TLR2 Signaling is Required for the Innate, but Not Adaptive Response to LVS clpB

Toll-like receptor 2 (TLR2) is the best-characterized pattern-recognition receptor for the highly pathogenic intracellular bacterium, Francisella tularensis. We previously identified a mutant in the live vaccine strain (LVS) of Francisella, LVS clpB, which is attenuated, but induces a protective immune response. We sought to determine whether TLR2 signaling was required during the immune response to LVS clpB. TLR2 knock-out (TLR2 KO) mice previously infected with LVS clpB are completely protected during a lethal challenge with LVS. Furthermore, the kinetics and magnitude of the primary T-cell response in B6 and TLR2 KO mice are similar indicating that TLR2 signaling is dispensable for the adaptive immune response to LVS clpB. TLR2 signaling was important, however, for the innate immune response to LVS clpB. We identified three classes of cytokines/chemokines that differ in their dependence on TLR2 signaling for production on day 3 post-inoculation in the bronchoalveolar lavage fluid. IL-1α, IL-1β, IL-2, IL-17, MIP-1α, and TNF-α production depended on TLR2 signaling, while GM-CSF, IFN-γ, and VEGF production were completely independent of TLR2 signaling. IL-6, IL-12, IP-10, KC, and MIG production were partially dependent on TLR2 signaling. Together our data indicate that the innate immune response to LVS clpB requires TLR2 signaling for the maximal innate response, whereas TLR2 is not required for the adaptive immune response.

One hallmark of pneumonic tularemia caused by wild-type Francisella is the near absence of an innate immune response in the lung despite high bacterial burdens (14). An in-frame deletion of the clpB gene in the live vaccine strain (LVS) of F. tularensis subsp.
holartica results in bacteria that lack the ability to inhibit host innate immune (15). ClpB is a highly conserved chaperone protein of the AAA+ superfamily of ATPases, which mediate protein disaggregation (16). ClpB has not been shown to be a TLR2 ligand, nor has it been shown to affect the expression of identified TLR2 ligands (17). Intranasal inoculation of C57Bl/6J and BALB/cJ mice with LVS clpB significantly increases the concentration of proinflammatory cytokines and chemokines in the bronchoalveolar lavage fluid (BALF) 3 days post-inoculation compared to LVS inoculated mice (15). Despite a robust innate immune response during LVS clpB infection, adaptive immunity is required for bacterial clearance and the frequency of IFN-γ producing CD4 + and CD8 + T-cells is similar in mice inoculated with LVS or LVS clpB (15). We and others have demonstrated that vaccination with clpB mutants in both LVS and the highly virulent F. tularensis subspecies tularensis (SchuS4) provide protection during lethal, wild-type challenge (15,(17)(18)(19).
Due to the well-characterized role of TLR2 during the immune response to wild-type Francisella, we sought to determine whether TLR2 was required during the immune response to LVS clpB. TLR2 KO mice were able to clear LVS clpB; clearance, however, was delayed compared to B6 mice inoculated with LVS clpB. Additionally, TLR2 KO mice previously infected with LVS clpB survived lethal LVS challenge. The ability of TLR2 KO mice to survive a lethal secondary challenge was not surprising given that the T-cell response in B6 and TLR2 KO mice was similar on days 7 and www.frontiersin.org 10 post-inoculation during the primary infection. Together, these data indicated that TLR2 signaling is dispensable during the primary and secondary T-cell response. However, TLR2 signaling was required for the maximal innate immune response to LVS clpB. We identified three classes of cytokines and chemokines in the BALF that differed in their requirement of TLR2 signaling for production on day 3 post-inoculation (TLR2 independent, TLR2 dependent, and TLR2 partially dependent). Together, these data indicated that while TLR2 is critical during the innate immune response, TLR2 signaling is dispensable during the primary adaptive immune response and a secondary challenge.

BACTERIA
Francisella tularensis subspecies holarctica LVS with an in-frame deletion of clpB (FTL_0094) was generated as previously described (15). Wild-type LVS was obtained from the CDC (Atlanta, GA, USA). Bacteria were grown at 37°C on chocolate agar supplemented with 1% IsoVitalex (Becton-Dickinson). Bacterial inoculations were prepared by re-suspending bacteria from a lawn grown on chocolate agar in sterile PBS at an OD 600 = 1 (equivalent to 1 × 10 10 CFU/mL). The number of viable bacteria was determined by serial dilution and plating on chocolate agar.

