Deletion Mutants of Francisella Phagosomal Transporters FptA and FptF Are Highly Attenuated for Virulence and Are Protective Against Lethal Intranasal Francisella LVS Challenge in a Murine Model of Respiratory Tularemia

Francisella tularensis (Ft) is a Gram-negative, facultative intracellular bacterium that is a Tier 1 Select Agent of concern for biodefense for which there is no licensed vaccine. A subfamily of 9 Francisella phagosomal transporter (fpt) genes belonging to the Major Facilitator Superfamily of transporters was identified as critical to pathogenesis and potential targets for attenuation and vaccine development. We evaluated the attenuation and protective capacity of LVS derivatives with deletions of the fptA and fptF genes in the C57BL/6J mouse model of respiratory tularemia. LVSΔfptA and LVSΔfptF were highly attenuated with LD50 values of >20 times that of LVS when administered intranasally and conferred 100% protection against lethal challenge. Immune responses to the fpt mutant strains in mouse lungs on day 6 post-infection were substantially modified compared to LVS and were associated with reduced organ burdens and reduced pathology. The immune responses to LVSΔfptA and LVSΔfptF were characterized by decreased levels of IL-10 and IL-1β in the BALF versus LVS, and increased numbers of B cells, αβ and γδ T cells, NK cells, and DCs versus LVS. These results support a fundamental requirement for FptA and FptF in the pathogenesis of Ft and the modulation of the host immune response.


Introduction
Francisella tularensis (Ft) is a Gram-negative, facultative intracellular, coccobacillus responsible for the zoonosis tularemia [1]. Ft has a broad host range and persists in the environment, causing an average of~250 natural cases of human tularemia in the U.S. each year [2]. While Ft can be contracted by various routes, most commonly through the bite of an insect (ticks) and contact with infected animal products or carcasses, it can be easily aerosolized, and inhalation of as few as 10 colony forming units of virulent Ft can cause disease with up to a 60% mortality rate if untreated [3][4][5][6][7][8][9]. Ft has a history of weaponization, and combined with the small inoculum, high mortality rate and ease of dissemination, has been listed by the Centers for Disease Control and Prevention as a Tier 1 Select Agent of concern for biodefense. Importantly, there is no licensed vaccine which could protect was highly attenuated in the C57BL/6J mouse model of tularemia, but disappointingly only provided partial protection against virulent Type A Ft challenge [37]. Because this candidate did not provide a high level of protection, our current study focuses on the promising but previously uncharacterized FptA and FptF transporters, which may prove to be better targets for attenuating mutations. In addition to furthering the advancement of an efficacious vaccine against Ft, we utilized the LVS∆fptA and LVS∆fptF mutant strains to understand the contributions of these Fpt transporters to the pathogenesis of Ft. We found that both fpt genes are critical for the induction of a pathological inflammatory response that facilitates pathogenesis and results in host death, and deletion of either gene results in vaccine candidates that provide protection against lethal challenge.

fpt Mutants Are Attenuated in the C57BL/6J Murine Model of Respiratory Tularemia Compared to Parental LVS
LVS strains with a complete deletion of the fptA or fptF gene products were constructed utilizing suicide plasmids as previously published [34] to generate the LVS∆fptA and LVS∆fptF mutant strains studied in this work. Previous in vitro studies demonstrated that LVS∆fptA and LVS∆fptF were not deficient for extracellular growth in broth culture or for intracellular replication in J774.1 macrophages, a BALB/c murine macrophage cell line [34]. However, both strains were attenuated for virulence following intraperitoneal (i.p.) injection in BALB/c mice; 100% of animals (n = 4 per group) survived infection with 3 × 10 3 or 6 × 10 3 CFU of LVS∆fptA or LVS∆fptF, respectively, compared to no survivors following LVS infection with~450 CFU (Supplementary Table S1). In this current work, we have moved to the C57BL/6J murine model because this mouse strain has been demonstrated to be the most stringent model for protection against lethal Ft and is prevalent in the literature for evaluating immune responses to Ft [14,38]. Additionally, we have utilized intranasal (i.n.) administration because this more closely mimics respiratory tularemia. In this model, we reassessed the levels of attenuation of the FptA-and FptF-based mutant strains. Groups of C57BL/6J mice were inoculated i.n. with similar doses of parental LVS, LVS∆fptA, or LVS∆fptF, or mock-infected with PBS, and survival was monitored for 28 days or until mice met criteria for euthanasia. An i.n. dose of 350 colony forming units (CFU) of LVS resulted in death of 100% of mice by day 10 post-infection ( Figure 1A). In contrast, 100% or 91% of mice survived infection with similar doses of LVS∆fptA or LVS∆fptF, respectively ( Figure 1A). 100% of mock-infected mice survived with no associated weight loss or other clinical signs of disease ( Figure 1A,B). Whereas LVS-infected mice exhibited lethargy and ruffling of fur and lost > 20% of their initial starting weight (which met euthanasia criteria) by day 10 post-infection with this dose, the LVS∆fpt mutant-infected mice inoculated with a similar dose only lost an average of~10% of their initial starting weight, exhibited little to no clinical signs, and recovered to mock-infected levels ( Figure 1B). Extended dose studies were performed to determine LD 50 values for each strain. Whereas the LD 50 value of LVS is <60 CFU, the LVS∆fptA and LVS∆fptF mutant strains were highly attenuated with LD 50 values of~1400 CFU, and between~2010 and 5167 CFU, respectively (Table 1). These results demonstrate that the LVS∆fptA and LVS∆fptF mutant strains are attenuated for virulence in this model compared to LVS, and highlights their potential as vaccine candidates. 1 Surviving mice were challenged i.n. with ~500 CFU of LVS on day 29 post-infection and followed for another 21 days post-challenge. 2 Vaccine efficacy was calculated using the following formula: (Attack rate unvaccinated − Attack rate vaccinated) ÷ (Attack rate unvaccinated) × 100. 3 Sterilizing immunity was determined by lung burden enumeration 21 days post-challenge. 4 One mouse was removed from analysis in the LVSΔfptA group receiving ~1120 CFU due to inefficient dose delivery. 5 One mouse for the intended LVSΔfptA group receiving ~1680 CFU died in the animal facility for unknown reasons prior to inoculation. 6 Three mice were removed from analysis in the LVSΔfptF group receiving ~670 CFU because these mice did not have adequate access to water and exhibited severe weight loss due to dehydration.   1 Surviving mice were challenged i.n. with~500 CFU of LVS on day 29 post-infection and followed for another 21 days post-challenge. 2 Vaccine efficacy was calculated using the following formula: (Attack rate unvaccinated − Attack rate vaccinated) ÷ (Attack rate unvaccinated) × 100. 3 Sterilizing immunity was determined by lung burden enumeration 21 days post-challenge. 4 One mouse was removed from analysis in the LVS∆fptA group receiving~1120 CFU due to inefficient dose delivery. 5 One mouse for the intended LVS∆fptA group receiving~1680 CFU died in the animal facility for unknown reasons prior to inoculation. 6 Three mice were removed from analysis in the LVS∆fptF group receiving~670 CFU because these mice did not have adequate access to water and exhibited severe weight loss due to dehydration.

