A Recombinant Attenuated Yersinia pseudotuberculosis Vaccine Delivering a Y. pestis YopE_Nt138-LcrV Fusion Elicits Broad Protection against Plague and Yersiniosis in Mice

In this study, a novel recombinant attenuated Yersinia pseudotuberculosis PB1+ strain (χ10069) engineered with ΔyopK ΔyopJ Δasd triple mutations was used to deliver a Y. pestis fusion protein, YopE amino acid 1 to 138-LcrV (YopE_Nt138-LcrV), to Swiss Webster mice as a protective antigen against infections by yersiniae.

ABSTRACT

In this study, a novel recombinant attenuated Yersinia pseudotuberculosis PB1+ strain (χ10069) engineered with ΔyopK ΔyopJ Δasd triple mutations was used to deliver a Y. pestis fusion protein, YopE amino acid 1 to 138-LcrV (YopE_Nt138-LcrV), to Swiss Webster mice as a protective antigen against infections by yersiniae. χ10069 bacteria harboring the pYA5199 plasmid constitutively synthesized the YopE_Nt138-LcrV fusion protein and secreted it via the type 3 secretion system (T3SS) at 37°C under calcium-deprived conditions. The attenuated strain χ10069(pYA5199) was manifested by the establishment of controlled infection in different tissues without developing conspicuous signs of disease in histopathological analysis of microtome sections. A single-dose oral immunization of χ10069(pYA5199) induced strong serum antibody titers (log₁₀ mean value, 4.2), secretory IgA in bronchoalveolar lavage (BAL) fluid from immunized mice, and Yersinia-specific CD4⁺ and CD8⁺ T cells producing high levels of tumor necrosis factor alpha (TNF-α), gamma interferon (IFN-γ), and interleukin 2 (IL-2), as well as IL-17, in both lungs and spleens of immunized mice, conferring comprehensive Th1- and Th2-mediated immune responses and protection against bubonic and pneumonic plague challenges, with 80% and 90% survival, respectively. Mice immunized with χ10069(pYA5199) also exhibited complete protection against lethal oral infections by Yersinia enterocolitica WA and Y. pseudotuberculosis PB1+. These findings indicated that χ10069(pYA5199) as an oral vaccine induces protective immunity to prevent bubonic and pneumonic plague, as well as yersiniosis, in mice and would be a promising oral vaccine candidate for protection against plague and yersiniosis for human and veterinary applications.

INTRODUCTION

The three human-pathogenic Yersinia species, Yersinia pestis, Yersinia enterocolitica, and Yersinia pseudotuberculosis (1), retain similar mechanisms of virulence but are radically variable in their pathogenicity (2, 3). Y. pestis, the causative agent of plague, is primarily a rodent pathogen transmitted to humans via an infected flea bite or direct contact with infected animals (4). The tier 1 select agent was responsible for the eradication of approximately 200 million human lives in three known historical deadly pandemics (the Plague of Justinian, the black death, and modern plague) (5). In the modern world, plague outbreaks are largely reduced due to improved understanding of Y. pestis transmission pathways, application of broad-spectrum antibiotics, and improving hygiene practices (6). Sporadic outbreaks of plague in Asia and the Americas (7, 8) and a recent large outbreak in Africa (9, 10) indicate that this old disease is still lurking today. High mortality rates and contagious properties, including the sustainability of Y. pestis in airborne droplets, may lead to possible illegitimate applications as a biowarfare agent. The threat of natural or fabricated plague outbreaks is strengthened further with the identification of drug-resistant Y. pestis isolates (11, 12), which may render antibiotic treatment ineffective. Despite vigorous efforts during past decades, no licensed plague vaccine is currently available for human application in the Western world. On the other hand, Y. enterocolitica and Y. pseudotuberculosis are zoonotic foodborne pathogens that spread through the fecal-oral route and cause yersiniosis, a generally self-limiting diarrheal illness in humans (2, 13, 14). Within the United States, Y. enterocolitica is primarily responsible for yersiniosis, with an estimated 2014 case incidence of 28 per 100,000 individuals (http://www.cdc.gov/foodnet). Yersiniosis is a highly prevalent disease in Northern Europe, Scandinavia, and Japan (15, 16) and is the third most commonly reported zoonosis in the European Union (17), resulting in high economic burdens (18). Until now, little attention has been dedicated to vaccine development against yersiniosis. Hence, design and development of a means for prophylactic vaccination are imperative for the long-term prevention and control of plague and yersiniosis.

Y. pestis, Y. enterocolitica, and Y. pseudotuberculosis share approximately 73% genetic identity based on nucleotide sequence analysis (19) and contain similar virulence plasmids (pCD1/pYV) encoding a type three secretion system (T3SS) (20, 21). Y. pseudotuberculosis is the closest ancestor of Y. pestis, sharing >90% genetic identity based on nucleotide sequence comparison and ∼75% protein amino acid sequence identity (22). Y. pseudotuberculosis shows greater genetic stability with a smaller number of insertion sequences than Y. pestis and has a broader host range (such as rodents, dogs, cats, cattle, rabbits, deer, and humans) (22–24). As a foodborne pathogen, Y. pseudotuberculosis is able to be developed as an oral vaccine that has several attractive features compared to parenterally administered vaccines and provides both social and economic advantages, especially in developing countries (25).

Our previous study implied that the attenuated Y. pseudotuberculosis strain χ10057 [Δasd ΔmsbB::P_msbB msbB_(EC) ΔP_crp::TT araC P_araBAD crp] harboring an Asd⁺ plasmid, pYA5199, for Y. pestis YopE amino acid 1 to 138-LcrV (YopE_Nt138-LcrV) antigen delivery by the T3SS induced significant protection (80% survival) against intranasal (i.n.) challenge with 240 median lethal doses (LD₅₀) (2.4 × 10⁴ CFU) of Y. pestis KIM6+(pCD1Ap), which suggested that the YopE_Nt138-LcrV fusion antigen was very immunogenic (26). The Yersinia virulence factors YopK and YopJ, which are associated with the pathogenicity of Y. pseudotuberculosis (27–29) and inhibit CD8⁺ T-cell priming (30), can be eliminated by gene deletion to attenuate the Y. pseudotuberculosis PB1+ strain (31) and enhance its immunogenicity. In this study, we generated a new attenuated Y. pseudotuberculosis strain (χ10069) combining delivery of the Y. pestis YopE_Nt138-LcrV fusion by the T3SS with ΔyopK ΔyopJ Δasd triple mutations to achieve balanced attenuation and enhancement of protective immunity. Mice vaccinated orally with χ10069(pYA5199) demonstrated potent humoral and cell-mediated immune responses, leading to effective protection of the mice against both subcutaneous (s.c.) and intranasal challenges with virulent Y. pestis and oral challenge with Y. enterocolitica WA and Y. pseudotuberculosis PB1+.

