YopP-Expressing Variant of Y. pestis Activates a Potent Innate Immune Response Affording Cross-Protection against Yersiniosis and Tularemia

Plague, initiated by Yersinia pestis infection, is a rapidly progressing disease with a high mortality rate if not quickly treated. The existence of antibiotic-resistant Y. pestis strains emphasizes the need for the development of novel countermeasures against plague. We previously reported the generation of a recombinant Y. pestis strain (Kim53ΔJ+P) that over-expresses Y. enterocolitica YopP. When this strain was administered subcutaneously to mice, it elicited a fast and effective protective immune response in models of bubonic, pneumonic and septicemic plague. In the present study, we further characterized the immune response induced by the Kim53ΔJ+P recombinant strain. Using a panel of mouse strains defective in specific immune functions, we observed the induction of a prompt protective innate immune response that was interferon-γ dependent. Moreover, inoculation of mice with Y. pestis Kim53ΔJ+P elicited a rapid protective response against secondary infection by other bacterial pathogens, including the enteropathogen Y. enterocolitica and the respiratory pathogen Francisella tularensis. Thus, the development of new therapies to enhance the innate immune response may provide an initial critical delay in disease progression following the exposure to highly virulent bacterial pathogens, extending the time window for successful treatment.


Introduction
Plague is a fatal disease that has caused millions of deaths in three world pandemics, and it remains a public health concern in some regions of the world [1] [2]. The disease, initiated by infection with the bacterial pathogen Yersinia pestis, is manifested in two main forms: pneumonic plague and bubonic plague. Primary pneumonic plague is the more fatal form of the disease, although it is less frequent in nature; it results from the inhalation of Y. pestiscontaining droplets or aerosols. Bubonic plague develops following transmission of the pathogen from rodent reservoirs to humans by infected fleas [1]. Bacteria migrate through the cutaneous lymphatics to regional lymph nodes, where they multiply rapidly, followed by colonization of the spleen and liver and, finally, bacteremia and systemic dissemination to other tissues. Nonspecific symptoms start several days after infection and are followed by the development of swollen lymph nodes, which are termed buboes [3].
The medical treatment of plague should be initiated as early as possible to be effective. Antibiotic administration is the wellestablished first-line therapy against plague. However, the existence of antibiotic-resistant Y. pestis strains highlights the need for the development of new therapeutic tools [4]. Three main strategies have been adopted to achieve this goal: a. Search for anti-infective agents that can inhibit the activity of specific essential virulence systems or virulence factors. These agents include small molecules that target the type III secretion system (T3SS) apparatus and specific T3SS effectors [5][6][7][8]. b. Development of subunit vaccine candidates that extend the spectrum of the immune response generated by the F1/LcrV-based vaccines are also promising [9,10]. c. Attenuated Y. pestis strains have been used in an effort to reveal potential targets for new therapeutic measures or to elucidate elements of the host immune system that are required for the generation of protective immunity against plague [11][12][13][14][15][16] We have recently shown that a newly engineered strain of Y. pestis, Kim53DJ+P, is dramatically attenuated in a mouse model of bubonic plague (LD 50 .10 7 cfu), whereas its virulence is retained in mouse models of pneumonic and systemic plague [17]. This strain was generated by the over-expression of the Y. enterocolitica 0:8 T3SS effector YopP in the virulent yopJ-deleted Kimberley53 Y. pestis strain. YopP is homologous to the endogenous Y. pestis YopJ; both are known to inhibit the pro-inflammatory response in target host cells [18][19][20]. However, the recombinant YopPexpressing strain was found to induce cell death more effectively than the wild-type strain, most likely due to differences in YopJ/ YopP translocation efficiency and allelic variations among the genes encoding these effectors [17,21]. These results are similar to the findings of Brodsky et al., who demonstrated that the ectopic expression of YopP from Y. enterocolitica in Y. pseudotuberculosis led to attenuation following the oral infection of mice [22]. Interestingly, the Y. pestis Kim53DJ+P strain induced a rapid and effective systemic immune response that provided protection against subcutaneous (s.c.), intranasal (i.n.) and intravenous challenges with a fully virulent Y. pestis strain [17].
In the present study, we further characterized the rapid immune response induced by the Kim53DJ+P recombinant Y. pestis strain in an effort to uncover a potential new avenue for the development of protective measures against plague. Using immunodeficient mouse strains, we observed the induction of a prompt but transient protective innate immune response that was interferon-c dependent. This response appeared to postpone disease progression and allow the development of adaptive immunity. Consistent with the relatively low specificity characterizing the rapid induction of the innate response, Kim53DJ+P-infected wild-type mice were also protected from a lethal challenge with the enteropathogen Y. enterocolitica and furthermore with a non-related respiratory pathogen, Francisella tularensis subspecies holarctica LVS.

