Combination of live-attenuated and adenovirus-based vaccines provides immediate and durable protection against Yersinia pestis in mice

Abstract

Yersinia pestis is the etiological agent of pandemic-causing plague with no Food and Drug Administration (FDA)-approved vaccines. The natural emergence or intentional release of antimicrobial-resistant strains are alarming, necessitating development of new-generation vaccines and immunization strategies that produce rapid and durable immunity. Herein, we evaluated immediate (within 7 days) or long-term (6 months) protection and immune responses by combining patented live-attenuated vaccines (LAVs) with a trivalent adenovirus-type 5-based vaccine (Ad5-YFV) in C57BL/6 mice. Animals were immunized with either a two-dose prime-pull strategy (intramuscular [i.m.] injection of LAVs followed by intranasal [i.n.] instillation of Ad5-YFV, 21 days apart) or a 1-dose strategy by simultaneous administration of LAVs (i.m.) and Ad5-YFV (i.n.). Animals immunized with both strategies were either challenged after 6 months or, with the simultaneous strategy, 7 days post-vaccination in lethal pneumonic or bubonic Y. pestis CO92 infection models. Mice immunized with both strategies were completely protected and elicited robust and durable humoral and cellular immune responses through 6 months. Simultaneous vaccination also protected animals against bubonic (100%) and pneumonic (80%) challenges 7 days post-vaccination with rapid induction of immunoglobulin M (IgM), IgG3, and total IgG humoral responses. Thus, these vaccines and strategies are suitable for reactive (emergency) and preventive (in endemic regions) use.

Graphical abstract

Immediate/long-term protective immunity in mice was evaluated by combining live-attenuated and trivalent-vector-based plague vaccines. Immunization with either two-dose prime-pull or one-dose simultaneous administration elicited long-term immunity and complete protection against pneumonic plague. Simultaneous vaccination also provided immediate protection against bubonic/pneumonic challenges, thus suitable for reactive and preventive use.

Introduction

Yersinia pestis infections have resulted in three major plague pandemics,¹ and the disease is re-emerging due to its recrudescence, especially in regions of endemicity. Though bubonic infections are most common in humans, recent outbreaks primarily of pneumonic plague with a case fatality rate (CFR) ranging from 8.6% to 36% in Madagascar were alarming.^2,3 Emergence of antimicrobial-resistant Y. pestis strains⁴ underscores the importance of developing and testing efficacious plague vaccines for reactive and preventative use. This was further enforced in 2024, as the World Health Organization (WHO) designated Y. pestis as a priority pathogen for epidemic and pandemic preparedness.⁵ However, no Food and Drug Administration (FDA)-approved plague vaccines exist.

Many vaccine platforms have been developed and all except for live-attenuated plague vaccines (LAVs) focus solely on two antigens: the fraction 1 capsule-like antigen (F1) and the low calcium response V antigen (LcrV or V) of the type 3 secretion system (T3SS).⁵ The F1 antigen stimulates rapid and robust adaptive humoral immune responses,^5,6 although not required for Y. pestis virulence in pneumonic infection. Thus, anti-F1 antibodies are inconsequential when infections are inflicted by F1-negative Y. pestis strains.^7,8 LcrV is an essential virulence factor and generates protective humoral and cellular immune responses in animal models.^9,10 However, LcrV alone does not confer complete protection against pneumonic infection,¹¹ and polymorphic clades of LcrV exist, which could compromise cross-protection of F1V-based vaccines against all circulating Y. pestis strains.¹²

LAVs for plague, e.g., EV76 and its derivatives that lack the pigmentation locus (pgm) for iron acquisition, have been successfully used in China, Mongolia, and the Former States of the Soviet Union¹³; however, the vaccine(s) can revert to the wild-type (WT) phenotype in the presence of excess iron, e.g., hemochromatosis,^14,15,16 and are associated with other adverse reactions and, hence, are not FDA approved.¹⁷

We developed three patented vaccine candidates, including two LAVs (LMA and LMP), and a human adenovirus type 5 vector (Ad5)-based vaccine incorporating three plague antigens (Ad5-YFV).^18,19,20 While LMA is deleted for genes encoding Braun lipoprotein (lpp), acetyltransferase B (msbB), and attachment-invasion locus (ail), LMP has deletions of lpp, msbB, and plasminogen-activating protease (pla). These targeted genes encoded either the bacterial outer membrane components (Lpp, Ail, and Pla) or an enzyme (MsbB) with the ability to modify bacterial membrane structures. Importantly, these all contribute to pathogenesis of Y. pestis. Specifically, deletion of lpp and msbB synergistically dampens detrimental inflammatory responses as Lpp binds to Toll-like receptor 2 (TLR-2), while deletion of msbB alters the lipid A structure of lipopolysaccharide (LPS) and reduces its endotoxicity by modifying its binding to TLR-4.^20,21,22 Deletion of ail reduces bacterial invasion by disrupting the ability to bind fibronectin, thereby affecting bacterial adhesion and Yersinia outer protein (Yop) delivery to host cells.²³ Deletion of pla disrupts degradation of extracellular matrices and fibrin clots, preventing bacterial dissemination and colonization of the lungs in high numbers.^24,25,26 Thus, these two LAV candidates not only systematically eliminate key Y. pestis virulence factors but also possess a highly immunogenic surface antigen profile.^20,27 These rationally designed avirulent mutants are approved to be used at biosafety level 2 (BSL-2) by the Centers for Disease Control and Prevention (CDC) and the National Institutes of Health (NIH).²⁸

Ad5-YFV is a replication-deficient Ad5-based vaccine incorporating genes for three plague antigens: Yersinia secretion protein F (YscF) of the T3SS, F1, and LcrV.^18,29,30 While F1 and LcrV are primary targets for neutralizing antibodies, inclusion of YscF, a highly conserved component of the T3SS injectosome needle, provides an essential element that disrupts injection of cytotoxic effector proteins (e.g., YopE) into host cells.^18,30 We have shown the trivalent antigen formulation overcomes the vulnerability of F1-negative Y. pestis strains that escape from a solely V-antigen-dependent immune response.¹¹ Both of the LAVs and the Ad5-YFV vaccine are fully characterized for safety and efficacy in various animal models, with each displaying their own unique immunological characteristics.^18,19,20 We have shown that mice immunized with Ad5-YFV by the intranasal (i.n.) route and LMA by either the i.n. or intramuscular (i.m.) route in various orders using either a two-dose prime-boost or a one-dose simultaneous regimen (Ad5-YFV [i.n.] and LMA [i.m.] at the same time) elicited robust immune responses with complete protection against lethal pneumonic challenge with WT and F1-negative Y. pestis CO92 strains at both 21 days and/or 3 months post-vaccination (p.v.).³¹ Although the one-dose regimen (Ad5-YFV [i.n.] and LMA [i.m.]) was 100% efficacious at 21 days, the protection rate dropped to 50%–60% when Ad5-YFV and LMA were administered via the same i.n. route, and the animals were challenged with either WT or F1-negative Y. pestis CO92 strains.³¹

We further evaluated the efficacy of the two-dose regimen and assessed the roles of interleukin (IL)-17 and interferon gamma (IFN-γ) in protection during vaccination as well as during infection.^28,32 We showed that our vaccines were still 100% protective against developing pneumonic plague in monoclonal antibody-induced IL-17 knockdown²⁸ and IFN-γ knockout mice.³² Treatment of unvaccinated WT mice with IFN-γ prior to and during pneumonic challenge resulted in 60%–80% animal survival.³² Likewise, treatment of WT mice with three doses (1 μg/dose) of IL-17A or IL-17F at one day prior challenge, at the time of challenge, and 1 day after challenge provided 30%–40% protection against pneumonic plague infection. Thus, while IL-17 and IFN-γ have a protective role during infection, immunization with our vaccines circumvents their necessity.