RESULTS

Live attenuated Y. pseudotuberculosis with pYA5199 constitutively synthesized YopE_Nt138-LcrV.

Y. pseudotuberculosis PB1+, χ10069(pYA3332), and χ10069(pYA5199) cultures were assessed by Western blot analysis for the synthesis and T3SS-mediated secretion of native LcrV and recombinant fusion YopE_Nt138-LcrV proteins. The molecular masses of the proteins were determined by using prestained protein markers on the nitrocellulose membrane and found to be consistent with the predicted values for LcrV (37.3 kDa) and the YopE_Nt138-LcrV fusion (52 kDa) (Fig. 1). The Y. pseudotuberculosis PB1+ strain was utilized as a positive control. Recombinant protein YopE_Nt138-LcrV was present in χ10069(pYA5199) cell lysates both in the presence and in the absence of Ca²⁺ in the culture medium, whereas the T3SS-mediated secretion of YopE_Nt138-LcrV protein occurred only in the medium deprived of Ca²⁺ and grown at 37°C for 5 h. A similar pattern of native LcrV synthesis and secretion was observed in cultures of Y. pseudotuberculosis PB1+, χ10069(pYA3332), and χ10069(pYA5199) (Fig. 1). The results showed that the recombinant protein YopE_Nt138-LcrV was constitutively synthesized and secreted via T3SS only under Ca²⁺-deprived conditions at 37°C.

FIG 1.

Determination of antigen synthesis and secretion in recombinant Y. pseudotuberculosis strains. The recombinant strains χ10069(pYA3332) and χ10069(pYA5199) were grown to exponential growth phase and induced for 4 to 5 h at 37°C in both the presence and the absence of Ca²⁺ ions to determine T3SS-mediated secretion of recombinant proteins. Y. pseudotuberculosis PB1+ was cultured under the same conditions as a positive control. Whole-cell lysates and proteins secreted in the culture supernatant were assessed by Western blot analysis using anti-LcrV polyclonal antibody.

Persistence of mutant Y. pseudotuberculosis.

To determine the dissemination of χ10069 mutant strains harboring an Asd⁺ plasmid, groups of Swiss Webster mice (9 per group) were orally administered ∼10⁹ CFU of wild-type (WT) Y. pseudotuberculosis PB1+, χ10069(pYA3332), or χ10069(pYA5199) to determine the bacterial loads in Peyer’s patches, livers, spleens, and lungs over a 9-day period. In Peyer’s patches, the numbers of the wild-type PB1+ strain steadily increased at 3, 6, and 9 days postinfection (Fig. 2a). Both χ10069(pYA3332) and χ10069(pYA5199) could rapidly colonize in Peyer’s patches and reached numbers similar to those of wild-type Y. pseudotuberculosis at 3 days postinfection, whereas the numbers of both mutant strains gradually decreased and were significantly lower than those of the wild-type PB1+ strain at days 6 and 9 postinfection (Fig. 2a).

FIG 2.

Assessment of bacterial burden and pathogenesis in mice infected with 10⁹ CFU Y. pseudotuberculosis PB1+, χ10069(pYA3332), or χ10069(pYA5199). (a) Kinetics of appearance of live bacterial titers (CFU) in lungs, livers, spleens, and Peyer’s patches of infected mice at days 3, 6, and 9 postinfection. The experiment was performed twice with equal numbers of animals (3 mice in each group) under similar experimental conditions; the graphs present mean values ± standard deviations (SD) of pooled results. (b) Histopathological analysis of microtomic sections of lung, liver, spleen, and ileum regions of intestine in representative mice from each group at day 6 postinfection. The sham (PBS)-inoculated mouse tissue sections were considered controls to analyze pathological changes in different organs during infections. Tissues were microscopically examined and photographed (bars, 500 nm). Statistical variation among groups at each time point was analyzed by one-way ANOVA applying Tukey’s test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. WP, white pulp; RP, red pulp; He, hemolysis; nec, necrosis; AST, alveolar septum thickening. The boxed areas are enlarged in the insets. The area within the dashed white line shows an inflammatory lesion.

The wild-type strain also effectively colonized the spleens and livers, and bacterial numbers steadily grew at 3, 6, and 9 days postinfection. The χ10069(pYA5199) strain rapidly disseminated into livers at 3 days postinfection, but the χ10069(pYA3332) strain could not. Although the numbers of χ10069(pYA5199) bacteria in livers were significantly lower than those of the wild-type strain, χ10069(pYA5199) delivering the YopE_Nt138-LcrV fusion colonized livers more efficiently than χ10069(pYA3332) at 3 and 6 days postinfection. The numbers of χ10069(pYA3332) and χ10069(pYA5199) bacteria in the livers significantly declined at 9 days postinfection in comparison to those of the wild-type strain, and no bacteria were recovered from livers in half the mice. The χ10069(pYA3332) and χ10069(pYA5199) strains could disseminate to spleens at 6 and 9 days postinfection instead of at 3 days postinfection, and bacterial numbers of the two mutant strains gradually declined in spleens at 6 and 9 days postinfection, following patterns similar to those in livers (Fig. 2a). Interestingly, no CFU of either strain were detectable in the lungs at 3 days postinfection. The detectable CFU of χ10069(pYA3332) and χ10069(pYA5199) appeared by day 6 and decreased by day 9 postinfection. Infection with χ10069(pYA3332) showed the highest bacterial numbers in lungs, but the wild-type PB1+ strain was able to reach the lungs of half the mice by day 9 postinfection (Fig. 2a).