Inoculation with LVS∆fptA or LVS∆fptF Confers Protection against Lethal Challenge with Parental LVS in the C57BL/6J Murine Model of Respiratory Tularemia
Given the high level of attenuation of the LVS∆fptA and LVS∆fptF mutant strains (Table 1), we hypothesized that these mutant strains would confer protection against lethal challenge with parental LVS. Surviving mice from the LD 50 experiments were challenged on day 29 post-infection with a lethal dose of LVS,~500 CFU i.n., to determine vaccine efficacy. At this challenge dose, there was a 95.8% mortality rate for the unimmunized control group. Control mice exhibited clinical signs of disease, including lethargy and ruffling of fur, and lost > 20% of their initial starting weight, requiring euthanasia, by day 10 post-challenge. In contrast, all immunized mice, regardless of the vaccinating strain and the dose, were protected against lethal challenge with LVS with no clinical signs of disease and little to no loss of weight. A subset of mouse lungs harvested at 21 days post-challenge was tested for challenge strain clearance. No challenge organisms were cultured from the lungs which supports sterilizing immunity conferred by these vaccine strains (Table 1).

fpt Mutants Have Reduced Bacterial Organ Burdens in the Lungs, Livers, and Spleens of C57BL/6J Mice versus Mice Infected with Parental LVS
Next, the mechanisms of attenuation of LVS∆fptA and LVS∆fptF were investigated in the mouse model by first evaluating their ability to colonize the host lung and disseminate to systemic organs, which are characteristics of tularemia that we hypothesized may be compromised in these mutant strains versus LVS. Groups of C57BL/6J mice were inoculated i.n. with~350 CFU of parental LVS, LVS∆fptA or LVS∆fptF, and organ burdens were enumerated from the lungs (primary site of infection), livers, and spleens on days 1, 3, 6, 14 and 21 post-infection. A dose of~350 CFU of LVS is 100% lethal but is a welltolerated dose of the fpt mutant strains; this target dose was maintained for all remaining experiments. One day post-infection, the bacterial burdens in the lungs were similar regardless of infecting strain ( Figure 2A) which suggests that FptA and FptF are not essential for initial colonization. By day 3 post-infection, the bacterial burdens in the lungs increased significantly as all strains have begun replicating (Figure 2A), and bacteria were present in the spleen ( Figure 2B) and liver ( Figure 2C) demonstrating that these strains are capable of dissemination to systemic organs. However, by day 6 post-infection, the organ burdens of both mutant strains in the lungs, livers, and spleens were markedly reduced compared to LVS which continued to replicate to high numbers between days 3 and 6 post-infection (Figure 2A-C). Mice that were infected with LVS did not survive past day 10 post-infection, and therefore organ burden data only exists for the fpt mutant strains from days 14 and 21 post-infection. Whereas the organ burdens of LVS-infected mice continued to rise through day 6 post-infection, after which time these mice meet euthanasia criteria, the organ burdens of fpt mutant-infected mice reached a peak organ burden at day 3 and then steadily declined through 21 days post-infection when most of the bacteria were cleared from mouse organs (Figure 2A A lack of symbols at any of the designated time points of harvest indicates an absence of bacteria in these organs. The limit of detection was 10 2 CFU/gram of tissue. All fpt strains had statistically significant lower bacterial burdens than LVS in mouse lungs on day 6 (p < 0.05).