Importantly, we also demonstrated significantly higher IgA in bronchoalveolar lavage fluid (BALF) of mice immunized with the LAVs (i.m.) first followed by Ad5-YFV (i.n.), irrespective of their IL-17 or IFN-γ status.^28,32 This prime-pull immunization strategy, in which a parenteral vaccination produces systemic immune responses (prime), followed by a mucosal boost (i.n.) to recruit primed antigen-specific immune cells to the mucosal site (pull),³³ was followed during these studies.

In 2018, the WHO released a target product profile (TPP) outlining the desired characteristics of a potential plague vaccine for both reactive and preventive use.³⁴ This included immunization with a single dose, preferentially, but limited to, a 2- to 3-dose schedule; ability to elicit both rapid and long-lasting protection; and a robust safety profile.³⁴ The current study is built upon our previous studies^19,28,32,35 and presented results demonstrating that a combination of LAVs and Ad5-YFV could provide both immediate (within 7 days) and long-lasting immunity (up to 6 months) and has the potential for both reactive and preventive use, adhering to the WHO TPP.

Results

Protection conferred to animals against pneumonic plague at 6 months post-vaccination

C57BL/6 mice (n = 15 per group) were vaccinated using either a prime-pull two-dose strategy, where an LAV (LMA or LMP; i.m.) was administered as the first dose on day 0 and Ad5-YFV (i.n.) as the second dose on day 21. In a simultaneous one-dose strategy, LAV (LMA or LMP; i.m.) and Ad5-YFV (i.n.) were administered at the same time on day 21 (Figure 1A). Mice (n = 15) receiving two doses of PBS were used as challenge controls. On day 200 (180 days p.v.), five mice from each group were allocated to BAL and flow cytometry assays, while the remaining 10 mice in each group were i.n. challenged with 100 LD₅₀ of Y. pestis CO92 on day 207 (187 days p.v.). The detailed vaccination regimens, timing, and allocation of mice were depicted in a timeline format (Figure 1A) and a table format (Figure 1B), respectively. Post-immunization, no local or systemic adverse effects, such as weight loss, ruffled fur, or injection site redness or tenderness in mice, were observed. All immunized mice, regardless of vaccination type or strategy, were completely protected against high-dose lethal pneumonic challenge, while all unimmunized control animals succumbed to infection by day 4 (Figure 1C).

Figure 1.

Combination of LAVs and Ad5-YFV provides long-lasting protection against lethal pulmonary Y. pestis CO92 infection

Mice (n = 15 per group) were immunized with LMA or LMP (i.m.) and Ad5-YFV (i.n.) either in a two-dose regimen 21 days apart (prime-pull) on days 0 and 21 or in a one-dose regimen where LMA or LMP (i.m.) and Ad5-YFV (i.n.) were delivered at the same time (Sim) on day 21. Mice receiving PBS were used as controls. The experimental time course and vaccination groups are illustrated in (A) and (B). Mice were separated into one of two cohorts: (1) bronchoalveolar lavage (BAL) and splenocyte collection for flow cytometry analysis on day 200 (n = 5 per group) and (2) 100 LD₅₀Y. pestis i.n. challenge on day 207 (n = 10 per group). The challenged animals were observed for morbidity and mortality for an additional 21 days. Survival percentages of mice are shown in (C). Kaplan-Meier analysis with log rank (Mantel-Cox) test was used to analyze survival. Asterisks represent statistical significance of vaccinated groups compared with PBS control mice. ∗∗∗∗p < 0.0001.

Long-lasting humoral immune responses were elicited in mice by vaccination

Serum samples were collected at various time points as depicted in Figure 1A. Dose 1 serum (prime bleed) of the two-dose regimen was collected on day 19. Dose 2 (two-dose regimen) and the simultaneous dose (Sim; one-dose regimen) sera were collected on day 39. Likewise, 2 mpv (months p.v.), 4 mpv, and 6 mpv sera were drawn on days 77, 142, and 196, respectively (Figure 1A). Immunoglobulin G (IgG) antibody titers were assessed in three biological replicates where each sample was composed of pooled sera from five randomly selected mice. Thus, three biological replicates were generated from 15 mice in each group at each time point. As shown in Figures 2A–2C, all vaccinated mice developed significantly higher IgG titers (p < 0.0001) against the three plague antigens F1, LcrV, and YscF, constituents of the Ad5-YFV vaccine, lasting up to 6 months p.v. compared with unimmunized control animals (receiving PBS). For the two-dose regimen (striped bars), prime dosing (dose 1) induced adequate IgG responses against F1 and YscF (Figures 2A and 2B), but not against LcrV (Figure 2C), corroborating previous reports that anti-LcrV antibodies required longer periods to develop compared with anti-F1 antibodies after immunization with recombinant LcrV (rLcrV) or rF1.^36,37 However, the boost vaccine dose significantly increased IgG titers against all three antigens (p < 0.0001) and reached peak levels either 18 days after dose 2 (day 39) for F1 or at day 77 (2 mpv) for YscF and LcrV (Figures 2A–2C; striped bars). The dominant nature of serum anti-F1 antibodies in animals immunized with either Ad5-YFV alone or in combination with LAVs was also reported after the boost dose in our previous studies.^11,31 Anti-LcrV IgG titers increased over time to levels comparable to anti-F1, while anti-YscF IgG titers were maintained at overall lower levels compared with both anti-F1 and anti-LcrV IgG titers (Figures 2A–2C). Overall, the greatest expansion of LcrV antibody titers occurred after the boost immunization (over 2 logs), followed by those for anti-F1 and anti-YscF antibodies (p < 0.0001), suggesting that the LcrV antigen generated a superior memory B cell response after the boost immunization.³⁸

Figure 2.