At the 6th day of infection, histopathological analysis of tissue sections in different Y. pseudotuberculosis-infected mice showed that the lung sections of χ10069(pYA5199)- or χ10069(pYA3332)-infected mice displayed moderate inflammation with spongy structures in representative lung sections, indicating alveolar thickening and decrease of airspace area, while Y. pseudotuberculosis PB1+-infected mice displayed mild alveolar inflammation in the lung architecture (Fig. 2b). Complete destruction of the normal architecture of intestinal villi, with ulcerated lumen and dispersed vacuolization with focal necrosis and tissue destruction in livers at day 6 postinfection (Fig. 2b) and severe hemolysis or necrotic cell death with a decrease in the lymphocyte population, was microscopically observed in the white pulp region of the spleen in Y. pseudotuberculosis PB1+-infected mice. In comparison, the χ10069(pYA3332)- and χ10069(pYA5199)-immunized mice demonstrated normal architectures in spleen, liver, and ileum sections with almost parallel results in tissue sections in sham-immunized mice. The histologic scores of each organ from mice orally administered each strain are listed in Table 1.

TABLE 1.

Histologic scores of organs from mice infected with Y. pseudotuberculosis PBI+, χ10069(pYA3332), or χ10069(pYA5199)

Tissue/organ	Infection group	Inflammationa	Hemorrhagea	Cell death or necrosisa
Lung	Naive	−	−	−
	χ10069(pYA3332)	+	+	+
	χ10069(pYA5199)	++	++	++
	PB1+	+	+	+
Spleen	Naive	−	−	−
	χ10069(pYA3332)	+	++	+
	χ10069(pYA5199)	++	++	++
	PB1+	+	+++	+++
Liver	Naive	−	−	−
	χ10069(pYA3332)	+	+	+
	χ10069(pYA5199)	+	+	+
	PB1+	++	++	++
Ileum	Naive	−	−	−
	χ10069(pYA3332)	+	−	−
	χ10069(pYA5199)	−	−	−
	PB1+	−	+++	+++

^a−, absent; +, slight; ++, moderate; +++, severe.

Stimulation of systemic and mucosal immune responses by live attenuated Y. pseudotuberculosis.

Sera and bronchoalveolar lavage (BAL) fluid obtained from χ10069(pYA3332)-, χ10069(pYA5199)-, and sham-immunized mice were evaluated for specific antibody titers and isotype profiles by enzyme-linked immunosorbent assay (ELISA) using purified LcrV as the target antigen. A single-dose oral immunization with χ10069(pYA5199) raised substantially higher anti-LcrV IgG titers in sera than immunization with χ10069(pYA3332) by days 14 and 28. The anti-LcrV IgG titers in mice immunized with χ10069(pYA5199) peaked on day 14 (log₁₀ mean value, 4.2) and stayed stable to the 28th day (Fig. 3a). Antibody titers were further characterized for the isotype profile. The titers of anti-LcrV IgG1, IgG2a, and IgG2b also rapidly increased at 14 days and were slightly higher at 28 days postimmunization. The ratio of IgG2a plus IgG2b to IgG1 was greater than 1.0 in mice immunized with χ10069(pYA5199) (Fig. 3b to d). A moderate increase in serum anti-LcrV IgA titers was recorded in both immunized groups (Fig. 3e). Immunization with χ10069(pYA3332) elicited a comparable pattern of induced antibody titers and isotypes at day 28 postimmunization (Fig. 3b to d). In parallel, the serum antibody titers targeting Y. pestis KIM6+(pCD1Ap) whole-cell lysate (YPL) displayed no substantial difference between χ10069(pYA3332)- and χ10069(pYA5199)-immunized mice for elicited humoral immunity at day 14 and day 28 postimmunization (see Fig. S1a in the supplemental material). The antibody titers against YPL also demonstrated the mixed composition of both IgG1 and IgG2a/IgG2b antibody subtypes (see Fig. S1b). Mixed IgG1 and IgG2a/IgG2b serum antibody responses to LcrV and YPL indicated coinduction of Th1- and Th2-mediated immune responses in mice immunized with these Y. pseudotuberculosis vaccine strains. To characterize the mucosal immune responses, BAL fluids isolated on the 42nd day after immunization were assessed for secretory IgA (sIgA). Mice immunized with χ10069(pYA5199) elicited prominent LcrV-specific IgA titers that were 1.4-fold higher than those induced by χ10069(pYA3332) (Fig. 3f). No conspicuous antibody titers or isotypes were observed in sera or BAL fluid from phosphate-buffered saline (PBS)-inoculated mice.

FIG 3.

Humoral immune responses in immunized mice. (a) Kinetics of LcrV-specific serum antibody titers in immunized mice at days 14 and 28 postimmunization. (b to e) Anti-LcrV antibody isotype profiles in immunized mouse sera collected on days 14 and 28 after prime immunization. (f) Anti-LcrV IgA concentrations in BAL fluid collected from three representative immunized mice on the 42nd day. The data represent mean values and SD. Sera collected from 10 mice (5 male and 5 female) on day 14 and day 28 postimmunization were individually analyzed. Statistical significance among groups at day 14 and day 28 was analyzed by two-way multivariant ANOVA applying Tukey’s post hoc test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

Oral immunization with live attenuated Y. pseudotuberculosis elicits cell-mediated immune responses.

Here, we determined the elicited levels of specific CD4⁺ and CD8⁺ T cells and corresponding cytokines in immunized mouse lungs and spleens using multiplex flow cytometer analysis (Fig. 4 and 5). After 72 h of in vitro induction with YPL or LcrV, lung cells isolated from χ10069(pYA3332)- and χ10069(pYA5199)-immunized mice demonstrated similar increases in both CD4 and CD8 T cell populations (Fig. 4a). The incubation of lung lymphocytes with both YPL and LcrV led to substantial induction of gamma interferon (IFN-γ), tumor necrosis factor alpha (TNF-α), interleukin 17 (IL-17), and IL-2 in both specific CD4⁺ and CD8⁺ T-cell populations (Fig. 4b to e). The increased levels of IL-2 cytokines in both T-cell subsets obtained from lungs of immunized mice supported the stimulation of specific CD4⁺ and CD8⁺ T-cell populations (Fig. 4d). The antigen-stimulated lymphocytes obtained from spleens of χ10069(pYA3332)- and χ10069(pYA5199)-immunized mice also showed heightened induction of CD4⁺ and CD8⁺ T cells (Fig. 5a). Splenocytes from χ10069(pYA5199)-immunized mice produced significantly higher levels of IFN-γ, TNF-α, IL-2, and IL-17 in response to restimulation with the LcrV antigen than cells from χ10069(pYA3332)-immunized mice, while cells from the PBS-immunized mice produced significantly lower levels of these cytokines (Fig. 5b to e). These results suggested that immunization with χ10069(pYA5199) delivering Y. pestis LcrV via the T3SS reinforced an LcrV-specific cellular immune response. However, IL-17A and IL-2 responses in spleen CD4⁺ and CD8⁺ T cells from χ10069(pYA5199)-immunized mice demonstrated less response to YPL than to LcrV (Fig. 5c and d). Splenocytes in the χ10069(pYA3332)-immunized group displayed limited induction of IL-17A- and IL-2-expressing CD4⁺ and CD8⁺ T-cell populations (Fig. 5c and d).