MICE
C57Bl/6J (B6) and B6.SJL-Ptprc a Pepc b /BoyJ (B6-CD45.1), and B6.129-Tlr2 tm1Kir /J (TLR2 KO) mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA). TLR2 KO mice were bred in-house and were age-matched with vendor-purchased B6 mice. Female B6 and TLR2 KO mice were between 6 and 10 weeks old at the time of inoculation. All mice were housed in specific pathogenfree conditions at the University of Arizona in accordance with the Institutional Animal Care and Use Committee (IACUC).

INOCULATION OF MICE
Mice were anesthetized with 575 mg/kg tribromomethanol (Sigma) and intranasally inoculated with 5 × 10 4 CFU LVS clpB. For lethal LVS challenge experiments, mice were anesthetized with 0.25 mL of 7.5 mg/mL ketamine and 0.5 mg/mL xylazine cocktail in PBS and intranasally inoculated with 5 × 10 3 CFU (5 × LD 50 ) LVS 35 days after the initial sub-lethal infection. Mice were weighed daily and sacrificed if they lost more than 25% of their starting weight.

BACTERIAL BURDEN DETERMINATION
Spleen, liver, and lung tissue were homogenized in sterile PBS using a Biojector (Bioject). Ten-fold serial dilutions of tissue homogenates were made using PBS and plated on chocolate agar. Resulting colonies were counted 72 h later. The limit of detection is 50 colony forming units (CFU) per organ.

SPLEEN, LUNG, AND BRONCHOALVEOLAR LAVAGE CELL ISOLATION
Spleens and lungs were harvested from mice and processed into single-cell suspensions as previously described (15). BALF was collected as previously described (15). Cells were removed from the BALF using centrifugation and resulting supernatant was stored at −80°C for multiplex cytokine/chemokine profiling.

CYTOKINE/CHEMOKINE QUANTIFICATION
A multiplex luminex bead-based approach was used to quantify cytokines/chemokines in the BALF as described (15). A 20-analyte assay panel was performed according to the manufacturer's protocol (Life Technologies) using a BioPlex array reader (Bio-Rad Laboratories) in the Duke Regional Biocontainment Laboratory Immunology Unit (Durhan, NC, USA).

INTRACELLULAR CYTOKINE STAINING
Intracellular cytokine staining was performed as previously described (15). Briefly, B6-CD45.1 splenocytes were inoculated with LVS at an MOI of 200:1. Two hours post-inoculation, cells were washed and fresh medium containing 5 µg/mL gentamicin (Sigma) was added. Infected splenocytes were incubated overnight in the presence of gentamicin. Infected splenocytes were washed extensively and then cultured at a 1:1 ratio with cells isolated from the spleen and lung of infected mice for 24 h. A total of 10 µg/mL Brefeldin A (Sigma) was added during the last 4 h of culture to stop cytokine secretion. Flow cytometry data were analyzed as previously described using FlowJo v10.0.6 (Treestar) (15).

STATISTICAL ANALYSIS
A one-way ANOVA with Tukey's post-test was used for BALF cytokine and chemokine concentrations. A Mann-Whitney test was used for ICS data to compare B6 and TLR2 KO mice on days 0, 7, or 10 post-inoculation. Bacterial burdens were log transformed and then a Student's t -test was applied. GraphPad Prism (v5.04) was used for analysis. Significance levels are indicated in the figures as follows: *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

TLR2 KO AND B6 MICE HAVE DIFFERENT DISEASE COURSES WHEN INOCULATED WITH LVS clpB
B6 mice clear LVS clpB infection by day 10 post-inoculation (15). We first sought to determine whether disease course and bacterial clearance were altered in TLR2 KO mice inoculated with LVS clpB. B6 or TLR2 KO mice were intranasally inoculated with 5 × 10 4 CFU LVS clpB and then sacrificed on days 3, 7, 10, or 14 post-inoculation to determine bacterial burdens. Weight loss profiles in B6 and TLR2 KO mice inoculated with LVS clpB differed in peak weight loss (−12% in B6 and −7% in TLR2 KO) and rate of weight gain after day 5 post-inoculation ( Figure 1A). Although peak bacteremia was the same in B6 and TLR2 KO mice (day 3 post-inoculation), LVS clpB clearance was delayed in the spleen, liver, and lung of TLR2 KO mice (Figures 1B-D). All B6 mice