Antibody responses were sustained across 6 months in vaccinated mice

Serum (n = 3 biological replicates per group) or BALF (n = 5 individual mice per group) were collected at various time points as described in Figure 1A. Each biological replicate was pooled sera from five randomly selected mice and thus three biological replicates were generated from 15 mice in each group at each time point. For the two-dose regimen (striped bars), serum samples from immunized animals were designated as “dose 1” and ”dose 2,” which were collected on day 19 and day 39, respectively. For the one-dose regimen (solid bars), vaccination serum collected on day 39 was designated as “Sim. Dose.” Subsequently, serum was collected on days 77, 142, and 196 for both one- or two-dose vaccination regimens and designated as 2, 4, and 6 mpv, respectively. Samples collected from unimmunized controls were labeled as PBS. Endpoint ELISAs were performed to evaluate serum IgG levels against individual plague antigens F1 (A), YscF (B), and LcrV (C). Neutralizing antibodies were assessed from pooled sera (n = 15/group) by competitive ELISA against a biotinylated LcrV Mab7.3. The level of neutralizing antibodies was determined as the highest dilution factor of the examined serum that inhibited binding of the biotinylated Mab7.3 to LcrV by 50% (NT₅₀; D). Likewise, an endpoint ELISA was used to measure IgA titers against the F1V fusion protein in BALF collected from individual animals (n = 5) at day 200 (E). The turquoise dashed line represents the baseline values of unimmunized control samples at NT₅₀ or at the dilution of 1:50 (IgA), 1:100 (anti-YscF and -LcrV IgG), or 1:500 (anti-F1 IgG). Antibody titers were log-transformed and presented as geometric means with or without (NT₅₀) standard deviations (SD). Statistical significance was determined by one-way ANOVA with Tukey’s post hoc test. Asterisks with comparison bars represent statistical significance between respective vaccinated groups (A–E). Statistical significance of vaccinated groups compared with unimmunized (PBS) control mice in (A)–(C) was not shown, as all of them were significant (p < 0.0001) except those denoted as not significant (n.s.), but was denoted with asterisks above the bars in (E). ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

For the one-dose regimen (solid bars), sufficient antibody levels against all three antigens were produced, and the titers were comparable to the two-dose regimen over a 6-month period (Figures 2A–2C), demonstrating long-lasting, humoral immune responses in immunized animals (Figures 2A–2C). To evaluate neutralizing antibody titers in the serum of vaccinated mice, a competitive enzyme-linked immunosorbent assay (ELISA) was performed utilizing the biotinylated LcrV monoclonal antibody (mAb) Mab7.3b. Mab 7.3b has confirmed neutralizing activity and binds to the amino acid region 135–275 of the Y. pestis LcrV antigen.³⁹ Levels of neutralizing antibodies (NT₅₀) were up to two logs higher in vaccinated mice compared with unvaccinated controls and remained elevated up to day 196, regardless of the vaccination strategy (Figure 2D). Furthermore, overall greater IgA titers against the F1V fusion antigen in BALF collected on day 200 were observed in immunized mice compared with control animals (Figure 2E). However, mice receiving the one-dose regimen of LAV and Ad5-YFV elicited significantly lower levels of IgA against the F1V fusion protein in the lungs compared with their prime-pull counterparts (LMA + Ad5-YFV, p < 0.001; LMP + Ad5-YFV, p = 0.0139; Figure 2E), which further demonstrated the advantage of the prime-pull regimen to induce long-term mucosal immunity. Interestingly, mice immunized with LMA displayed greater anti-F1V IgA titers than mice immunized with LMP in both vaccination regimens, but the differences were not significant (Figure 2E).

To further assess the humoral immune response, splenocytes were isolated from immunized and unimmunized animals at day 200 and stained for germinal center activity (CD3⁻CD138⁻CD19⁺CD38⁻GL7⁺) and memory B cell production (CD3⁻CD138⁻CD19⁺CD38⁺GL7⁻IgD⁻; Figure 3A). Compared with unvaccinated control animals, all vaccinated mice generated significantly higher populations of germinal center B cells (p < 0.05–0.01; Figure 3B) and memory B cells (p < 0.05–0.01; Figure 3C). Although slightly lower memory B cells were observed in mice immunized with LMP and Ad5-YFV (turquoise and teal bars), significant levels were not reached compared with LMA- and Ad5-YFV-immunized animals (pink and burgundy bars; Figure 3C).

Figure 3.

Germinal center and memory B cells elicited in long-term vaccinated mice

Splenocytes were collected on day 200 (Figure 1A) from individual immunized and unimmunized control mice (n = 5 per group). Single-cell splenocyte suspensions were then surface stained with various B cell markers and subjected to flow cytometry. The gating strategy is shown in (A). The percentage of Germinal Center B cells (B) were gated as CD3⁻CD138⁻CD19⁺CD38⁻GL7⁺, while memory B cells (C) were identified as CD3⁻CD138⁻CD19⁺CD38⁺GL7⁻IgD⁻. Outliers were excluded using the ROUT method, and the data points shown range from n = 3–5. Data are presented as arithmetic means ± SD. Each dot represents samples from individual mice. Statistical significance was determined by one-way ANOVA with Tukey’s post hoc test. Asterisks directly above the bars represent statistical significance of vaccinated groups as compared with unimmunized (PBS) control mice. ∗p < 0.05; ∗∗p < 0.01.

Cellular immune responses were elicited at 6 months post-vaccination

To assess cell-mediated immune responses, splenocytes collected on day 200 were stimulated with rF1V and stained for various cytokines and T cell markers to evaluate cytokine-producing T cells, as well as central (T_CM: CD44⁺CD127⁺CD62L⁺) and effector (T_EM: CD44⁺CD127⁺CD62L⁻) memory T cells. In addition, an activation-induced marker (AIM) assay was employed to identify overall activated T cell populations (CD44⁺CD25⁺CD134⁺/CD69⁺) regardless of cytokine production.⁴⁰ As expected, cellular immune responses determined by levels of activated and memory T cells in all immunized mice were higher than their unimmunized counterparts (Figures 4A–4R). Furthermore, CD4 cells positive for Th1 cytokines (IFN-γ, IL-2, or tumor necrosis factor alpha [TNF-α]) were at comparable levels to Th2 cytokines (i.e., IL-4), indicating that the Th response transitioned from an initial Th2 bias at around 2 weeks post-immunization^28,32 toward a more balanced response at day 200 (Figures 4A, 4C, 4E, and 4I). Surprisingly, slightly better cellular immune responses were generally observed with the one-dose regimen than the corresponding two-dose regimen, particularly in T cells positive for IL-2, TNF-α, IL-4, IL-17A, and memory T cells (Figures 4C–4F, 4I–4L, and 4O–4R). These differences were statistically higher for IL-4⁺ T cells in mice immunized with LMP (Figures 4I and 4J; p = 0.0410 and p = 0.0265) and for effector memory CD8⁺ T cells in mice immunized with LMA (Figure 4R; p = 0.0426).

Figure 4.