FIG 4.

Analysis of antigen-specific T cells from lungs and associated cytokine responses. After the 42nd day of immunization, lymphocytes were aseptically isolated from mice and in vitro stimulated with Y. pestis LcrV purified protein or YPL at 20 μg/ml for 72 h to assess specific CD4⁺ and CD8⁺ T cells encoding IFN-γ, TNF-α, IL-2, and IL-17. The sham-immunized mouse lung cells were considered controls. (a) CD4⁺ and CD8⁺ T-cell numbers in lungs. (b) CD4⁺ IFN-γ⁺ and CD8⁺ IFN-γ⁺ cell numbers. (c) CD4⁺ IL-17A⁺ and CD8⁺ IL-17A⁺ cell numbers. (d) CD4⁺ IL-2⁺ and CD8⁺ IL-2⁺ cell numbers. (e) CD4⁺ TNF-α⁺ and CD8⁺ TNF-α⁺ cell numbers. Each symbol represents data obtained from an individual mouse. The bars represent mean values plus SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; two-way ANOVA with Bonferroni post hoc tests.

FIG 5.

Analysis of antigen-specific T cells from spleens and associated cytokine responses. After the 42nd day of immunization, splenocytes were aseptically isolated from mice and in vitro stimulated with Y. pestis LcrV purified protein or YPL at 20 μg/ml for 72 h to assess specific CD4⁺ and CD8⁺ T cells encoding IFN-γ, TNF-α, IL-2, and IL-17. The sham-immunized mouse spleen cells were considered controls. (a) CD4⁺ and CD8⁺ T-cell numbers in spleens. (b) CD4⁺ IFN-γ⁺ and CD8⁺ IFN-γ⁺ cell numbers. (c) CD4⁺ IL-17A⁺ and CD8⁺ IL-17A⁺ cell numbers. (d) CD4⁺ IL-2⁺ and CD8⁺ IL-2⁺ cell numbers. (e) CD4⁺ TNF-α⁺ and CD8⁺ TNF-α⁺ cell numbers. Each symbol represents data obtained from an individual mouse. The bars represent mean values plus SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; two-way ANOVA with Tukey’s post hoc tests.

Oral immunization with χ10069(pYA5199) confers comprehensive protection against plague and yersiniosis.

Forty-two days after oral immunization, groups of mice (10 in each group) were challenged with Y. pestis KIM6+(pCD1Ap) by subcutaneous or intranasal administration. Mice immunized with χ10069(pYA5199) were afforded 80% protection against 2.6 × 10⁵ LD₅₀ (2.6 × 10⁶ CFU) of virulent Y. pestis by s.c. infection (Fig. 6a), which was similar to the level of protection observed for mice immunized with χ10069(pYA3332). The χ10069(pYA5199)-immunized mice were found to be 90% protected against 50 times the LD₅₀ (5 × 10³ CFU) of virulent Y. pestis by the intranasal infection route (Fig. 6b). The protective efficacy of the χ10069(pYA5199) strain was further assessed using higher doses of lethal Y. pestis intranasal infection with 500 times the LD₅₀, where up to 70% to 80% of immunized animals remained alive (mean value, 75% survival) (Fig. 6c). At the 14th day postinfection, all surviving animals were found to be healthy without any external signs of disease. Pathogen clearance and localization studies to address the live bacterial burden in lungs, livers, and spleens showed no viable Y. pestis in surviving animals, thus demonstrating induction of sterilizing immunity. In contrast, highly variable protection efficacies (13.3% to 80%) were observed following χ10069(pYA3332) immunization at different challenge dosages (Fig. 6a to c). All sham-immunized mice administered PBS succumbed to infection within 3 to 5 days. The animal study was expanded further to assess protective efficacy against other human-pathogenic yersiniae (Y. pseudotuberculosis and Y. enterocolitica). Vaccination with either χ10069(pYA3332) or χ10069(pYA5199) provided comprehensive protection against yersiniosis, with 100% survival in mice (Fig. 6d and e).

FIG 6.

Survival of mice challenged with virulent Y. pestis KIM6+(pCD1Ap), Y. enterocolitica WA, or Y. pseudotuberculosis PB1+. Groups of mice were orally vaccinated with a dose of 10⁹ CFU of χ10069(pYA3332), 10⁹ CFU of χ10069(pYA5199), or PBS as a negative control. On the 42nd day after immunization, immunized and PBS (sham)-immunized groups of Swiss Webster mice were s.c. challenged with a high infection dose (2.6 × 10⁶ CFU) of Y. pestis (a), Swiss Webster mice were intranasally challenged with Y. pestis at a medium dose (5.0 × 10³ CFU) (b), Swiss Webster mice were intranasally challenged with Y. pestis at a high dose (5.0 × 10⁴ CFU) (c), or groups of mice were orally challenged with Y. enterocolitica WA (2.4 × 10⁹ CFU) (d) or Y. pseudotuberculosis PB1+ (1.05 × 10⁹ CFU) (e). Mortality and morbidity were recorded in surviving mice for 15 days postinfection. Ten mice (5 male and 5 female) in each group were used in the experiment. The LD₅₀ of Y. pseudotuberculosis PB1+ was ∼5 × 10⁷ CFU, and the LD₅₀ of Y. enterocolitica WA was ∼2 × 10⁷ CFU in Swiss Webster mice. The survival and P value were calculated by Kaplan-Meier analysis. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Additionally, we compared the protective efficacies for WT mice, IgG-deficient BALB/c mice, and sIgA-deficient mice immunized with live attenuated Y. pseudotuberculosis strains to further probe the protective roles of IgG and sIgA against pneumonic plague challenge. The results showed that IgG-deficient mice immunized with χ10069(pYA5199) displayed drastically reduced survival versus WT mice (Fig. 7a). Intriguingly, polymeric IgR-deficient (pIgR^−/−) mice lacking sIgA and immunized with strain χ10069(pYA5199) had protection equal to that of WT BALB/c mice immunized with strain χ10069(pYA5199) (Fig. 7b). No protection was observed in sham-immunized mice and in χ10069(pYA3332)-immunized mice (Fig. 7a and b). Both male and female immunized mice were found not to be biased for protection against pathogenic yersiniae among the groups.