Bacterial strains
The bacterial strains used in this study are listed in Table 1.

Ethics statement
This study was carried out in strict accordance with the recommendation in the Guide for the Care and Use of Laboratory Animals of the National Institute of Health. All animal experiments were performed in accordance with Israeli law and were

Animal studies
Yersinia strains were grown on BHIA plates (Difco) for 48 hours at 28uC. Several colonies were suspended in a sterile saline solution, diluted to the desired infectious dose and used for the challenge. Mice (6/group, unless stated otherwise), were injected with 0.1 ml of suspension containing the indicated amount of bacterial cells s.c. in their lower or upper back or intraperitoneally (i.p.). Bacterial cell counts were performed by serial dilution plating on BHIA. Unless otherwise indicated, the challenged animals were monitored daily for 35 days. The F. tularensis LVS strain was grown at 37uC to the mid-log phase (optical density of 0.1-0.2 at 660 nm) in TSBC (TSB [Difco] supplemented with 0.1% cysteine). The bacteria were washed and then re-suspended at the desired concentration in PBS. The bacteria were instilled i.n. (25 ml) to ketamine/xylazine-anesthetized mice or injected i.p. (0.5 ml). Experiments evaluating the ability of Y. pestis Kim53DJ+P to protect against Y. enterocolitica and F. tularensis were performed with OF1 outbred mice. Results from animal studies were obtained from at least two repeats.

Antibody titer analysis
Titers of anti-F. tularensis antibodies in serum samples were determined by ELISA in 96-well microtiter plates coated with 100 ml of 10 8 CFU/ml formalin-inactivated LVS. Antibody titers were expressed as reciprocal geometric mean titers (GMTs). The limit of detection for this assay was the reciprocal GMT value of 40.

Statistical analyses
GraphPad statistical software was used for the statistical analyses. Survival curves were compared using the log-rank test. In all analyses, P values equal to 0.05 served as the limit of significance. Results and Discussion

Y. pestis over-expressing YopP activates a transient innate protective response
In mice, the s.c. administration of Kim53DJ+P induces a highly rapid systemic resistance against Y. pestis strains [17]. To characterize the contribution of innate immunity to the Kim53DJ+P-activated protective response, we used immunodeficient RAG1 -/mice, which lack the capacity to mount an adaptive immune response.
Parental C57BL/6J mice and their isogenic RAG1 -/mice were infected s.c. with Kim53DJ+P (10 4 cfu) or with the fully virulent Y. pestis strain Kim53pGFP (10 2 cfu) as a control. As expected, all wild-type C57BL/6J mice survived the Kim53DJ+P infection, whereas those infected with Kim53pGFP succumbed to the disease within seven days ( Figure 1A). All immunodeficient RAG1 -/mice infected with Kim53pGFP succumbed to the disease within nine days, whereas high survival rates of 100-89% were observed for Kim53DJ+P-infected RAG1 -/mice during the initial nine days post-infection ( Figure 1B). This result suggests that following the s.c. infection of RAG1 -/mice with Kim53DJ+P, a rapid protective innate immune response was evoked. However, the protective response was transient as the survival rate declined gradually to 14% by day 35 post-infection (mean time to death of 19 days in the Kim53DJ+P-infected mice compared with 6.6 days in the RAG1 -/mice infected with Kim53pGFP).
In a subsequent cross-protection experiment, we assessed the ability of the innate immune response induced by Kim53DJ+P to provide rapid systemic protection against the challenge with a fully virulent Y. pestis strain inoculated s.c. at a different remote site. RAG1 -/mice were infected s.c. with 10 4 cfu of Kim53DJ+P and were challenged s.c. with a lethal dose of 10 2 cfu of the virulent Kim53pGFP strain three days later. Indeed, the Kim53DJ+Pinduced protective response in the RAG1 -/mice was also effective in postponing the time to death following Kim53pGFP infection ( Figure 1C). Importantly, the survival curve of co-infected mice was statistically indistinguishable from that of mice infected with Kim53DJ+P alone (P = 0.97 by the log-rank test, Figures 1B and  1C). These results indicate that a potent protective innate response can delay mouse mortality following infection with a highly virulent Y. pestis strain and even cure a small but distinct proportion of Y. pestis-infected mice. One may speculate that in wild-type mice this initial delay in disease progression provides the host with the time needed to develop an effective adaptive response against Y. pestis infection.
In an effort to characterize the distinct innate immune response induced rapidly following infection with Kim53DJ+P, we compared the levels of various classes of innate immune cells in the draining inguinal lymph node (ILN) and spleen at 24 and 48 hours post-infection with Kim53DJ+P and the virulent Kim53pGFP strain using flow cytometry. However, no significant differences in the levels of neutrophils, macrophages and dendritic cells were observed between mice infected with Kim53DJ+P and Kim53pGFP (data not shown). Mast cells (MCs) are another type of pro-inflammatory cell located throughout the skin, and they are considered to be among the first cells encountered by invading pathogens. MCs can induce rapid innate immune responses through the release of mediators [23][24][25]. Interestingly, histamine, which is one of these MC mediators, has been shown to be important for controlling Y. enterocolitica infection in mice [26]. These mast cell features led us to evaluate their involvement in the specific immunity induced by s.c. infection with Kim53DJ+P. However, the infection of MC-deficient mice (Kit wsh-/-) with 10 4 cfu of Kim53DJ+P via the s.c. route resulted in 100% survival. Furthermore, all Kim53DJ+P-infected Kit wsh-/mice were protected against the challenge with a lethal dose of the virulent Kim53pGFP strain, which was administered s.c. three days later. These results indicate that MC activation is dispensable for the rapid induction of protective immunity by Kim53DJ+P.