Cellular immunity elicited in long-term vaccinated mice

Splenocytes were collected on day 200 from individual immunized and unimmunized control mice (n = 5 per group; Figure 1A). The isolated splenocytes were stimulated with rF1V (100μg/mL) for 48 h at 37°C and then treated with Brefeldin A for an additional 4 h. Splenocytes were surface and intracellularly stained with various cytokine and T cell markers for flow cytometry analysis. The percentage of cytokine-positive T cells (A–L), antigen-activated T cells measured by activation-induced marker assay (AIM; M and N), central memory (T_CM; O and P), and effector memory (T_EM; Q and R) T cells were acquired and plotted. Outliers were excluded using the ROUT method, and the data points shown range from n = 3–5. Data are presented as arithmetic means ± SD. Each dot represents samples from individual mice. Statistical significance was determined using one-way ANOVA with Tukey’s post hoc test. Asterisks with comparison bars represent statistical significance between respective vaccinated groups, while asterisks directly above the bars represent statistical significance of vaccinated groups as compared with unimmunized (PBS) control mice. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

On the other hand, when comparing LMA + Ad5-YFV- vs. LMP + Ad5-YFV-immunized mice (one-dose regimen) or LMA − Ad5-YFV vs. LMP − Ad5-YFV mice (two-dose regimen), those vaccinated with LMA developed greater IFN-γ⁺ T cell populations and polyfunctional T cells positive for all three Th1 cytokines (IFN-γ⁺IL-2⁺TNF-α⁺; Figures 4A, 4B, 4G, and 4H). The one-dose regimen of LMA + Ad5-YFV stimulated statistically higher CD8⁺ polyfunctional Th1 cytokines compared with LMP + Ad5-YFV (p = 0.0037; Figure 4H). Likewise, LMP-vaccinated animals displayed greater populations of T cells positive for IL-2, TNF-α, and IL-4 (Figures 4C–4F, 4I, and 4J). Based on these data, all evaluated regimens elicited robust and sustained cellular immune responses at 6 mpv. Simultaneous vaccination with LMP (i.m.) and Ad5-YFV (i.n.) elicited the best responses overall (Figures 4A–4R).

Simultaneous vaccination provided rapid protection against bubonic and pneumonic plague

As depicted above, simultaneous administration of our vaccines evoked superior long-term humoral and cell-mediated immune responses and provided complete protection to immunized animals against pneumonic plague (Figures 1, 2, 3, and 4). We then analyzed their ability to induce rapid and sufficient immunity in response to an emergent plague outbreak or bioterror scenario. Mice (n = 10) were vaccinated with a single dose of LAVs (i.m.) in combination with Ad5-YFV (i.n.), while animals that received PBS were used as controls. Seven days after vaccination, animals were split into two equal cohorts (n = 5) and challenged with either 15 LD₅₀ (4,200 CFU) by the i.n. route or 100 LD₅₀ (1,000 CFU) by the subcutaneous (s.c.) route of WT Y. pestis CO92. Blood was drawn at various time points to analyze levels of IgM, IgG3, and total IgG. The detailed vaccination regimens, timing, and allocation of mice were depicted in both a timeline and a table format (Figure 5A). As shown in Figure 5B, simultaneous vaccination conferred 40% (2 out of 5) and 80% (4 out of 5) protection against pneumonic plague to animals immunized with LMA + Ad5-YFV and LMP + Ad5-YFV, respectively, at 7 days post-immunization, while all control animals succumbed by day 4 post-infection. Impressively, all immunized animals survived bubonic infection 7 days p.v., while all control animals succumbed to infection by day 6 post-infection (Figure 5C).

Figure 5.

Simultaneous vaccination provided immediate and long-lasting protection to immunized animals against lethal pneumonic and/or bubonic challenge

Mice (n = 10 per group) were immunized with LMA or LMP (i.m.) and Ad5-YFV (i.n.) in a one-dose regimen delivered at the same time. Mice receiving PBS were used as controls. The experimental time course and vaccination groups are illustrated in (A). Seven days post-vaccination, mice were equally divided into two cohorts (n = 5 per group) and either intranasally (i.n.) or subcutaneously (s.c.) challenged with 15 LD₅₀ or 100 LD₅₀ of Y. pestis CO92, respectively. The morbidity and mortality of animals were observed for 21 days. Serum was collected from individual animals 3 days before immunization and on days 3, 7, 21, 28, 35, and 63. Survival percentages of animals from either the initial i.n. or s.c. challenge are shown in (B) and (C), respectively. On day 95, animals surviving from both the initial pneumonic and bubonic challenges (a total of seven and nine survivors from LMA or LMP involved vaccination, respectively), and a cohort of age-matched control animals (n = 5), were rechallenged i.n. with 100 LD₅₀ of Y. pestis CO92. Animals were monitored for additional 21 days, and animal survival percentages are displayed in (D). Kaplan-Meier analysis with log rank (Mantel-Cox) test was used to analyze survival. Asterisks represent statistical significance of vaccinated groups compared with PBS control mice. ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

To further evaluate the protection elicited by simultaneous administration of our vaccines during subsequent re-infection, the immunized animals that survived the initial pneumonic and bubonic infections (seven and nine total survivors from either LMA + Ad5-YFV- or LMP + Ad5-YFV-immunized mice, respectively), along with age-matched unvaccinated control mice (n = 5), were rechallenged i.n. with 100 LD₅₀ Y. pestis CO92 88 days after the initial infection (day 95; Figure 5A). One animal out of seven vaccinated with LMA + Ad5-YFV succumbed to infection (86% survival rate), while complete protection was achieved in the animals vaccinated with LMP + Ad5-YFV. All unvaccinated control animals succumbed to infection by day 4 post-rechallenge (Figure 5D). Animals that succumbed to infection had an average lung bacterial load of 3 × 10⁸ CFU and spleen bacterial load of 1.2 × 10⁹ CFU/per organ, while those surviving to the end of the study (21 days post-rechallenge) had fully cleared the plague bacilli.

Rapid antibody responses were stimulated by simultaneous vaccination

In addition to T cell-dependent innate immunity, T-independent (TI) innate-like immunity elicited by vaccination plays a significant role in early protection against bacterial pathogens.^41,42,43 TI responses are characterized by rapid extrafollicular plasma cell expansion that produces IgM, followed by isotype switching to IgG3 in mice.^41,42 For these reasons, we measured serum levels of IgM, IgG3, and total IgG against individual plague antigens F1 and LcrV in vaccinated mice. Serum was collected on days 3 and 7 p.v. and days 14, 21, 28, and 56 post-initial infection (Figure 5A). As shown in Figures 6A and 6C, on day 3 p.v., anti-F1 and -LcrV IgM antibody titers (checkered bars) had already reached over three logs higher in all vaccinated mice compared with unvaccinated controls receiving PBS. On day 7 p.v., anti-F1 IgM antibody titers in animals vaccinated with LMA + Ad5-YFV continued to rise (Figure 6A). While, the IgM titers slightly declined from days 3 to 7 in animals vaccinated with LMP + Ad5-YFV and reached significantly lower anti-LcrV IgM levels at day 7 as compared to levels on day 3 p.v. (p = 0.0073; Figure 6D). After the initial challenge, irrespective of bubonic or pneumonic infection, anti-F1 IgM antibody titers generally declined (Figures 6A and 6B), while anti-LcrV IgM antibody titers initially rose and either remained stable or declined thereafter (Figures 6C and 6D), suggesting Y. pestis infection did not significantly alter the IgM response. Nevertheless, all anti-F1 and -LcrV IgM antibody titers were maintained at approximately three logs above the unimmunized controls through day 56 post-initial infection (Figure 6).