Pathology Is Less Severe in the Lungs of C57BL/6J Mice Infected with fpt Mutants than in the Lungs of Mice Infected with Parental LVS
The initial host response to respiratory tularemia is characterized by an acute lung infiltration of macrophages and neutrophils in an attempt to clear infection prior to dissemination [39,40,41,42]. This is followed by the development of necrotic foci in the lungs, liver, and spleen, which is one of the hallmark features of end-stage tularemia and indicative of extensive cell death and disseminated disease [43]. Due to the reduced bacterial organ burdens of the fpt mutants at day 6 post-infection in mouse lungs, livers, and spleens ( Figure 2), we hypothesized that the fpt mutants may be attenuated due to a differential host response that is able to clear the fpt mutants with reduced resultant pathology. Accordingly, the pathological consequences of infection and the corresponding changes in the murine lung tissue in response to infection with the fpt mutants compared to parental LVS were assessed. To measure the extent of pulmonary involvement in the lungs after i.n. inoculation with LVS, LVSΔfptA, or LVSΔfptF, or mock ✱ Figure 2. Colonization and dissemination of fpt mutant strains in the murine model of respiratory tularemia. The bacterial organ burdens in the lungs (A), spleens (B) and livers (C) of eight-week-old male and female C57BL/6J mice following infection with LVS or LVS∆fpt mutant strains were enumerated. Groups of 4 mice per strain per time point were inoculated i.n. with~350 CFU of either LVS, LVS∆fptA or LVS∆fptF and organs were harvested for bacterial enumeration on days 1, 3, 6, 14, and 21 post-infection. Symbols indicate burdens in individual mice. *, p < 0.05 by a two-tailed t test. A lack of symbols at any of the designated time points of harvest indicates an absence of bacteria in these organs. The limit of detection was 10 2 CFU/gram of tissue. All fpt strains had statistically significant lower bacterial burdens than LVS in mouse lungs on day 6 (p < 0.05).

Pathology Is Less Severe in the Lungs of C57BL/6J Mice Infected with fpt Mutants than in the Lungs of Mice Infected with Parental LVS
The initial host response to respiratory tularemia is characterized by an acute lung infiltration of macrophages and neutrophils in an attempt to clear infection prior to dissemination [39][40][41][42]. This is followed by the development of necrotic foci in the lungs, liver, and spleen, which is one of the hallmark features of end-stage tularemia and indicative of extensive cell death and disseminated disease [43]. Due to the reduced bacterial organ burdens of the fpt mutants at day 6 post-infection in mouse lungs, livers, and spleens ( Figure 2), we hypothesized that the fpt mutants may be attenuated due to a differential host response that is able to clear the fpt mutants with reduced resultant pathology. Accordingly, the pathological consequences of infection and the corresponding changes in the murine lung tissue in response to infection with the fpt mutants compared to parental Pathogens 2021, 10, 799 7 of 24 LVS were assessed. To measure the extent of pulmonary involvement in the lungs after i.n. inoculation with LVS, LVS∆fptA, or LVS∆fptF, or mock infection with PBS, lungs were harvested at day 6 post-infection (the time point at which there were statistically significant differences in organ burdens in the lung between LVS and the fpt mutants, which was maintained for the remainder of the studies). Hematoxylin and eosin (H&E) stained lung tissues were prepared for analysis. The histopathologic parameters examined and the scoring criteria are described in Supplementary Table S2. Representative images are shown for the scored slides which highlight the more prominent acute inflammatory changes and increased extent of tissue destruction in the LVS group, and to a lesser extent, the LVS∆fptA group compared to the LVS∆fptF group ( Figure 3). LVS caused significant damage by all measures assessed; histopathology was characterized by extensive inflammation, diffuse alveolar destruction, and the presence of neutrophils, macrophages, and lymphocytes ( Table 2). Diffuse alveolar destruction due to inflammation caused by neutrophilic, macrophagic, and lymphocytic infiltrates is evident in the LVS-infected lungs ( Figure 3A,B). In contrast, infection with LVS∆fptA or LVS∆fptF was characterized by reduced destruction and pulmonary involvement ( Table 2). Mice infected with LVS∆fptA had intermediate histopathologic scores that were lower than LVS but higher than LVS∆fptF; compared to LVS, there was reduced inflammation and alveolar destruction and reduced numbers of neutrophils and lymphocytes ( Table 2). Despite lower overall scores than LVS, LVS∆fptA induced microscopic manifestations that included some parenchymal destruction due to inflammation, interstitial inflammation in areas of preserved parenchyma, chronic inflammatory infiltrates, and fibrin deposits ( Figure 3C,D). Infection with LVS∆fptF induced the lowest histopathology scores in all measures assessed, especially in terms of inflammation, alveolar destruction, fibrin deposition and neutrophilic infiltration ( Table 2). Images of lungs infected with LVS∆fptF show less prominent inflammatory manifestations that were limited to interstitial lymphocytic inflammation with hyperplasia of type II pneumocytes, and small foci of inflammation, but unlike the lungs infected with LVS or LVS∆fptA, lung architecture has been preserved ( Figure 3E,F). Mock infection did not generate any inflammation or destruction of the lung ( Figure 3G,H; Table 2). In summation, mouse lungs infected with the LVS∆fptF mutant showed the least pathology when compared to the levels of tissue destruction caused by LVS, followed by LVS∆fptA. These histopathology results are consistent with their respective levels of attenuation ( Table 1) which confirms that virulence is linked to the inflammatory response.

fpt Mutants Induce Altered Proinflammatory Cytokine Responses in the Bronchoalveolar Lavage Fluid of C57BL/6J Mice Compared to Parental LVS
The pathology associated with LVS infection is partly due to the ability of LVS to cause a cytokine storm, which is one of the hallmark features of tularemia [44][45][46]. Because