LVS clpB VACCINATION OF TLR2 KO MICE PROTECTS AGAINST LETHAL LVS CHALLENGE
Prior infection (i.e., vaccination) with LVS clpB protects B6 mice during a lethal LVS challenge (15,20). We therefore sought to determine whether TLR2 was required for protection during a secondary challenge. B6 and TLR2 KO mice were challenged with a lethal dose of LVS 35 days after inoculation with LVS clpB. 100% of the B6 and TLR2 KO mice previously inoculated with LVS clpB survived the LVS lethal dose challenge, whereas all naïve mice succumbed to infection (Figure 2A). Vaccinated B6 and TLR2 KO mice also had similar weight loss profiles during lethal challenge ( Figure 2B). Peak weight loss in both groups occurred on day 3 post-rechallenge and was approximately −8% of the starting weight in both groups ( Figure 2B). The weight loss curve for naïve TLR2 KO mice has increased variability because not all mice lost >25% of their starting weight on the same day post-inoculation ( Figure 2B). When vaccinated B6 and TLR2 KO mice were sacrificed on day 14 post-rechallenge, no culturable bacteria were present in the spleen, liver, or lung (data not shown), indicating that vaccination with LVS clpB provided sterilizing immunity during LVS rechallenge and that TLR2 signaling was not required during this memory response.

THE T-CELL RESPONSE IN THE LUNG IS SIMILAR IN B6 AND TLR2 KO MICE INOCULATED WITH LVS clpB
The ability of TLR2 KO mice to survive a lethal LVS challenge suggested that TLR2-deficient mice are able to mount an effective T-cell response to LVS clpB. To determine whether the absence of TLR2 signaling affected the kinetics or magnitude of the T-cell response, we used ICS to enumerate three T-cells subsets (IFN-γ + CD4 + (Th1), IL-17A + CD4 + (Th17), and IFN-γ + CD8 + cytotoxic T-cell) on days 7 and 10 post-inoculation. TLR2 KO mice had a significant increase in lung cellularity compared to B6 mice on day 10, but not day 7, post-inoculation ( Figure 3A). The difference observed on day 10 is likely due to the presence of bacteria in TLR2 KO, but not B6 mice. There was no difference in the absolute number of Th1 cells or percentage of IFN-γ +/ CD4 + Tcells in the lungs of LVS clpB inoculated B6 or TLR2 KO mice (Figures 3B,C). There was also no difference in the absolute number of Th17 cells in the lung of B6 and TLR2 KO mice or percentage of IL-17A + /CD4 + T-cells (Figures 3D,E). Finally, there was no difference in the absolute number of IFN-γ + CD8 + T-cells or percentage of IFN-γ + /CD8 + T-cells in B6 or TLR2 KO mice inoculated with LVS clpB (Figures 3F,G). Together, these data indicate that the kinetics and magnitude of the T-cell response in the lung during LVS clpB infection is similar in B6 and TLR2 KO mice suggesting that TLR2 signaling is not required to mount an adaptive immune response to LVS clpB.

THE T-CELL RESPONSE IN THE SPLEEN IS SIMILAR IN B6 AND TLR2 KO MICE INOCULATED WITH LVS clpB
In addition, we used ICS to identify Th1, Th17, and IFN-γ + CD8 + T-cells in the spleen on days 7 and 10 post-inoculation. There was no difference in total spleen cellularity in B6 and TLR2 KO mice inoculated with LVS clpB (Figure 4A). There were also no significant differences in the absolute number or percentage of cytokine positive Th1, Th17, or IFN-γ + CD8 + T-cell subsets in LVS clpB inoculated B6 or TLR2 KO mice (Figures 4B-G). These www.frontiersin.org data indicate that like the lung, the kinetics and magnitude T-cell response in the spleen is similar in B6 and TLR2 KO mice, suggesting that TLR2 signaling is dispensable for the adaptive immune response to LVS clpB.