Figure 6.

Early IgM antibody responses elicited in simultaneously vaccinated mice

Serum was collected from both individual immunized and unimmunized mice at various time points as described in Figure 5A, and endpoint ELISAs were performed to evaluate their anti-F1(A and B) and anti-LcrV (C and D) IgM levels. There were 10 animals in each immunized and unimmunized group at the start of the study. As the post-vaccination day 3 and day 7 bleeding time points were too close, five mice designated at each time point were bled to prevent animal over bleeding. On post-initial challenge days (21, 28, 35, and 56), each individual surviving mouse was bled. However, there were only two or five LMA + Ad5-YFV-immunized mice that survived from the pneumonic or bubonic challenge, respectively. Likewise, there were only four or five LMP + Ad5-YFV-immunized mice that survived from the pneumonic or bubonic study, respectively. Therefore, each data point represents individual animals ranging from 2 to 5, matching the number of live animals at each bleeding time point. Checkered bars represent IgM antibody titers at days 3 and 7 post-vaccination. Solid and striped bars represent IgM antibody titers in immunized animals either post-pneumonic or -bubonic infection. The turquoise dashed line represents the baseline values of unimmunized control samples at the dilution of 1:20. Antibody titers were log-transformed and presented as geometric means with standard deviations (SD). Statistical significance was determined by one-way ANOVA with Tukey’s post hoc test. Asterisks with comparison bars represent statistical significance between respective vaccinated groups, while asterisks directly above the bars represent statistical significance of vaccinated groups compared with unimmunized (PBS) control mice. ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

In contrast to IgM antibodies, anti-F1 and -LcrV IgG3 antibody titers in vaccinated mice increased slightly at day 3 p.v. to around 0.5 logs above unimmunized controls; however, titers quickly rose 2–3 logs higher in all vaccinated mice compared with unvaccinated animals at day 7 p.v. (Figure 7, checkered bars). After initial challenge, irrespective of bubonic or pneumonic infection route, IgG3 antibody titers against F1 and LcrV continued to rise and reached peak levels around days 14–28 post-initial infection. The IgG3 antibody titers against F1 and LcrV steadily declined but remained 3–4 logs higher than unimmunized controls through day 56 post-initial infection (Figure 7). Compared with IgM, peak IgG3 antibody titers against F1 and LcrV appeared later in the vaccination period (day 3 vs. day 7), indicting isotype switching from IgM to IgG3 (Figures 6 and 7).

Figure 7.

Early IgG3 antibody responses elicited in vaccinated mice

Serum was collected from individual immunized and unimmunized mice at various time-points as described in Figure 5A and detailed in the legend of Figure 6. Endpoint ELISAs were performed to evaluate the anti-F1(A and B) and anti-LcrV (C and D) IgG3 antibody levels for serum samples collected from immunized individual animals on days 3, 7, 21, 28, 35, and 56. Checkered bars represent titers at days 3 and 7 post-vaccination. Solid and striped bars represent IgM antibody titers in immunized animals either post-pneumonic or -bubonic infection. The turquoise dashed line represents the baseline values of unimmunized control samples at the dilution of 1:20. Antibody titers were log-transformed and presented as geometric means with standard deviations (SD). Statistical significance was determined by one-way ANOVA with Tukey’s post hoc test. Asterisks with comparison bars represent statistical significance between respective vaccinated groups, while asterisks directly above the bars represent statistical significance of vaccinated groups compared with unimmunized (PBS) control mice. ∗∗∗∗p < 0.0001; n.s., not significant.

A similar pattern was also observed for anti-F1 and -LcrV total IgG antibody titers compared with IgG3 during the vaccination period, indicating that IgG3 was the dominant isotype in the total IgG population (Figure 8). Both IgG3 and total IgG antibody titers in the serum of all immunized mice increased after the initial Y. pestis assault and reached peak levels at days 14–28 post-initial infection. However, total IgG antibody titers were maintained at 4–6 logs above unimmunized controls through day 56 post-initial infection, in contrast to the decline of IgG3 by that time point (Figures 7 and 8). Clearly, Y. pestis infection boosted the production of IgG3 and total IgG antibodies, indicating the involvement of memory responses that were not observed for IgM (Figures 6, 7, and 8). Generally, no differences regarding serum levels of IgM, IgG3, and total IgG were observed in immunized mice, irrespective of the LAVs received (LMA or LMP) during simultaneous vaccination (Figures 6, 7, and 8).

Figure 8.

Early IgG antibody responses elicited in vaccinated mice

Serum was collected from individual immunized and unimmunized mice at various time points as described in Figure 5A and detailed in the legend of Figure 6. Endpoint ELISAs were performed to evaluate the anti-F1 (A and B) and anti-LcrV (C and D) IgG antibody levels for serum samples collected from immunized individual animals on days 3, 7, 21, 28, 35, and 56. Checkered bars represent titers at days 3 and 7 post-vaccination. Solid and striped bars represent IgM antibody titers in immunized animals either post-pneumonic or -bubonic infection. The turquoise dashed line represents the baseline values of unimmunized control samples at the dilution of 1:20. Antibody titers were log-transformed and presented as geometric means with standard deviations (SD). Statistical significance was determined by one-way ANOVA with Tukey’s post hoc test. Asterisks with comparison bars represent statistical significance between respective vaccinated groups, while asterisks directly above the bars represent statistical significance of vaccinated groups compared with unimmunized (PBS) control mice. ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

Discussion

For pathogens associated with periodic outbreaks and severe disease sequalae, such as Y. pestis, development of pre- and post-exposure prophylaxis is urgently needed. Our previous studies combining LAVs and Ad5-YFV were mainly mechanistic in scope, demonstrating protective efficacy with a standard 3- to 4-week immunization schedule and revealing that IFN-γ and IL-17 signaling were largely dispensable for vaccine-mediated clearance of Y. pestis.^28,31,32 Earlier, a long-term study was performed only when LMA was combined with Ad5-YFV in a two-dose regimen for a period of 3 months.³¹ The current study expands upon those findings by examining the real-world deployment capabilities of our vaccines. Here, we further evaluated the durability of immune responses conferred by our LAVs in combination with the Ad5-YFV vaccine using either a two-dose prime-pull strategy or a one-dose simultaneous strategy for a 6-month period. Importantly, we have demonstrated that the heterologous one-dose vaccination strategy also provided rapid protection within 7 days after vaccination against both lethal pneumonic and bubonic Y. pestis challenge. Evaluating the immediate (reactive) and long-term (preventive) efficacy of our vaccination regimens links our pre-clinical proof of concept to viable multi-use countermeasures against current and future plague outbreaks.