fpt Mutants Induce Altered Proinflammatory Cytokine Responses in the Bronchoalveolar Lavage Fluid of C57BL/6J Mice Compared to Parental LVS
The pathology associated with LVS infection is partly due to the ability of LVS to cause a cytokine storm, which is one of the hallmark features of tularemia [44][45][46]. Because of the reduced lung pathology and inflammatory infiltrates associated with infection with the fpt mutants ( Figure 3, Table 2), we hypothesized that the fpt mutants induce modified cytokine responses in the lungs versus parental LVS. Secreted cytokines (IFN-γ, TNF-α, IL-1β, IL-10, IL-12p70, KC/GRO, IL-6, IL-5, IL-4, and IL-2) in the bronchoalveolar lavage fluid (BALF) were quantified on day 6 post-infection following i.n. inoculation with LVS, LVS∆fptA, or LVS∆fptF, or mock infection with PBS using a 10-plex electrochemiluminescent detection method. Despite some variation between individual mice, BALF samples from LVS-and fpt mutant-infected mice showed significantly elevated levels of IFN-γ, TNF-α, IL-1β, IL-12p70, IL-10, IL-2, and KC/GRO compared to mock infection ( Figure 4). Levels of IL-4, IL-5, and IL-6 were not significantly elevated in any infection group versus mock infection (Supplementary Figure S1). There were significant decreases in the levels of secreted IL-1β and IL-10 in the fpt mutant-infected BALF samples versus LVS; both fpt mutant strains induced significantly less IL-1β and IL-10 secretion in the lungs at day 6 post-infection versus LVS ( Figure 4A,C). While not statistically significant, the median levels of IL-2 and IL-12p70 were elevated in the BALF from the fpt mutant-infected mice versus LVS ( Figure 4B,D). Furthermore, the median level of KC/GRO was reduced in the BALF from fpt mutant-infected mice versus LVS ( Figure 4G). The levels of IFN-γ and TNF-α were similar between fpt mutant-infected and LVS-infected BALF samples ( Figure 4E,F). Upon closer analysis of cytokine secretion in individual mice, the mice that produced high levels of IFN-γ had comparably high levels of TNF-α, IL-1β, etc., and the mice that had minimal levels of IFN-γ also had low levels of the other cytokines. The differences in proinflammatory cytokines elicited by the fpt mutants versus LVS further support a role for these Fpts in the virulence of Ft and the modulation of host immune responses to infection.  Bars represent means with SEM from duplicate measurements from one experiment. *, p < 0.05 by a two-tailed t test. All treatment samples were statistically significantly upregulated versus the mock-infected control. One BALF sample from the LVS∆fptF group was excluded from analyses due to inefficient BALF harvest.

fpt Mutants Induce Altered Host Immune Responses in the Lungs of C57BL/6J Mice Compared to Parental LVS
Given the modified cytokine profiles ( Figure 4) and histopathological evidence ( Figure 3, Table 2) of an altered immune response to the fpt mutants, we next hypothesized that early immune cell infiltration to the lungs of fpt mutant-infected mice may differ from LVS-infected mice. To interrogate this, immune cell infiltration in the lungs of mice infected with LVS, LVS∆fptA, or LVS∆fptF was evaluated using flow cytometry. Mouse lungs were harvested at day 6 post-infection following i.n. inoculation with LVS, LVS∆fptA, or LVS∆fptF, or mock infection with PBS. Single cell suspensions were processed for analysis using the gating strategy shown in Supplementary Figure S2. Analysis revealed changes in the overall cell numbers and composition in the LVS-and fpt mutant-infected lungs compared to the mock-infected animals. As expected, there was a significant increase in the number of live CD45.2+ immune cells in the LVS-and fpt mutant-infected versus mock-infected lungs ( Figure 5A). This was mostly due to increased numbers of non-T and non-B cells in the infected lungs ( Figure 5D). There was an increase in the number of T cells ( Figure 5B) and B cells ( Figure 5C) present in the lungs of LVS∆fptA-infected mice, and to a lesser, non-statistically significant extent, in the lungs of LVS∆fptF-infected mice, compared to LVS-infected lungs. Among the T cell subsets, LVS infection caused a decrease in αβ T cells that was not observed with LVS∆fptA or LVS∆fptF infection ( Figure 5E). The numbers of γδ T cells and NKT-like cells were not different in LVS-versus mock-infected lungs, but were increased in LVS∆fptA-infected lungs, and the numbers of γδ T cells in the LVS∆fptF-infected lungs were intermediate between the numbers in LVS∆fptA-and LVS-infected lungs ( Figure 5F,G).
infected with LVS, LVSΔfptA, or LVSΔfptF was evaluated using flow cytometry. Mouse lungs were harvested at day 6 post-infection following i.n. inoculation with LVS, LVSΔfptA, or LVSΔfptF, or mock infection with PBS. Single cell suspensions were processed for analysis using the gating strategy shown in Supplementary Figure S2. Analysis revealed changes in the overall cell numbers and composition in the LVS-and fpt mutant-infected lungs compared to the mock-infected animals. As expected, there was a significant increase in the number of live CD45.2+ immune cells in the LVS-and fpt mutant-infected versus mock-infected lungs ( Figure 5A). This was mostly due to increased numbers of non-T and non-B cells in the infected lungs ( Figure 5D). There was an increase in the number of T cells ( Figure 5B) and B cells ( Figure 5C) present in the lungs of LVSΔfptA-infected mice, and to a lesser, non-statistically significant extent, in the lungs of LVSΔfptF-infected mice, compared to LVS-infected lungs. Among the T cell subsets, LVS infection caused a decrease in αβ T cells that was not observed with LVSΔfptA or LVSΔfptF infection ( Figure 5E). The numbers of γδ T cells and NKT-like cells were not different in LVS-versus mock-infected lungs, but were increased in LVSΔfptA-infected lungs, and the numbers of γδ T cells in the LVSΔfptF-infected lungs were intermediate between the numbers in LVSΔfptA-and LVS-infected lungs ( Figure 5F,G). Among the non-B and non-T cell populations, the numbers of neutrophils ( Figure 6B) and non-B Class II+ cells ( Figure 6C) were increased in LVS-and fpt mutant-infected versus mock-infected lungs; however, while not significant, there was a gradient of neutrophil influx where LVS induced the most neutrophils, followed by LVS∆fptA and LVS∆fptF, which corresponds to the histopathology data ( Figure 3, Table 2). The number of NK cells was increased in the LVS∆fptA-and LVS∆fptF-infected lungs compared to both mock-and LVS-infected lungs ( Figure 6A). When the non-B Class II+ cell populations were subtyped, LVS failed to induce DC infiltration in contrast to the LVS∆fptA and LVS∆fptF mutants ( Figure 6D). The numbers of CD11b+ macrophages ( Figure 6E) and Class II+ neutrophils ( Figure 6F) were increased in the LVS-and fpt mutant-infected lungs versus mock-infected. Additionally, the numbers of CD11b− macrophages were reduced in LVS-infected versus fpt mutant-infected lungs ( Figure 6G). Finally, when the CD11b− macrophages were further characterized as alveolar macrophages based on the expression of CD68 and CD11c, there were significantly fewer alveolar macrophages present in the LVS-infected lungs versus mock; this decrease did not occur in LVS∆fptA-and LVS∆fptF-infected lungs ( Figure 6H). Whereas infection with all strains resulted in significant innate immune responses, differential cell infiltration in response to the fpt mutants versus LVS reflects the attenuation of these mutant strains and the induction of immune responses that promote host survival, bacterial clearance, and eventual protection as opposed to the pathological immune response to LVS that results in death of the infected animal.  Figure S4H) by the total number of CD11b− macrophages (G).