FIG 7.

Survival of antibody-deficient mice challenged with virulent Y. pestis KIM6+(pCD1Ap). Groups of mice were orally vaccinated with a dose of 10⁹ CFU of χ10069(pYA3332) or 10⁹ CFU of χ10069(pYA5199) or sham vaccinated as a negative control. On the 42nd day after immunization, Igh-6^−/− mice (a) and pIgR^−/− mice (b) were intranasally challenged with Y. pestis at a medium dose (4.4 × 10³ CFU). Mortality and morbidity were recorded in surviving mice for 15 days postinfection. Ten mice (5 male and 5 female) in each group were used in the experiment. The survival and P value were calculated by Kaplan-Meier analysis. **, P < 0.01; ***, P < 0.001.

DISCUSSION

Live attenuated vaccines are among the most effective means of prophylaxis to control deadly Y. pestis infections in humans and animals (5). The Y. pestis EV series of vaccines are able to induce protective immune responses against bubonic and pneumonic plague in humans, but pathogenesis in rodent and nonhuman primate models and reactogenicity in humans (32–35) limit acceptance of the vaccines worldwide. As an alternative, a single-dose oral administration of a live attenuated vaccine at mucosal inductive sites could be ideal to stimulate humoral and cell-mediated protective immunity against pneumonic plague (31, 36, 37). The transit recirculating lymphocytes (B cells and T cells) at mucosa-associated lymphoid tissues (MALT) may help in the establishment of both systemic and mucosal immunity, rather than an exclusive systemic immune response usually induced by parenteral immunization with vaccines (38, 39). In this study, we sought to engineer a safer Y. pseudotuberculosis strain as an oral live vaccine based on our previous findings (31).

Immune responses tailored to live attenuated Y. pseudotuberculosis vaccines could be augmented by successful infection at a priming site and dissemination to other tissues, but with limited ability to cause infection. We therefore assessed the kinetics of invasion and colonization of live attenuated Y. pseudotuberculosis strains and the pathogenic impacts on lungs, livers, spleens, and intestines of mice. The controlled infection with vaccine strains at the early stage (day 3) and progressive decline at day 9 without developing significant macroscopic or microscopic signs of infection justified the attenuation of strains χ10069(pYA3332) and χ10069(pYA5199). In contrast, the Y. pseudotuberculosis PB1+ strain titers steadily increased in all tissues during the entire course of infection, causing severe tissue injuries (Fig. 2a and b), which was similarly observed in previous studies (40, 41). The Y. pseudotuberculosis PB1+ cells were less prone to escape the confines of the gastrointestinal tract during early infection stages than χ10069(pYA3332) and χ10069(pYA5199) cells (Fig. 2a). This episode was consistent with the findings of Barnes et al. (40), who showed that hepatosplenic colonization by Y. pseudotuberculosis was intimately connected with replication in the intestinal lumen (40). Interestingly, colonization of Peyer’s patches at day 3 postinfection and dissemination to lungs at day 6 postinfection by χ10069(pYA3332) or χ10069(pYA5199) were faster than those by Y. pseudotuberculosis PB1+ (Fig. 2a). These observations are similar to findings reported by Thorslund et al., in which a Y. pseudotuberculosis yopK mutant that is attenuated in mouse infection but fails to cause systemic infection was found to colonize Peyer’s patches and mesenteric lymph nodes more rapidly than the wild-type Y. pseudotuberculosis strain (28). The χ10069 strain with the attribute of fast dissemination might therefore be due to inclusion of the yopK mutation. Similar occurrences were observed in various Y. pseudotuberculosis mutants that we are currently studying (unpublished data). However, the exact mechanism for this observation is not yet clear.

Serum antibody titers, as important correlates of plague protection (41–43), were assessed in immunized mice at different times. Robust LcrV-specific IgG and IgA titers (Fig. 3) induced by χ10069(pYA5199) immunization were correlated with their protective efficacy against lethal challenge by three pathogenic Yersinia strains (Fig. 6). The protective effects of the humoral immune response were further supported by pneumonic plague challenge in IgG-deficient mice immunized with χ10069(pYA5199). Drastically reduced survival of immunized IgG-deficient mice indicated that humoral immune responses were indispensable for achieving comprehensive protection against pneumonic plague (Fig. 7a). A previous study also showed that the passive transfer of convalescent-phase sera obtained from mice infected with Y. pestis could protect both wild-type and B-cell-deficient μMT mice against lethal intranasal Y. pestis infection (36). In the case of bacterial pathogenesis at mucosal surfaces, sIgA plays a central role in host defense by neutralizing and agglutinating pathogens to cause their elimination. Our localization data suggested that the rapid dissemination of Y. pseudotuberculosis to the lungs possibly aided the substantial induction of LcrV-specific sIgA in BAL fluid from χ10069(pYA5199)-immunized mice (Fig. 3c), which might act synergistically in the comprehensive eradication of Y. pestis from surviving mice. However, abrogation of sIgA alone did not weaken protection in pIgR^−/− mice immunized with χ10069(pYA5199) (Fig. 7b), implying that sIgA was dispensable for the protection of animals against pneumonic plague. These results were consistent with a previous claim that IgG, rather than sIgA, was required for a high degree of protection against pneumonic plague challenge (44). However, the natural route of Y. pestis infection is aerosol inoculation rather than i.n. administration to anesthetized mice (45), so the artificial infection route (intranasal instillation) might mask the bona fide role of sIgA associated with protection. Currently, we are unable to carefully examine this due to the limitation of our facilities. In addition, studies reported that IgM could compensate on mucosal surfaces for the lack of IgA (46, 47) and facilitate the removal of foreign pathogens due to its efficient agglutination (48). Moreover, IgM-mediated protection was reported against infection with different pathogens: Haemophilus influenzae (49), Ehrlichia muris (50), Borrelia hermsii (51), Streptococcus pneumoniae (52), and Y. pestis (53). So far, it is unknown whether IgM induced by χ10069(pYA5199) immunization and deposited on respiratory mucosal surfaces in mice could correlate with protection against pneumonic plague. We will pursue this hypothesis in future studies.