During the long-term study, robust humoral and cellular immune responses were observed in all immunized animals through day 200 (Figures 2, 3, and 4), which were similar in magnitude and quality observed between 2 and 12 weeks post-vaccination in our previous studies,^28,31,32 indicating sustained immunity elicited by our vaccines. Slightly greater immune responses were observed in animals immunized with the one-dose regimen compared with the two-dose regimen, which may be attributed to several advantages of a simultaneous strategy. First, simultaneous vaccination with both an LAV and an Ad5-YFV delivered a greater quantity and a more diverse repertoire of antigens to mice at the initial immunization; second, this co-vaccination was simultaneously adjuvanted by both bacteria and adenovirus; and third, co-vaccination avoided issues commonly associated with prime-boost strategies, in which immune responses elicited after the boost dose do not reach desirable levels due to preexisting antibodies developed during priming. These preexisting antibodies could neutralize a live-attenuated vaccine(s) before successful replication in vivo or form antigen–antibody complexes against non-live vaccines, thereby reducing antigen availability.³⁸ Consequently, a one-dose regimen might confer a more robust initial immune response in animals and could potentially prolong the replication of LAVs before their clearance in vivo, leading to development of long-lasting immunity. Indeed, significantly increased IFN-γ⁺ and IL-17⁺ T cell populations at 3 weeks post-vaccination were also observed in mice immunized with LMA (i.m.) and Ad5-YFV (i.n.), with a one-dose compared with a two-dose regimen during our previous study.³¹ Despite the one-dose regimen optimizing the overall magnitude of the cellular immune response in immunized animals, the endpoint protection conferred by both vaccination strategies was equivalent in both our current and previous studies.³¹

In contrast, the simultaneous regimen elicited significantly lower secretory IgA (sIgA), which is dependent upon cognate help from T helper (Th) cells in mucosal-associated tissues⁴⁴ and is specifically associated with the production of IL-17, IL-4, and IFN-γ.^45,46,47 Interestingly, splenic IFN-γ⁺ (CD4), IL-4⁺, and IL-17A⁺ T cell populations were generally lower in mice immunized with the two-dose regimen, suggesting that the associated increase in sIgA was most likely due to the ability of prime-pull strategies to induce local mucosal immunity, rather than a contribution from cytokines. However, the possibility for greater induction of these cytokines within local respiratory mucosal surfaces by the two-dose regimen cannot be ruled out and warrants further investigation.

In addition, LMA and Ad5-YFV stimulated greater anti-F1V sIgA in BALF than LMP and Ad5-YFV in both the one- and two-dose vaccination regimens, which aligned with our previous studies utilizing C57BL/6 mice,³² but not Swiss Webster mice,²⁸ indicating that the strain of mouse also affects sIgA in response to vaccination. Indeed, Swiss Webster mice typically mount stronger sIgA responses at mucosal surfaces compared to C57BL/6, which is linked to their unique microbiota and enhanced capacity for antibody secretion.⁴⁸

During our previous immunization study utilizing two-dose homologous administration of LAVs alone, both LMA and LMP elicited robust immune responses in immunized mice.¹⁹ Compared with LMA, LMP induced a more substantial population of CD4⁺ IFN-γ⁺ or IL-17⁺ T cells in mice after immunization, as well as during the post-infection recall immune response.¹⁹ Slightly better protection was observed in LMP-immunized mice (100%) than after LMA immunization (80%) at 4 mpv. Furthermore, LMP induced superior antibody responses and greater CD4⁺ IFN-γ⁺ or IL-17⁺ T cells than LMA in immunized rats.¹⁹ However, these advantages of LMP vaccination were not apparent when the LAVs and Ad5-YFVvaccine were combined in a two-dose strategy.^28,31,32 Therefore, the addition of Ad5-YFV alleviated any differences in immunogenicity elicited by LMA and LMP and balanced the overall immune response in the two-dose regimen, indicating a dominant immunological role for Ad5-YFV in the two-dose regimen. However, simultaneous immunization with LMP and Ad5-YFV elicited the most optimal overall T cell responses, suggesting that the type of LAV plays an important role in stimulating the T cell response in the one-dose regimen.

The rapid induction of anti-F1 and -LcrV IgM and IgG3 antibodies and the immediate onset of protection against bubonic and pneumonic plague in mice vaccinated with the one-dose strategy of LAVs and Ad5-YFV (Figures 5, 6, and 7) suggest the involvement of innate-like B cells (e.g., B1 B cells) in a T cell-independent (TI) manner.⁴⁹ B1 B cells exhibit properties resembling the innate immune phenotype by mounting quick, specific responses to TI antigens and are characterized by the production of IgM followed by IgG3.⁴⁹ TI type I (TI-I) antigens, such as LPS, induce both specific and non-specific antibody responses through direct B cell stimulation, while TI-II antigens include haptens, such as the haptenated Ficoll experimental antigen, that engage B cell receptors in parallel.⁴² The native F1 antigen forms a capsule-like polymer on the surface of Y. pestis. Functional LcrV also exists as a pentameric structure on the bacterial surface, while the fusion YFV protein remains in the monomeric form in Ad5-YFV-infected cells.¹⁸ Therefore, the observed rapid induction of both anti-F1 and -LcrV IgM and IgG3 antibodies was potentially due to the multimeric nature of F1 and LcrV on the surface of the LAVs and the dispensable role of Ad5-YFV in eliciting immediate immune response for the 1-dose strategy. This assumption was further supported by the observation that one-dose LMP + Ad5-YFV vaccination provided more robust rapid-onset protection against pneumonic plague than LMA + Ad5-YFV (Figure 5B). This was not only in line with the greater observed immunogenicity of LMP over LMA during our previous homologous vaccination study¹⁹ but also reemphasized the dominant role of LAVs in tailoring the T cell response in the one-dose regimen.

Slightly greater antibody responses were observed after vaccination with LMA + Ad5-YFV than LMP + Ad5-YFV on the day of initial challenge (day 7; Figures 6, 7, and 8) contrasting the superior rapid-onset protection that was conferred by LMP + Ad5-YFV against pneumonic plague (Figure 5B). This discrepancy between antibody response and protection suggested the involvement of other protective innate mechanisms of neutrophils and macrophages, and antibody-independent functions of innate-like B cells, such as antigen presentation and co-stimulation of T cells, secretion of soluble mediators, and stimulation of lymphoid tissue development.⁴⁹ In addition, the live attenuated Y. pestis EV76 vaccine strain induced immediate protection to immunized animals by triggering iron nutritional defense mechanisms that abrogated siderophore-mediated virulence by competing for iron ions in the host.⁵⁰ While target-specific antibodies and adaptive T cell responses are vital for long-term protective efficacy, the immediacy of protection observed here suggests that a local innate or trained immunity may act as a critical first line of defense.⁵¹ Live attenuated and vector-based vaccines stimulate short-term inflammatory responses that drive metabolic and epigenetic shifts, thereby enhancing protection against subsequent infections.^52,53 In addition, mucosal delivery of replication-deficient adenoviral vectors engages pattern recognition receptors, thereby stimulating rapid pro-inflammatory cytokines and type I interferons.⁵⁴ Recent evidence suggests that localized viral vectors can induce trained immunity, resulting in heightened non-specific antimicrobial states.⁵⁵ Similar protective mechanisms may also be stimulated by our LAVs with Ad5-YFV, and further investigation is warranted.