Discussion
The 2001 anthrax attacks in the U.S. emphasized the need for the development of diagnostic, preventative, and therapeutic measures against Select Agents of concern for use in biowarfare, including Ft. LVS was extensively studied in volunteers where it demonstrated incomplete protection against aerosol challenge with the virulent Type A SchuS4 strain. While great effort has been placed on developing alternative vaccines, it is generally accepted that any new vaccine candidate needs to outperform LVS in terms of safety, immunogenicity, and protective efficacy against aerosolized Type A Ft

Discussion
The 2001 anthrax attacks in the U.S. emphasized the need for the development of diagnostic, preventative, and therapeutic measures against Select Agents of concern for use in biowarfare, including Ft. LVS was extensively studied in volunteers where it demonstrated incomplete protection against aerosol challenge with the virulent Type A SchuS4 strain. While great effort has been placed on developing alternative vaccines, it is generally accepted that any new vaccine candidate needs to outperform LVS in terms of safety, immunogenicity, and protective efficacy against aerosolized Type A Ft [11,12,14,15,[47][48][49][50]. Few LVS-based vaccines have shown protective efficacy against respiratory tularemia caused by heterologous Type A challenge [51][52][53]. Even though there is evidence to suggest that an effective live attenuated tularemia vaccine would likely need to be engineered in the Type A strain, LVS still proves a useful tool to identify targeted mutations that could be generated in the Type A background. The generation of live attenuated bacterial vaccines with defined mutations in metabolic genes has proven to be a successful strategy for vaccine development against Shigella and Salmonella Typhi, among others [54][55][56][57][58][59], and has been utilized in the generation of vaccine candidates against Ft. Several distinct metabolic mutations have been introduced into Ft to generate live attenuated, auxotrophic mutants with varying degrees of attenuation and protective efficacies [60][61][62][63][64][65][66][67][68]. Other metabolic targets for attenuation include MFS transporters such as Fpts [34,37]. In this study, we investigated two MFS transporters, FptA and FptF, which have not yet been fully characterized, as attenuating targets for vaccine development.
By all criteria tested, both LVS∆fptA and LVS∆fptF were attenuated for virulence and the host responses to infection with these mutants were substantially modified. The reported LD 50 of LVS varies depending on route of infection, mouse strain, and methodology [69][70][71]. Using a protocol that incorporates a 50 µL dose volume and deep anesthesia [72,73], the LD 50 of LVS was determined to be < 60 CFU by the i.n. route in C57BL/6J mice which highlights the extreme virulence of LVS in the mouse model and validates LVS as a useful surrogate model to study the pathogenesis of virulent Type A Ft. The low LD 50 of LVS also emphasizes the attenuation levels of the LVS∆fptA and LVS∆fptF mutant strains; the LD 50 values of both fpt mutants in this model were > 20 times that of LVS. These values supported the evaluation of their ability to confer protection against challenge, and we found that a single inoculation of LVS∆fptA or LVS∆fptF, regardless of the vaccination dose, provided 100% vaccine efficacy against a lethal challenge of~500 CFU of LVS, which is >8 times the LD 50 .
Infection with Ft via the respiratory route begins with bacterial entry into and replication within alveolar macrophages, neutrophils, dendritic cells, monocytes and alveolar type II cells [40,[74][75][76] for the first 48-hours followed by dissemination to the liver and spleen [40][41][42], which is consistent with our results. The fpt mutant strains were able to colonize the lungs and disseminate to the livers and spleens, initially replicating to the same levels as LVS by day 3 post-infection. However, the fpt mutants failed to reach the same peak organ burdens as LVS by day 6 post-infection at which time the numbers of mutant bacteria declined. Furthermore, the fpt mutants were almost cleared by day 21 post-infection; clearance is an important consideration for live attenuated vaccine studies, as a strain that is unable to be cleared would not be a viable vaccine candidate.
The host initially responds to Ft with an acute lung infiltration of macrophages and neutrophils, which become infected shortly after exposure [39][40][41][42]. During severe respiratory tularemia, the alveolar spaces become filled with neutrophils, macrophages, necrotic cellular debris, and bacteria [42], which may contribute to the resultant severe pneumonia and pathology. However, this acute immune response is insufficient to clear Ft from the lungs and prevent dissemination, and instead, contributes to pathogenesis; uncontrolled bacterial replication and neutrophil infiltration contribute to the extensive tissue destruction and the unrestrained production of proinflammatory cytokines in what is known as the cytokine storm, which is associated with significant pathology, extensive organ damage, multiple organ failure, the development of severe sepsis, and eventual host death [38,39,45,71,[77][78][79][80][81][82][83]. This response is consistent with our organ burden, histopathology and flow cytometry data showing increasing organ burdens (or the failure to control replication and dissemination) and extensive inflammation, diffuse alveolar destruction, and the presence of neutrophils, macrophages, and lymphocytes in LVS-infected lungs on day 6 post infection. However, fpt mutant-infected lungs showed reduced pathology and altered host responses characterized by reduced inflammation, including reduced neutrophil infiltration, and reduced alveolar destruction. This evidence suggests that modified host responses to the fpt mutant strains occurred that (1) controlled the bacterial burden as demonstrated by the reduced organ burdens at day 6 post-infection, (2) induced minimal tissue pathology and damage, which reduced the severity of disease, and (3) promoted the survival of mice infected with fpt mutants.
The pathology associated with the cytokine storm during LVS infection is one of the hallmark features of tularemia [44][45][46]84]. In good agreement with prior reports [85], we observed that Ft induced high levels of the potent proinflammatory cytokines IFN-γ, TNF-α, IL-1β, and KC/GRO. Mouse knockout models have demonstrated critical roles for IFN-γ and TNF-α in the immune response against Ft as knockout mice succumb to tularemia more rapidly than wildtype mice [86][87][88][89][90][91][92][93]. It is therefore speculated that elevated IFN-γ and TNF-α are necessary but not sufficient for survival against Ft infection. Encouragingly, we found that the fpt mutants induced similarly elevated levels of IFN-γ and TNF-α in the BALF as LVS, which may contribute to the development of protective immune responses in fpt mutant-infected mice that are not seen in the LVS-infected mice that succumb by day 10 post-infection, which is not sufficient time for the development of the adaptive response. KC/GRO is a marker of the severe sepsis associated with tularemia [78] which was correspondingly elevated in the LVS-infected BALF. However, the fpt mutants induced reduced levels of KC/GRO in the BALF versus LVS, suggesting a reduction of severe sepsis in the fpt mutant-infected mice and providing an additional measure of their attenuation. The role for IL-1β in the defense against Ft is not yet well understood; however, IL-1β in the BALF of C57BL/6J mice i.n. infected with a lethal dose of LVS