CYTOKINE AND CHEMOKINE PRODUCTION FOLLOWING LVS clpB INOCULATION HAVE DIFFERENTIAL REQUIREMENTS FOR TLR2 SIGNALING
Although TLR2 appears to be dispensable for the adaptive immune response to LVS clpB, we sought to determine whether the innate immune response required TLR2 signaling for maximal cytokine and chemokine production. If so, we can use LVS clpB to identify host signaling pathways that are altered during LVS infection, and interrogate the requirement of signaling moieties for the production of specific cytokines and chemokines. Three days postinoculation with LVS clpB, mice were sacrificed and the BALF was collected. The concentration of 20 different cytokines and chemokines in the BALF was determined using a multiplex bead assay and data reported as fold-change ( Figure 5). The absolute concentrations of the cytokines and chemokines are listed in Table  S1 in Supplementary Material.
We identified three classes of clusters and chemokines: those that were partially dependent on TLR2, those that were dependent on TLR2, and those that were independent of TLR2. Cytokines and chemokines that partially depended on TLR2 signaling for their production are IL-6, IL-12 (p40/p70), KC, MIG, and IP-10. Cytokines and chemokines that depended on TLR2 signaling for their production (i.e., are not made at increased levels in infected TLR2 KO compared to uninfected mice) were IL-1α, IL-1β, IL-2, IL-17, MIP-1α, and TNF-α. GM-CSF, IFN-γ, and VEGF were made at similar levels in B6 and TLR2 KO mice indicating that their production was independent of TLR2 signaling. These data indicated that while TLR2 signaling is responsible for the induction of some cytokines and chemokines, other innate signaling molecules may also contribute to the overall innate response to infection.

TLR2 IS REQUIRED FOR MAXIMAL CELLULAR INFILTRATION INTO BRONCHOALVEOLAR LAVAGE FLUID
We speculated that the differences observed in BALF cytokine/chemokine milieu between LVS clpB inoculated B6 and TLR2 KO mice could impact airspace infiltration by innate immune cells. We therefore used flow cytometry to identify immune cell subsets within the BALF on day 3 post-inoculation. TLR2 KO mice have decreased BALF cellularity compared to B6 mice ( Figure 6A). When the cellular composition of the BALF was compared, TLR2 KO mice had fewer neutrophils as a percentage of live cells compared to B6 mice ( Figure 6B). TLR2 KO mice had an increased percentage of DCs compared to B6 mice ( Figure 6C). There was no difference in the percentage of alveolar macrophages (AMs) or interstitial macrophages (IMs) when LVS clpB inoculated B6 and TLR2 KO mice were compared (Figures 6D,E). There was, however, a significant decrease in the frequency of AMs in the BALF of infected animals compared to uninfected control mice ( Figure 6D). The frequency of AMs changed in B6 and TLR2 KO mice inoculated with LVS clpB because there was an influx of infiltrating neutrophils. When the total number of AMs was compared in uninfected or B6 and TLR2 KO mice inoculated with LVS clpB, we did not observe any significant differences between groups (data not shown). Together, these data indicated that differences in the BALF cytokine/chemokine milieu in TLR2 KO mice correlate with changes in BALF cellular composition. Furthermore, these data in conjunction with BALF cytokine and chemokine profiles suggested that TLR2 signaling is required during the innate immune response to LVS clpB.