The natural polymeric form of the capsular F1 antigen triggers rapid and robust B1-cell-dependent immune responses and elicits complete protection to immunized animals against lethal subcutaneous (bubonic) Y. pestis challenge, even several hours after vaccination.^37,43 This polymeric F1 also provided adequate protection to immunized mice against pulmonary Y. pestis infection 7 days post-vaccination.⁵⁶ However, the monomeric LcrV antigen was incapable of inducing rapid IgM and IgG3 antibody responses in immunized mice and did not confer protection to animals against bubonic plague.³⁷ Interestingly, immunization of mice with both polymeric F1 and monomeric LcrV antigens shortened the time required for achieving rapid protective immunity against pneumonic plague.⁵⁶ In this regard, the pentameric form of LcrV on the surface of our LAVs could induce rapid, protective immunity more efficiently than monomeric LcrV alone. However, it is unclear if the LAVs or Ad5-YFV alone would be sufficient to elicit rapid protective immunity and needs to be further examined.

Our vaccines provided more robust immediate protection against bubonic than pneumonic plague. Disregarding the more severe nature of pneumonic infection, the relatively higher anti-F1 antibody titers compared with anti-LcrV titers in immunized mice at the initial challenge day potentially contributed to this difference in protection. Previous reports indicated a dispensable role of LcrV in mediating rapid, protective immunity against bubonic plague; however, LcrV assisted the F1 antigen in mounting a quicker immune response against pneumonic plague.^37,56 In addition, the pivotal role of anti-LcrV antibodies against pneumonic plague infection was emphasized by the fact that a killed whole-cell (KWC) plague vaccine induced anti-F1, but not anti-LcrV, antibodies and therefore was ineffective in providing immunity against pneumonic plague.³⁶ Indeed, full protection against pneumonic plague was achieved when anti-LcrV antibody titers were comparable with anti-F1 antibody titers in our long-term study. The ability of the polymetric forms of F1 and LcrV antigens to induce both immediate and durable protection suggests they have both TI and T cell-dependent (TD) properties. Polymeric F1 interacted with B1b cells to elicit a rapid immune response,⁴³ but whether the LcrV acts in the same manner requires investigation. B1b cells preferentially recognize protective bacterial antigens.⁴² Therefore, interaction of various plague antigens with B1 cells may delineate their roles and identify new immunogens for plague vaccine development and will be included in future studies.

While this study provides valuable insight into the immediate and long-term protective efficacy of our vaccines against plague, several limitations must be considered. First, while the mouse models used here were Ad5-naive, humans develop high rates of pre-existing neutralizing antibodies against Ad5 (anti-Ad5) after natural infection or vaccination,⁵⁷ which can severely blunt vaccine immunogenicity. Although our Ad5-YFV vaccine still displayed adequate protection against bubonic and/or pneumonic Y. pestis infection in both mice and nonhuman primates (NHPs) with pre-existing anti-Ad5,¹⁸ this resilience to anti-Ad5 may not be seen in humans with unique immunologic characteristics. Next, we acknowledge that the intranasal delivery volume (50 μL) of our Ad5-YFV vaccine in mice could result in deposition of the vaccine in nasal passages as well as the lungs.⁵⁸ The robust systemic and mucosal immune responses observed in our mouse model would then likely reflect a combination of upper and lower respiratory responses, as opposed to a strictly local nasal response in humans. We have previously intranasally instilled the Ad5-YFV vaccine to NHPs via mucosal atomizer in the form of aerosol mist, and robust immune response was elicited.¹⁸ A similar commercially available respiratory fine mist sprayer (Aptar Pharma) may be used for delivering the vaccine in rodents. Despite these limitations, our findings provide the foundation for optimization of mucosal vaccine delivery. Future studies utilizing humanized mouse models or human clinical trials would validate the translational efficacy of our vaccination platforms. Likewise, a mechanistic study on our rationally designed LMA and LMP vaccines will illuminate their distinct immunological profiles. In addition, development of a comprehensive safety profile including formal biodistribution and vector shedding analyses is a necessary next step within the regulatory pipeline for these vaccine platforms.

In summary, this study demonstrated that our LAVs in combination with Ad5-YFV provided both immediate and long-lasting immunity in mice against bubonic and pneumonic plague. Given the socioeconomic burden, the public perception, and the bioterror threat of plague, these vaccination strategies offer assurance through durability and protective efficacy in both reactive and preventive scenarios and should be further evaluated in NHP and human clinical trials.

Materials and methods

Animals

Female 6-week-old inbred C57BL/6 mice (RRID:IMSR_JAX:000664) were purchased from the Jackson Laboratory (Bar Harbor, ME). All animal experiments were ethically performed under an approved Institutional Animal Care and Use Committee (IACUC) protocol in Animal Biosafety Level 3 (ABSL3) facilities of the Galveston National Laboratory (GNL)-Keiller complex at University of Texas Medical Branch at Galveston (UTMB).

Bacterial strains and vaccines

The fully virulent human Y. pestis isolate CO92 was obtained from BEI Resources (NR-641; Manassas, VA), while the F1-negative CO92 strain was a laboratory stock. The live-attenuated vaccine candidates LMA and LMP (Y. pestis CO92; ΔlppΔmsbBΔail [LMA] and ΔlppΔmsbBΔpla [LMP]) and Ad5-YFV were prepared in our laboratory as previously described.^18,19,32 All in vitro studies were performed in a Tier 1 select agent Biosafety Level 3 (BSL-3) laboratory at UTMB.

Vaccination and challenge of animals

In the long-term study, mice (n = 15 per group) were immunized with either a two-dose strategy—LMA or LMP (i.m.) as the prime dose (day 0) and Ad5-YFV (i.n.) as the boost dose (pull; day 21), denoted with a hyphen (e.g., LMA – Ad5-YFV)—or were vaccinated with a 1-dose strategy—LMA or LMP (i.m.) and Ad5-YFV (i.n.) were given simultaneously on day 21, denoted with a plus (e.g., LMA + Ad5-YFV). Mice receiving two-dose PBS were used as unvaccinated controls. LMA and LMP were administered in the hindlimb at a dose of 2 × 10⁶ CFU/50 μL, while Ad5-YFV was administered in the nasal vestibule at a dose of 5 × 10¹⁰ v.p./50 μL. Serum was collected from retro-orbital bleeds 18–19 days after the prime and boost immunizations and at 2-month intervals up to day 196. Collected serum was then pooled from five randomly selected mice, thus generating three biological replicates from 15 mice in each group for each time point for ELISA analysis. On day 200, spleens and BALF from five animals per group were harvested for flow cytometry and IgA antibody analysis, respectively. On day 207, the remaining 10 animals per group were challenged with 100 LD₅₀ of Y. pestis CO92, where 1 LD₅₀ = 280 CFU (colony forming units) for pneumonic challenge in C57BL/6 mice, and observed for mortality for 21 days.

