized mice demonstrated robust stimulation of all cytokines measured when stimulated ex vivo with 4CMenB. Linear regression analysis examining relationships between the burden of N. gonorrhoeae during infection and splenocyte cytokine production at baseline ( Figure 7A ), with N. gonorrhoeae OMV stimulation ( Figure 7B ), and with 4CMenB stimulation ( Figure 7C ) was conducted. Interestingly, the level of 4CMenB stimulated secretion of the immunosuppressive cytokine IL-10 was correlated with increased N. gonorrhoeae burden in 4CMenB immunized mice. 4CMenB stimulated IL-2 secretion was also associated with increased N. gonorrhoeae burden ( Figure 7 ; Supplementary Figure 15 ).

Our data illustrate that vaccination with 4CMenB and MC58 ΔABR is associated with enhanced gonococcal clearance in a mouse infection model, but immune components (immunological factors) common to both immunizations that were associated with this protection were not identified using simple linear regression analysis to determine the relationship of each parameter with infection intensity. To test whether multiple parameters in combination with one another might yield a common immunologic mechanism for N. gonorrhoeae clearance, we conducted unsupervised principal coordinate analysis ( Figure 8A ). The result showed that mice immunized with Alum, 4CMenB or MC58 ΔABR formed clusters that were distinct from each other, suggesting that different vaccinations exhibited difference in polarization of immune responses, either due to the antigenic make up of each vaccine or due to the route of vaccine administration. When Alum immunized animals were taken out of the PCoA, 4CMenB or MC58 ΔABR continued to cluster based on vaccination group ( Figure 8B ). However, in both these analyses, there was not clustering of mice that cleared infection separate from mice that did not clear infection. PCoA was also conducted on each group of immunized mice separately to test whether there was within group parameters that were associated with N. gonorrhoeae clearance ( Figures 8C, D ). Neither vaccination group was found to have clustering of mice that cleared or failed to clear N. gonorrhoeae infection. Taken together, the PCoA data showed that vaccination yielded distinct immunological profile and that there was not a distinct profile that separated mice based on their clearance of N. gonorrhoeae infection.