Both IgGI and IgG2a/IgG2b antibody isotypes in mice immunized with Y. pseudotuberculosis mutant strains (Fig. 3b to d; see Fig. S1) demonstrated coinduction of mixed Th1 and Th2 immune responses (31). Titers of LcrV-specific IgG2a plus IgG2b (3.6 plus 4.3 log₁₀) in χ10069(pYA5199)-immunized mice were significantly higher than those of IgG1 (4.3 log₁₀) at day 28 postimmunization, indicating a Th1-biased immune response to LcrV (Fig. 3b to d). The presence of CD4⁺ and CD8⁺ T cells expressing IFN-γ, TNF-α, and IL-17A in conjunction with specific neutralizing antibodies (Fig. 3 and 5) could explain the rapid clearance of Y. pestis from immunized mice (Fig. 6). Similar circumstances in previous studies have depicted the pivotal role of T-cell-mediated synergistic induction, along with proinflammatory cytokines (IFN-γ, TNF-α, and IL-17A) and specific antibody titers, in attaining comprehensive protection against pneumonic plague (26, 37, 54–56). Additionally, studies showed that Th17 cells and IL-17A-producing CD4 T cells played an important role in bacterial infections at mucosal sites (57, 58). High-IL-17A-producing CD4 T cells from lungs and spleens of χ10069(pYA5199)-immunized mice were observed when cells were stimulated with LcrV antigen (Fig. 4c and 5c), which might be correlated with high levels of protection against pneumonic plague challenge. Recently, antigen-specific Th17 memory cells in the lung elicited by a sublethal dose of H. influenzae provided broad protection against reinfection by both typeable and nontypeable H. influenzae (59). Lung colonization by χ10069(pYA5199) in a controlled manner (Fig. 2a) might therefore prime Th17 memory cells associated with protection against pulmonary infection with Y. pestis.

Increased secretion of the proinflammatory cytokines IFN-γ and TNF-α in lung cells of χ10069(pYA5199)-immunized mice was observed (Fig. 4b to e) and is classically associated with CD8⁺ T-cell-mediated defense against Y. pestis infection by promoting macrophage microbicidal phagocytosis of intracellular Y. pestis at (lung) sites of infection (60). Intriguingly, we also observed the significant presence of CD8⁺ T cells expressing IL-17A in both lungs and spleens of χ10069(pYA5199)-immunized mice (Fig. 4c and 5c). Current information suggests that IL-17A CD8⁺ T cells may play a potential pathogenic role in human inflammatory disease (61), but the exact roles of IL-17A CD8⁺ T cells in protection or pathogenicity are not well defined.

Mice immunized with χ10069(pYA5199) showed comprehensive protection that exhibited 90% and 80% protection against 50 times the LD₅₀ of Y. pestis by i.n. challenge and 2.6 × 10⁵ LD₅₀ of Y. pestis by s.c. challenge, respectively (Fig. 6a). However, the protective efficacy of χ10069(pYA3332) was less stable, with 15% survival at a very high dose (500 times the LD₅₀) of Y. pestis pneumonic plague challenge, than that of χ10069(pYA5199) (Fig. 6c). The suboptimal concentration of secretory IgA in BAL fluid from χ10069(pYA3332)-immunized mice, along with inferior antibody titers and expression of CD4⁺ CD8⁺ T cells with IFN-γ, TNF-α, and IL-17A, could be responsible for the variable survival rate (15% to 80%) against pulmonary Y. pestis challenge, depending upon the infection doses (Fig. 6b and c).

Overall, the findings in this study defined χ10069(pYA5199) as a superior candidate vaccine against bubonic and pneumonic plague, as well as yersiniosis, with significant capability to concurrently stimulate humoral and cell-mediated immune responses and sterilizing immunity. The comprehensive protection with total bacterial clearance from lungs and spleens was further supported by histopathological analysis and demonstrated that χ10069(pYA5199) with balanced attenuation would be an effective plague and yersiniosis vaccine precursor, but it will need to be validated by further studies in other animal models, including nonhuman primates. Therefore, a safe and effective live attenuated Y. pseudotuberculosis-based oral vaccine would be a promising vaccine candidate for both humans and animal reservoir species, thereby directly and indirectly preventing plague and yersiniosis.

MATERIALS AND METHODS

Ethical statement.

All animal procedures were approved by the Animal Care and Use Committee at Albany Medical College (IACUC protocol 17-02004).

Bacterial strains, plasmids, and growth media.

All the bacterial cultures and plasmids used in this study are listed in Table 2. Escherichia coli χ7213, used as a suicide plasmid donor strain, was grown routinely at 37°C in Luria Bertani (LB) broth (62) or on LB agar plates supplemented with chloramphenicol (50 μg/ml) (Cm) or diaminopimelic acid (50 μg/ml) (DAP) as appropriate. The Y. pseudotuberculosis and Y. enterocolitica WA strains used in this study were routinely grown in LB medium at 28°C. LB plates containing 5% sucrose were used for sacB gene-based counterselection in allelic-exchange experiments for Y. pseudotuberculosis attenuated-strain (χ10069) construction. MacConkey agar supplemented with 1% galactose was used for Y. pseudotuberculosis enumeration in different organs of immunized or infected mice. Y. pestis KIM6+(pCD1Ap) was used for challenge studies under animal biosafety level 3 (ABSL3) containment, as previously reported (63). Y. pestis cells were grown routinely in heart infusion broth (HIB) or HIB plus Congo red agar plates at 28°C (64). HIB-Congo red agar plates were used to confirm the pigmentation (Pgm) phenotype of the Y. pestis pathogenic strain (65).