In the short-term emergency use study, mice (n = 10 per vaccination strategy) were immunized simultaneously with LMA or LMP (i.m.) and Ad5-YFV (i.n.) as detailed above. Serum was collected from retro-orbital bleeds of individual animals on days 3 and 7 post-immunizations. On day 7 post-immunization, animals were split into two equal cohorts (n = 5 per group in each cohort) and were challenged with 15 LD₅₀ (1 LD₅₀ = 280 CFU) Y. pestis CO92 by intranasal route (i.n.) and with 100 LD₅₀ (1 LD₅₀ = 10 CFU) of Y. pestis CO92 by subcutaneous route, respectively. Challenged mice were observed for mortality for about 12 weeks. At around 12 weeks (day 95) post-initial challenge, the remaining surviving mice, along with new age0matched naive control mice, were challenged i.n. with 100 LD₅₀ of Y. pestis CO92 and observed for mortality for three additional weeks. Lungs and spleens of challenged animals were harvested for bacterial enumeration after succumbing to infection or at the end of the study.

Antibody titer analysis

Antibody titers were measured via indirect ELISAs as we previously described.^31,32 Absorbance was measured with a VersaMax tunable microplate reader (Molecular Devices, San Jose, CA) at 450 nm. Antibody titers were defined as the inverse of the highest serum dilution giving an optical density (OD) reading above the mean optical OD of negative control (pre-immune) samples. Individual samples or biological replicates were used for the assays.

Neutralizing antibodies against LcrV were measured with a competitive ELISA using LcrV monoclonal antibody Mab7.3 as previously described.^28,32,59 The inverse of the highest dilution of serum that inhibited binding of the biotinylated Mab7.3 to rV by 50% or greater (NT₅₀) represented neutralizing antibody titer.

Flow cytometry analysis

Spleens were collected from individual control and immunized mice as previously described.^28,31,32 For B cell phenotypes, collected cells were directly stained with the following antibodies: CD16/32 Fc Receptor Block (BioLegend, San Diego, CA), fixable viability dye eFluoe 780 (Invitrogen, Carlsbad, CA), FITC anti-mouse CD19 (BioLegend), Alexa Fluor 700 anti-mouse CD3 (BioLegend), Brilliant Violet 650 anti-mouse CD138 (BD Biosciences, San Jose, CA), PE-Cy7 anti-mouse CD38 (BioLegend), PerCP/Cyanine 5.5 anti-mouse GL7 (BioLegend), and Brilliant Violet 510 anti-mouse IgD (BioLegend).

For T cell phenotypes, collected cells were first stimulated with rF1V (100 μg/mL, BEI Resources) for 48 h at 37°C and then treated with 1 × Brefeldin A (BioLegend) for an additional 4 h. Cells were then surface stained with the following antibodies: CD16/32 Fc Receptor Block (BioLegend), fixable viability dye eFluor 780 (Invitrogen), Alexa Fluor 700 anti-mouse CD3 (BioLegend), Brilliant Violet 785 anti-mouse CD4 (BioLegend), FITC anti-mouse CD8 (BioLegend), Brilliant Violet 510 anti-mouse CD44 (BioLegend), Brilliant Violet 711 anti-mouse CD62-L (BioLegend), PE anti-mouse CD127 (BioLegend), Brilliant Violet 605 anti-mouse CD25 (BioLegend), and APC anti-mouse CD134 (BioLegend). Cells were stained intracellularly with Brilliant Ultraviolet anti-mouse CD69 (BD Biosciences), PE/Cyanine7 anti-mouse IL-17A (BioLegend), PerCP/Cyanine5.5 anti-mouse IFNγ (BioLegend), eFluor 450 anti-mouse TNFα (eBioscience), PE-CF594 anti-mouse IL-2 (BD Biosciences), and Brilliant Violet 650 anti-mouse IL-4 (BD Biosciences).

Cells were fixed with 1% paraformaldehyde before analyzing via flow cytometry on the BD FACSymphony A5. Flow cytometry data were prepared via FlowJo software v.10.10.0.

Statistical analysis

Power analysis calculations confirmed that challenging at least five animals per group while assuming 0% survival in unvaccinated controls and 100% survival in vaccinated animals will result in 99% confidence of detecting statistical difference (p ≤ 0.05) between groups. Kaplan-Meier analysis with log rank (Mantel-Cox) test was used to analyze survival. ELISAs were log transformed and presented as geometric means ± standard deviation. Flow cytometry data were presented as arithmetic means ± standard deviation. Statistical analysis was conducted using one-way analysis of variance (ANOVA) with Tuley’s post hoc test. Data were assessed for normality with a parametric Shapiro-Wilk test using an α level of 0.05. Outliers were confirmed using the ROUT method. All data were analyzed using GraphPad Prism 10 software.

Data and code availability

All data have been provided with this paper.

Acknowledgments

These studies were supported by grants from the National Institute of Allergy and Infectious Diseases at the National Institutes of Health (AI153524 and AI071634), The National Institute of Allergy and Infectious Diseases T32 Biodefense Training Program (AI060549), and The University of Texas Medical Branch (UTMB Technology Commercialization Program, John S. Dunn Foundation, and the Kempner Predoctoral Fellowship). Flow cytometry analysis was performed on instruments in the UTMB Flow Cytometry and Cell Sorting Core Lab. Graphical abstract and methodology figures were created in BioRender by Hendrix, E. (2026). The following reagents were obtained through BEI Resources, NIAID, NIH: Yersinia pestis F1-V Fusion Protein, Monomer-Enriched Antigen, Recombinant from Escherichia coli (NR-2561); Y. pestis, Strain CO92 (NR-641); and Y. pestis LcrV Protein, Recombinant from Escherichia coli (NR-32875).

Author contributions

E.K.H., J.S., and A.K.C. designed studies. E.K.H., J.S., P.B.K., B.H.N., A.K.V., and A.K.C. executed the studies. E.B.R. performed ELISAs. E.K.H. analyzed data. J.S. and A.K.C. provided project management. E.K.H., J.S., A.K.C. wrote the manuscript.

Declaration of interests

The described technologies are protected by US Patent numbers RE49514 assigned to The Board of Regents of the University of Texas System, Westport Bio, LLC invented by A.K.C., V.L.M., and E.R. and 10328139 assigned to The Board of Regents of the University of Texas System invented by A.K.C. and J.S..