is higher at day 3 post-infection than at day 6 post-infection [94]. Therefore, one could speculate that early IL-1β secretion is insufficient to promote bacterial clearance and host survival. Others have shown that hypersecretion of IL-1β in mice infected with attenuated mutant strains of LVS was protective and associated with survival [95]; we observed the opposite phenomenon wherein our fpt mutant strains induced significantly less IL-1β than LVS, a difference that highlights the incomplete understanding of the role of IL-1β in the response against LVS. Furthermore, LVS induced the production of the anti-inflammatory cytokine IL-10 which has been shown to be produced by activated lung DCs early in the course of respiratory tularemia [74,80,84] and by regulatory T cells (Tregs) [96] which are present in the lungs as soon as one day post-infection [80]. It seems confounding that a pathogen capable of producing a cytokine storm would also induce antiinflammatory mediators. However, the timeline of cytokine secretion is an important factor; it has been speculated that Ft induces IL-10 early during the course of respiratory tularemia to dampen or delay the initial pro-inflammatory cytokine response [84], which may explain the subsequent reduction in IL-1β secretion, facilitating bacterial growth and survival [80]. We found that the fpt mutants failed to induce comparably elevated levels of IL-1β and IL-10 as LVS. As IL-1β is released from immune cells upon pyroptosis, it is possible that the reduced histopathology in the fpt mutant-infected lungs can be attributed to decreased pyroptosis and IL-1β secretion compared to LVS. Additionally, the reduced levels of IL-10 in the fpt mutant-infected lungs may be indicative of the fpt mutants' inability to invoke an anti-inflammatory cytokine milieu, as observed during LVS infection, providing further evidence that FptA and FptF are important factors required for the pathogenesis of Ft. These differences in pro-and anti-inflammatory cytokines elicited in response to the fpt mutants versus LVS underscore the balance between pathological inflammation (as seen in LVS-infected mice) and protective inflammation (as seen in fpt mutant-infected mice) and support the contribution of the fptA and fptF gene products in the modulation of the host inflammatory response by LVS.
In addition to reduced lung histopathology and altered cytokine responses in the BALF of fpt mutant-infected mice, immune cell infiltration in lungs of fpt mutant-infected mice was modified compared to LVS. LVS infection was characterized by an increase in non-T and non-B cells, specifically neutrophils and non-B Class II+ cells (including CD11b+ macrophages and Class II+ neutrophils), with a concomitant decrease in T cells and alveolar macrophages compared to mock-infected lungs. Others have reported similar innate immune cell profiles in a murine model of respiratory tularemia [40]. Mature neutrophils typically do not express Class II, but a novel phenomenon in which granulocytes (including neutrophils) acquire antigen presenting capabilities that influence T cell responses has been reported [97][98][99]. A significant proportion of the neutrophil population was Class II+ in the LVS-and fpt mutant-infected lungs and may represent the first evidence of this phenomenon occurring during Ft infection. Generally, responses to the fpt mutant strains, especially the LVS∆fptA mutant strain, and to a lesser extent, the LVS∆fptF mutant strain, were characterized by increases in B cells, T cells, NK cells, DCs, and CD11b− macrophages versus LVS. It has been shown that strong and early protection against sublethal doses of LVS is dependent on B cells through limiting bacterial growth early after infection, and this protective mechanism is dependent on IFN-γ [100]. Therefore, it is possible that the increased presence of B cells in the fpt mutant-infected lungs on day 6 post-infection may contribute to the observed reduction in organ burdens among these mice at this time point. Furthermore, the increased presence of DCs in the fpt mutant-infected lungs may also contribute to the reduction of fpt mutant organ burdens as DCs have been shown to help limit early pathogen replication [101]. The increased levels of IL-10 in the LVS-infected BALF may underlie the reduced levels of T cells in the LVS-infected lungs; IL-10 suppresses T cell proliferation. Conversely, the reduced levels of IL-10 in the fpt mutant-infected BALF versus LVS may explain the increased presence of T cells in the fpt mutant-infected lungs. Further studies are needed to identify the source of IL-10 in the murine lung during respiratory tularemia. There is evidence of NK regulation of T cell activity during respiratory tularemia [102]. Additionally, it is thought that another role for NK cells in respiratory tularemia is to lyse infected alveolar macrophages [102]. We speculate that the severe lung pathology and reduced population of alveolar macrophages in the LVS-infected lungs may be due in part to lysis of these macrophages, perhaps by NK cells, which have become heavily infected with Ft. It has also been shown that γδ T cells can directly activate NK cells [103], and therefore the increase in NK cells in the fpt mutant-infected lungs versus LVS-and mock-infected lungs may be reflective of the increased presence of γδ T cells in the fpt mutant-infected lungs. Taken together, B cells in fpt mutant-infected mice may be receiving more T cell help as a result of increased antigen presentation by DCs and higher NK cell activity in mutant-infected lungs. The differential host responses to the fpt mutants reflect their levels of attenuation and may indicate a protective response, perhaps due to less pathological inflammation. Additional studies at later time points are needed to identify the implications of these early responses on the development of protective immunity against subsequent challenge.
This highly sensitive murine model of respiratory tularemia allowed the dissection of many parameters of the host pulmonary responses to LVS and to the fpt mutants following i.n. inoculation. In agreement with the literature, LVS induced a highly pathogenic immune response that contributes to the lethality of LVS in mice. The differences between the histopathology, BALF cytokines, and the immune cell composition between LVS-and fpt mutant-infected lungs allowed us to evaluate the contribution of FptA and FptF to the pathogenesis of Ft. We speculate that the respective (yet unknown) substrates of the FptA and FptF transporters are essential for Ft to reach a critical threshold of bacterial burden in the host to facilitate full virulence and pathogenesis. Alternatively, the fpt mutants may be impaired such that they are unable to induce the required virulence gene profiles for pathogenesis and are instead cleared through a more protective and less pathological immune response by the host. Further studies involving the analysis of the in vivo transcriptomic profiles of the fpt mutants versus LVS are needed to identify the specific effects of deleting the fpt gene products on virulence gene expression. Ultimately, the differential host responses to the fpt mutant strains versus to LVS support the essential roles for FptA and FptF in the modulation of the host immune response such that LVS can establish conditions to permit bacterial replication, spread, pathogenesis and eventual host death. These data support the further interrogation of fptA and fptF genes in the virulent Type A SchuS4 strain towards the ultimate goals of developing live attenuated tularemia vaccines and understanding the contributions of Fpts in the pathogenesis of Ft.