TABLE 2.

Strains and plasmids used in this study

Strain or plasmid	Genotype or relevant characteristics	Source or reference
Strains
E. coli χ6212	F− λ− ϕ80 Δ(lacZYA-argF) endA1 recA1 hsdR17 deoR thi-1 glnV44 gyrA96 relA1 ΔasdA4	31
E. coli χ7213	thi-1 thr-1 leuB6 fhuA21 lacY1 glnV44 ΔasdA4 recA1 RP4 2-Tc::Mu [λpir]; Kmr	31
Y. enterocolitica WA	Serotype O:8	Robert Perry
Y. pseudotuberculosis PB1+	Serotype O:1b	Robert Perry
Y. pseudotuberculosis χ10067	ΔyopJ315 ΔyopK108	31
Y. pseudotuberculosis χ10069	Δasd206 ΔyopJ315 ΔyopK108	This study
Y. pestis KIM6+(pCD1Ap)	Pgm+ pMT1 pPCP1 pCD1Ap	65
Plasmids
pRE112	Suicide vector; Cmr mob− (RP4)R6K ori sacB	26
pYA3332	Asd+ p15A ori	26
pYA5154	Δasd DNA fragment of Y. pseudotuberculosis PB1+ into XmaI and KpnI sites of pRE112	26
pYA5199	sycE-yopE′ amino acid 1 to 138-lcrV fragment cloned into pYA3332	26

Construction of Y. pseudotuberculosis mutants.

All the primers used in this study are listed in Table S1 in the supplemental material. The construction of a Y. pseudotuberculosis attenuated strain (χ10069) with yopJ and yopK deletions and an Δasd mutation to enable its use as a selective marker for maintenance of Asd⁺ plasmids was performed as described previously (26). In brief, primer sets Asd1/Asd2 and Asd3/Asd4 were used for the PCR amplification of upstream and downstream flanking regions of the Y. pseudotuberculosis PB1+ asd gene. Then, the Δasd DNA fragment was amplified using primer set Asd1/Asd4 by overlapping PCR and cloned into the KpnI/XmaI sites of a suicide plasmid, pRE112, to generate pYA5154 (Table 2). The suicide plasmid pYA5154 (Δasd) was conjugationally transferred from E. coli χ7213 to strain χ10067 (Table 2). Single-crossover insertion strains were isolated on LB agar plates containing Cm. Loss of the suicide vector after the second recombination between allelic exchanges was selected by using a sacB-based sucrose sensitivity counterselection system. The colonies were screened for Cm^s and verified by PCR using primers Asd1/Asd4. A recombinant plasmid (pYA5199) containing Y. pestis yopE_Nt138 (encoding amino acids 1 to 138 of YopE) and full-length lcrV in the pYA3332 plasmid for synthesis of the chimeric protein YopE_Nt138-LcrV was adapted from our previous study (26). χ10069 competent cells transformed with recombinant plasmids pYA3332 and pYA5199 containing the wild-type asdA gene were selected on LB agar plates to establish complementation and antigen delivery.

Western blot analysis.

Cultures of Y. pseudotuberculosis strains grown overnight were reinoculated in freshly prepared LB broth and Ca²⁺-deprived medium (1% tryptone, 0.5% yeast extract, 50 mM MOPS [morpholinepropanesulfonic acid], pH 7.0, 16 mM sodium oxalate, 160 mM magnesium chloride) to grow to an optical density at 600 nm (OD₆₀₀) of 0.8 to 0.9 at 28°C. The T3SS-mediated secretion of native and recombinant proteins was induced by shifting cultures to 37°C for 5 h with constant agitation. Bacterial cells were harvested by centrifugation at 4,000 × g, and secreted proteins in culture media were recovered by trichloroacetic acid (TCA) precipitation. SDS-PAGE followed by Western blot analysis was performed, with equal amounts of protein normalizing the final bacterial concentration in cultures and the secreted-protein concentration from the culture supernatant. An LcrV-specific polyclonal antibody was used to analyze YopE_Nt138-LcrV fusion protein synthesis and secretion (31).

Mice and immunization.

Male and female Swiss Webster outbred mice aged 6 weeks were purchased from Charles River. BALB/c background pIgR^−/− mice from Jackson Laboratories and BALB/c background Igh-6^−/− mice from Taconic Laboratories were maintained in a pathogen-free animal facility at Albany Medical College. The mice were allowed 1 week for environmental adaptation before immunization or infection experiments. Morbidity, including weight loss, and mortality were monitored daily in all studies inside ABSL2 and ABSL3 facilities.

The Y. pseudotuberculosis χ10069(pYA3332) and χ10069(pYA5199) cultures grown overnight were freshly inoculated into LB broth at 28°C with constant agitation (180 rpm) until the exponential growth phase (OD₆₀₀, 0.8 to 0.9). The bacterial cells were collected by centrifugation and concentrated 100-fold using PBS, pH 7.4. Mice deprived of food and water for 6 h were orally administered a single dose (∼10⁹ CFU; 20 μl) of bacterial culture. Sham-immunized mice were given a volume of sterile PBS equal to that given to negative controls (31).

In vivo assessment of bacterial localization.

Mice were orally administered ∼10⁹ CFU (20 μl) of Y. pseudotuberculosis strain χ10069(pYA3332), χ10069(pYA5199), or PB1+ resuspended in PBS to assess the kinetics of bacterial localization. At days 3, 6, and 9 postinfection, 3 mice from each group were euthanized, and the bacterial burdens in different vital organs, i.e., lung, liver, spleen, and Peyer’s patches, were evaluated by the conventional plate count method. Organs removed aseptically were weighed and washed 2 or 3 times with sterile PBS to remove blood and free surface bacteria. The precisely weighed organs (lung, liver, spleen, and Peyer’s patches) were homogenized in ice-cold PBS (pH 7.4) using a bullet blender (Bullet Blender Blue; Averill Park, NY, USA) at power level 7 for 2 min. Tissue homogenates from Y. pseudotuberculosis-infected animals were serially diluted and plated on MacConkey agar plates. Each count was confirmed on duplicate plates with a minimum of 2 dilutions to determine the titers in bacteria per gram of tissue.