Interferon-gamma is essential for the rapid induction of the Kim53DJ+P-mediated protective response
Interferon-gamma (IFNc) is a pleiotropic cytokine that modulates the development of both the innate and adaptive immune responses [27]. Based on the key role that IFNc plays in the immune system and its previously demonstrated role in the defense against Yersinia species [28][29][30][31][32][33][34][35], we evaluated the contribution of IFNc to Kim53DJ+P-mediated immunity. All mice infected s.c. with 10 4 cfu of Kim53DJ+P succumbed to the disease within four days, which was similar to mice infected with the fully virulent Y. pestis strain (Figure 2). Restoration of the virulent phenotype was unique to Kim53DJ+P, as IFNc-deficient mice infected with other attenuated Y. pestis strains, such as the vaccine strain EV76 or the pCD1 2 /pPCP1 2 derivative of Kim53, did not develop any disease symptoms (Figure 2). These observations clearly indicate that IFNc is a key element of Kim53DJ+P-mediated immunity. Moreover, the rapid kinetics of disease progression and mortality of all Kim53DJ+P-infected IFNc -/mice within four days postinfection suggests that IFNc is important for the induction of protective innate responses by Kim53DJ+P prior to the development of adaptive immunity.
Numerous studies have demonstrated that IFNc can activate innate immune cells to exert a potent antimicrobial effect against intracellular pathogens, including Legionella pneumophila [36], F. tularensis [37,38] and Mycobacterium tuberculosis [39]. However, in the case of Y. pestis, the ripA gene located within the pgm locus was shown to contribute to pathogen survival in macrophages activated by IFNc in vitro following infection [40]. Recent in vivo studies have shown that neutralization of tumor necrosis factor alpha (TNFa) and IFNc increases the lethality and bacterial burden of pgm 2 negative Y. pestis strain D27 in naïve mice infected i.n. [41]. In addition, these cytokines were found to play a role in the F1/LcrV-mediated protective response in various mouse models of plague and are proposed to be released from T cells primed following vaccination [31,34,41]. In the present study, we further extend the observations regarding the contribution of IFNc to host defense against plague and demonstrate that this cytokine plays an essential role in the rapid development of protective immunity against s.c. infection with the pgm-positive Kim53DJ+P Y. pestis strain in naïve mice (Figure 2). The rapid time course of disease progression in Kim53DJ+P-infected IFNc -/mice suggests that adaptive cellular immunity, which typically requires a longer duration to develop, is not involved in the IFNc-mediated early protective response against Kim53DJ+P. To further address this assumption, TCRbd -/mice, which lack functional T cells, were infected s.c. with 10 4 cfu of Kim53DJ+P, and survival was monitored for 35 days. All TCRbd -/mice survived for two weeks after s.c. infection with 10 4 cfu of Kim53DJ+P, suggesting that T cells are not required for the IFNc-mediated early protective response against Kim53DJ+P ( Figure 3A). However, by day 35 post-infection, 50% of the TCRbd -/--infected mice succumbed to the infection ( Figure 3A), indicating that T cell activity is required to augment the early protective innate response against Kim53DJ+P and that this activity makes an important contribution to host defense against Kim53DJ+P infection. Notably, the ability of T cells to help B cells produce antibodies appears to play a minor role in the Kim53DJ+P-mediated defense response, as 90% of mMT B cell-deficient mice ( Figure 3B) and all CD4 T celldeficient mice survived for 35 days following s.c. infection with 10 4 cfu of Kim53DJ+P.
Currently, the exact mechanism by which IFNc is involved in the Kim53DJ+P-mediated early innate immune response is unclear, and requires further studies. It should be noted, however, that the levels of IFNc measured in the spleens of mice during the first 72 hours after infection with Kim53DJ+P were similar to their levels in mice infected with the virulent Kim53pGFP strain [17], suggesting that IFNc might act locally following s.c. infection with Kim53DJ+P. Another possibility is that the expression of IFNc is necessary, but not sufficient for Kim53DJ+P-mediated rapid protective response, and that the ability of the host to overcome the infection depends on another, yet unknown, factor(s).