Mice
Wildtype C57BL/6J mice were purchased from The Jackson Laboratory (The Jackson Laboratory, Bar Harbor, ME, USA) for use in i.n. attenuation, organ burden enumeration, and protective capacity studies. For Supplementary Table S1; wildtype BALB/c mice were purchased from Charles River Laboratories (Charles River Laboratories, Frederick, MD, USA) for use in i.p. attenuation studies. Infected mice were housed in the University of Maryland animal biohazard safety level 2 (ABSL-2) facility for the duration of the studies. All animal experiments were conducted according to protocols approved by the UMB Institutional Animal Care and Use Committee (IACUC).

Attenuation of fpt Mutant Strains in Mice
Six-to eight-week-old female BALB/c mice (Charles River Laboratories, Frederick, MD, USA) were initially used to test the fpt mutant strains in vivo for attenuation. Groups of 4 BALB/c mice per group were injected i.p. with either LVS or the fpt mutant, or mockinfected with PBS. To reassess the attenuation levels of fpt mutants in a more stringent model of protection against lethal Ft following vaccination, groups of 4 or 6 eight-to nineweek-old male and female C57BL/6J mice per dosage concentration were anesthetized i.p. with a cocktail of 90 mg/kg of ketamine and 9 mg/kg of xylazine and inoculated i.n. with either LVS or the fpt mutant suspended in 50 µL of PBS, or mock-infected with PBS. Studies utilizing dosage groups near the LD 50 values were repeated to confirm the LD 50 values. Mice were monitored daily for survival and clinical signs of infection (weight loss, lethargy, and ruffling of fur), and weights were monitored every other day for 28 days post-infection. Mice were assigned health assessment scores based on general physical appearance (skin, fur, etc.), ambulation, potential dehydration, response to stimuli, movement, evidence of food consumption (feces in bedding), grooming and any clinical sign that may indicate disease or illness (lethargy, hunched posture, head tucked into abdomen, rough hair, visible tumors, etc.). Mouse health was scored based on the following criteria: condition 1 for normal activity, where mice are bright, alert, reactive, healthy, shiny, have normal posture and lack dehydration; condition 2 for mice that are quiet, alert, reactive with early stage piloerection, normal posture and mild dehydration; condition 3 for mice that are quiet and not very reactive, and exhibit mild piloerection, dull coats, hunched posture, difficult ambulation, moderate dehydration, and scruff remains tented (lifted) for 2-3 s; and condition 4 for mice that are non-reactive with severe piloerection, dull and dirty coats, hunched and squinting, severely dehydrated, tented scruff that when lifted does not return, severe weight loss (>20%) and respiratory distress. Mice reaching a clinical score of 4 or losing >20% of their body weight were euthanized as required by the UMB IACUC.