Antibody titer and isotype profiling.

On days 14 and 28 postimmunization, each mouse was bled by submandibular vein puncture. The collected sera were stored at −20°C until further use. LcrV-specific antibody titers in sera were determined by the following standard protocol. The 96-well LcrV-coated plates were incubated with 2-fold serially diluted serum samples in the range from 1:200 to 1:51,200. The colorimetric reaction was developed by alkaline phosphatase-conjugated secondary antibodies in the presence of nitrophenyl phosphate disodium substrate (Sigma). Sera at the highest dilution giving OD₄₀₅ values a minimum of 2-fold higher than those for sham-immunized mouse serum were considered the antibody titer (66). For BAL fluid collection, 3 representative mice from each group were euthanized on the 42nd day postimmunization, and BAL fluid was collected by performing tracheotomies according to the method of Chen et al. with slight modifications (67). In brief, euthanized mice were exsanguinated from the heart chambers to minimize blood contamination. The mice were lavaged 3 times with 500 μl of ice-cold sterile PBS (pH 7.4) via a tracheal tube using a tubing adaptor (BD Biosciences, USA). The collected BAL fluid was spun at 1,500 rpm for 5 min at 4°C to remove cellular contents and stored at –20°C until further analysis. Antibody isotype profiles in serum and the sIgA titer in BAL fluid were determined by using a Southern Biotech isotyping kit (5300-01; mouse immunoglobulin panel) following the manufacturer’s protocol.

Flow cytometric analysis of T cells.

Spleens and lungs taken aseptically from euthanized animals were dissociated using 70-μm strainers to obtain single cells. The lysed individual red blood cell (RBC) populations (2 × 10⁶) were seeded in 12-well cell culture plates and in vitro stimulated for 72 h with 20 μg/ml YPL or purified Y. pestis LcrV protein. Four hours before the collection of cells, the culture medium in each well was supplemented with brefeldin A and monensin cocktail (1:1 ratio) to plug Golgi apparatus-mediated cytokine secretion. For the flow cytometric analysis of T-cell populations and their corresponding cytokines, induced cells were harvested and resuspended in fluorescence-activated cell sorter (FACS) staining buffer containing CD16/32 antibodies (1:200) for 10 min on ice. T-cell-specific markers were stained using anti-mouse CD3 (fluorescein isothiocyanate [FITC]), CD4 (phycoerythrin [PE]), and CD8 (allophycocyanin [APC]) antibodies, followed by intracellular-cytokine (IFN-γ, peridinin chlorophyll protein [PerCP] Cy5.5; TNF-α, HV510; IL-2, PECy7; IL-17A, APC Cy7) staining using BioLegend Perm-fix solution and buffer according to the manufacturer’s protocol. The entire staining process was performed at 4°C or on ice with 1 to 2 h of incubation at each step. Events acquired on BD flow cytometers (LSRII) with FACS Diva software were analyzed using FlowJo v.10.

In vivo protection study and survival.

Forty-two days after immunization, the mice were shifted to ABSL3 to perform lethal challenges to mimic bubonic and pneumonic plague. For pneumonic plague infection, the Y. pestis KIM6+(pCD1Ap) culture grown overnight was reinoculated in 10 ml HIB medium with xylose (0.2% [wt/vol]), CaCl₂ (2.5 mM), and ampicillin (100 μg/ml) and grown at 37°C with constant agitation to an OD₆₀₀ of 0.8. The bacterial cells harvested by centrifugation at 4,000 rpm for 12 min at room temperature were resuspended in PBS at appropriate dilutions. Ketamine (xylazine-ketamine, 1:5 ratio)-anesthetized mice were given 40 μl Y. pestis (10³ to 10⁴ CFU) challenge via the nostril, while bubonic plague infection was established in the mice by subcutaneous injection with 2.6 × 10⁶ CFU of Y. pestis harvested from freshly cultured HIB containing xylose (0.2% [wt/vol]) and CaCl₂ (2.5 mM) at 28°C. The mortality and morbidity of the infected mice were observed daily for the next 15 days.

Cultures of Y. enterocolitica WA or Y. pseudotuberculosis PB1+ virulent strains grown overnight were reinoculated in freshly prepared LB broth and grown to exponential growth phase at 28°C. Then, the cultures were shifted to 37°C for 2 h to induce the T3SS-mediated secretion of effector components. The bacterial cells were harvested and diluted to the proper concentrations with PBS. The immunized and sham-immunized groups of mice were orally challenged with 2.4 × 10⁹ and 1.05 × 10⁹ CFU of Y. enterocolitica WA or Y. pseudotuberculosis PB1+ culture, respectively. The infected mice were observed daily for 15 days postinfection to assess the protective efficacies of the vaccine strains.

Histopathological analysis.

Tissue samples obtained from representative immunized, sham-immunized control, and Y. pseudotuberculosis-infected groups of mice were immediately fixed in a 10% neutral buffered formaldehyde solution. After 24 h of fixation, paraffin-embedded tissue blocks were microtome sectioned and hematoxylin-eosin stained following a standard protocol. Each tissue section was observed and scored by a semiquantitative method (−, absent; +, slight; ++, moderate; and +++, severe) under a light microscope with a minimum 5-point assessment of pathological conditions (68). Spleens were scored for the enlargement of white pulp areas, lymphocytolysis and cell death in the white pulp, and congestion and hemorrhage in the red pulp. Lungs were photographed and evaluated for alveolar wall thickening with decreased airspace areas and hemorrhage. Livers were scored for hepatocellular necrosis, hepatocellular degeneration, and hemorrhage causing discoloration. The distal small intestine (ileum) was evaluated for inflammation in mucosa or submucosa, along with degradation of villus structures.

Statistical analysis.

Each experiment included a significant number (a minimum of 3) of biological replicates with 2 or 3 repeats performed in a synchronized fashion to establish reproducibility. Statistical analyses of data among groups were evaluated by one-way analysis of variance (ANOVA)/univariate or two-way ANOVA using Tukey’s post hoc tests, whereas the significance of IgA titers in BAF samples was assessed using paired t tests. The data were analyzed using GraphPad Prism 8.0 software. The data are represented as mean values ± standard deviations, with significance indicated.