Kim53DJ+P activates cross-protection against other bacterial pathogens
Non-specific protective responses against secondary infection with heterologous bacterial pathogens have previously been described in the pioneering studies of Mackaness and Henderson in the 1960s [42,43], and in later publications by Killar [44] and Elkins [45]. Additional work by Elkin et al., also described rapid induction of non-specific, IFN-c-mediated protection against F. tularensis infection using bacterial DNA [46]. The rapid activation of innate immunity by Kim53DJ+P raised the intriguing question of whether s.c. infection with Kim53DJ+P would also elicit heterologous protective immunity against other bacterial pathogens.
We first evaluated whether Kim53DJ+P-mediated cross-protection could be induced against another member of the Yersinia genus, the food-born pathogen Y. enterocolitica, which causes gastroenteritis [47]. Mice were infected s.c. with 10 5 cfu of Kim53DJ+P and were challenged i.p. three days later with a lethal dose of 2610 3 cfu of Y. enterocolitica WA O:8 strain. Although all mice infected with 2610 3 cfu of Y. enterocolitica alone succumbed to the disease within 11 days, a high survival rate of 80% was observed when the mice were challenged with Y. enterocolitica three days after infection with Kim53DJ+P ( Figure 4A). This result indicated that the Kim53DJ+P-activated response can indeed provide rapid and effective cross-protection against another pathogen. However, because Y. enterocolitica and Y. pestis belong to the same genus and may share common antigenic determinants, we extended the evaluation of the protective immunity induced by Kim53DJ+P toward a non-related Gram-negative pathogen, F. tularensis, the etiologic agent of tularemia. Mice were infected i.p. and i.n. with a lethal dose of the F. tularensis subspecies holarctica LVS attenuated strain (500 cfu and 5,000 cfu, respectively). All infected mice exhibited disease symptoms (weight loss), and succumbed within ten days (Figures 4B and 4C). However, when the mice were first infected s.c. with Kim53DJ+P and challenged three days later with F. tularensis LVS, high survival rate of 70% was observed following the i.p. challenge and most notably, 63% survived the i.n. challenge (Figures 4B and 4C). In addition, substantial anti-LVS titers (10 3 -10 5 ) were measured in the sera of mice surviving the infection with Kim53DJ+P and secondary  tularensis LVS (square, ten mice/group) or infected s.c. with 1610 4 cfu of Kim53DJ+P and challenged i.p. with 500 cfu of F. tularensis LVS three days later (circle, ten mice/group). (C) Mice were either only infected i.n. with 5000 cfu of F. tularensis LVS (square, eight mice/group) or infected s.c. with 1610 4 cfu of Kim53DJ+P and challenged i.n. with 5000 cfu of F. tularensis LVS three days later (circle, ten mice/group). Mortality was monitored for 16 days post-challenge. doi:10.1371/journal.pone.0083560.g004 challenge with F. tularensis LVS. These findings suggest that the initial non-specific immune response evoked by Y. pestis Kim53DJ+P delayed disease progression, allowing the development of a slower adaptive immune response against the F. tularensis LVS.
Notably, s.c. infection with the facultative intracellular pathogen Y. pestis Kim53DJ+P induced a prompt protective response against an unrelated intracellular pathogen that can replicate in the lung, spleen and liver. One can speculate that IFNc, which was suggested in this study to be essential for the development of rapid protection against Kim53DJ+P, might also provide initial antigenindependent cross-protection against F. tularensis LVS due to its ability to activate macrophages. Indeed, it has been reported that rapid generation of a strong innate immune response mediated by macrophages and NK cells and dependent on IFNc and TNFa, protects naïve mice against lethal infection with F. tularensis LVS [48].
Taken together, the present data indicate that the s.c. infection of mice with Kim53DJ+P activates an extremely rapid and potent protective response in which innate immunity appears to play a critical although transient role. IFNc was found to be a key player in mediating this early, effectively induced protective response. Consistent with the relatively low specificity characterizing the rapid induction of the innate response, Kim53DJ+P-infected wildtype mice were also protected from a lethal challenge with other bacterial pathogens, including the enteropathogen Y. enterocolitica and the respiratory pathogen F. tularensis LVS. Thus, the development of new therapies to enhance the innate immune response may provide an initial critical delay in disease progression following the exposure to highly virulent bacterial pathogens such as Y. pestis, extending the time window for successful treatment.