Protective Capacity of fpt Mutant Strains in Mice
The surviving mice from the "Attenuation of fpt Mutant Strains in Mice" studies listed above were utilized for challenge studies to determine protective capacity of the fpt mutant strains. Twenty-eight days post-vaccination, the surviving mice were anesthetized and inoculated i.n. with~500 CFU of LVS suspended in 50 µL of PBS. After challenge, all mice were monitored daily for clinical signs and weights were monitored every other day as described above for 21 days post-challenge. Vaccine efficacy was calculated using the following formula: (Attack rate unvaccinated − Attack rate vaccinated) ÷ (Attack rate unvaccinated) × 100.

Quantification of Bacterial Organ Burdens in Mice
To determine the bacterial organ burdens of fpt mutant strains in vivo, groups of 4 eight-week-old male and female C57BL/6J mice per strain per time point were anesthetized and inoculated i.n. with~350 CFU of either LVS or the fpt mutant suspended in 50 µL of PBS. Mice were euthanized at the designated time points post-infection, and organs were harvested to enumerate bacterial load. Organs were halved for use in quantification of bacteria organ burdens and histopathology analysis. Lungs, livers, and spleens from infected mice were homogenized using the BeadBug Microtube Homogenizer (Benchmark Scientific, Sayreville, NJ, USA) according to manufacturer's protocol in PBS. The bacterial load was enumerated by serial plating on MHA plates.

Histopathology Analysis of Infected Mouse Lungs
To assess the pathological consequences and changes in lung tissue in response to infection with the fpt mutants, groups of 4 eight-week-old male and female C57BL/6J mice were anesthetized and inoculated i.n. with~350 CFU of either LVS or the fpt mutant suspended in 50 µL of PBS. Mice were euthanized at day 6 post-infection and lungs were harvested for histopathology analysis. Lungs were fixed in 10% Formalin, neutral buffered (Sigma-Aldrich, St. Louis, MO, USA) for a minimum of 24 h. After fixation, the tissues were processed and embedded in paraffin. Tissue blocks were sectioned at 5 µm and stained with hematoxylin and eosin by the University of Maryland School of Medicine Histology Core Laboratory. Stained tissues were examined on a Aperio ScanScope CS (Aperio, Vista, CA, USA) and assessed by two pathologists. Both observers were blinded regarding the type of infection and performed their scoring independently. The histopathologic parameters examined and scoring criteria are described in Supplementary Table S2.

Quantification of Secreted Cytokines in Murine Bronchoalveolar Lavage Fluid
To measure the secreted cytokines in the BALF in response to infection with the fpt mutants, LVS, or mock infection with PBS, groups of 5 eight-week-old male and female C57BL/6J mice were anesthetized and inoculated i.n. with~350 CFU of either LVS or the fpt mutant suspended in 50 µL of PBS, or mock-infected with PBS. Mice were euthanized on day 6 post-infection by i.p. anesthetic overdose (cocktail of 360 mg/kg of ketamine and 36 mg/kg of xylazine) followed by secondary exsanguination. Bronchoalveolar lavage fluid was harvested in 1 mL of sterile PBS via an intratracheal catheter attached to a 1mL syringe, and BSA (Reagent Diluent Concentrate 2, R&D Systems, Minneapolis, MN, USA) was added for a final concentration of 1%. BALFs were frozen at −80 • C until cytokines were analyzed using the V-PLEX Proinflammatory Panel 1 Mouse Kit assay based on an electrochemiluminescent (ECL) detection method (Meso Scale Diagnostics, LLC, Rockville, MD, USA) according to the manufacturer's recommendations. Data were collected using the MESO QUICKPLEX SQ 120 instrument and analyzed with DISCOVERY WORKBENCH ® V4.0 data analysis software (Meso Scale Diagnostics, LLC, Rockville, MD, USA). Briefly, samples were thawed on the day of analysis and standards and BALF samples (50 µL) were added to appropriate wells and incubated for two hours at room temperature with shaking (700 rpm). The fluid was then removed, and the wells were washed 3 times with wash buffer (0.05 % Tween-20 in PBS). Detection antibodies were added to each well and incubated for 2 h at room temperature with shaking, followed by washing 3 times. After washing, 150 µL of 2× reading buffer was added to each well. The plate was analyzed on the MSD instrument immediately. For the purposes of statistical analyses, any value that was below the lowest limit of detection (LLOD) for the assay was replaced with half of LLOD of the assay.

Statistical Analysis
A two-tailed t test was used to determine statistical significance in organ bacterial burden experiments. Results were considered statistically significant at a p value of <0.05. A two-tailed t test was used to determine statistical significance in cytokine secretion in BALF samples. Results were considered statistically significant at a p value of <0.05. A one-way ANOVA with a Tukey's post-test was used to determine statistical significance in the flow cytometric analyses. Results were considered statistically significant as follows: ****, p < 0.0001; ***, p = 0.001; **, p < 0.001; *, p < 0.05. Data analysis was performed using GraphPad Prism 9.0 (GraphPad Software, Inc., San Diego, CA, USA).

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/10 .3390/pathogens10070799/s1, Table S1: Attenuation of fpt mutant strains in BALB/c mice, Table S2: Scoring criteria for histpathology analysis, Figure S1: Cytokine secretion in the bronchoalveolar lavage fluid (BALF) of mice infected with fpt mutant strains, Figure S2: Gating strategy for flow cytometric analysis of infected lung samples, Figure S3