DISCUSSION
Pattern-recognition receptors, such as TLRs, play a critical role in initiating an innate immune response to microbial pathogens. TLRs except TLR3 and TLR4 require the adaptor protein MyD88 for signaling (5). MyD88-deficient mice are highly susceptible to Francisella infection indicating PRRs are critical to the host's immune response during infection (21). The role of several TLRs has been studied in the context of a Francisella infection. Although Francisella is a gram-negative pathogen, it has an altered lipid A structure that fails to induce signaling through TLR4 (22)(23)(24)(25). TLR4 knock-out (TLR4 KO) mice are not more susceptible than wild-type mice to Francisella infection (7,26,27). The importance of TLR5 and TLR9 has also been tested in the context of Francisella infection, but no phenotype was observed for either PRR (9,21). To date, TLR2 is the best-characterized PRR for F. tularensis (6)(7)(8)(9)(10)(11)(12)(13). Pro-inflammatory cytokine production requires TLR2 signaling and TLR2 KO mice are more susceptible to sub-lethal infection with LVS (6)(7)(8)(9)(10). In order to evade the TLR2-mediated host immune response, Francisella actively inhibits the early innate immune response in vivo (14,15). Lipids derived from SchuS4 inhibit E. coli LPS-induced TNF-α and IL-6 production in the lungs of B6 mice, but not TLR2 KO mice indicating that the lipids depend on TLR2 signaling to inhibit the pro-inflammatory response (12). Not only does Francisella directly inhibit host signaling via TLR2, it also uses the CRISPR/Cas system to regulate expression of its own bacterial lipoprotein (FTN_1103) that could be sensed by host TLR2 (28).
We have previously shown that LVS clpB fails to inhibit the early innate immune response and unlike inoculation with wildtype LVS, a robust pro-inflammatory innate immune response is detected in the BALF on day 3 post-inoculation (15). Despite LVS clpB's attenuation, it elicits a robust T-cell response and previous infection with LVS clpB protects 100% of mice challenged with a lethal dose of LVS (15,20). We were therefore interested in whether TLR2, a key host sensor for detecting Francisella, was required during the various phases of the immune response to LVS clpB. www.frontiersin.org Toll-like receptor 2 KO mice exhibited delayed clearance of LVS clpB (Figure 1). B6 and TLR2 KO mice had similar peak lung bacterial burdens on day 3 post-inoculation indicating that the delayed clearance was not simply due to an initial increase in bacterial burdens that persists during the course of infection. One possible explanation for the delayed clearance is a delay in the T-cell response in TLR2 KO mice. TLR2 has been shown to be required for CD80, CD86, and MHCII up-regulation in bone marrow-derived DCs inoculated with LVS (8). We did not observe any defects in the T-cell response in TLR2 KO mice on days 7 or 10 post-inoculation, suggesting that a poor T-cell response was not the cause of the delayed bacterial clearance in TLR2 KO mice. Another possible explanation for delayed clearance in TLR2 KO mice is the requirement of both IFN-γ and TNF-α during Francisella infection (29)(30)(31). Production of IFN-γ during the innate immune response against LVS clpB was completely independent of TLR2 signaling (Figure 5). Likewise, TLR2 KO mice intranasally inoculated with wild-type LVS produce significantly more IFN-γ on day 7 post-inoculation compared to B6 mice indicating that TLR2 is not required for IFN-γ production (6). TNF-α Frontiers in Immunology | Microbial Immunology FIGURE 5 | TLR2 signaling is required for maximal cytokine and chemokine production in the lung after LVS clpB inoculation. B6 or TLR2 KO mice were intranasally inoculated with 5 × 10 4 CFU LVS clpB or were left naive. Three days post-inoculation, mice were sacrificed and BALF was collected, and cytokine and chemokine concentrations were determined using a Luminex-based assay. For each pair of groups, a fold-changed was determined based on the average cytokine or chemokine concentrations. Data are combined from two independent experiments. n = 7-12 mice/group. Statistical significance was determined using ANOVA with Tukey's post-test on the absolute concentration of each analyte. production after LVS clpB inoculation, however, required TLR2 signaling (Figure 5). TNF-α production in the lungs of TLR2 KO mice inoculated with LVS is delayed and the overall concentration of TNF-α is lower when measured in lung homogenate or by in situ TNF-α staining (6,7). Together, these data indicate that while the IFN-γ-mediated immune response is intact in TLR2 KO mice, there could be defects in the TNF-α-mediated response, which results in delayed LVS clpB clearance.
The ability of TLR2 KO mice to survive a lethal LVS secondary challenge suggested that these mice mount a robust adaptive immune response since T-cells are required for survival during a secondary infection (32). Indeed, B6 and TLR2 KO mice had similar absolute numbers and frequencies of Th1, Th17, and CD8 + IFN-γ + T-cells in the lung and spleen (Figures 3 and 4) (34)(35)(36)(37). In our model, TLR2 KO mice produce significantly less IL-12, a cytokine required for the polarization of Th1 cells, compared to B6 mice (38). Despite this defect or other defects in TLR2-deficient antigen presenting cells that we did not investigate, the T-cell response in TLR2 KO mice is very similar to the response in B6 mice were TLR2 signaling is intact indicating that TLR2 is dispensable for the adaptive immune response to LVS clpB.
We next investigated the requirement of TLR2 during the innate immune response to LVS clpB. Because LVS clpB fails to inhibit the early innate immune response (15), we could use LVS clpB as a tool to identify host signaling pathways that are inhibited during wildtype infection. Intranasal inoculation of B6 and TLR2 KO mice with LVS clpB followed by collection of the BALF 3 days postinoculation, revealed three groups of cytokine and chemokine production: dependent on TLR2, independent of TLR2, and partially dependent on TLR2 ( Figure 5). IL-1α, IL-1β, IL-2, IL-17, MIP-1α, and TNF-α production required TLR2 signaling. Notably, the failure of TLR2 KO mice to produce IL-1β suggests that TLR2 signaling provides the first signal that leads to up-regulation of pro-IL-1β mRNA, which is later cleaved by active caspase-1. The requirement of TLR2 signaling for mouse IL-1β production was also demonstrated by Li et al. (9). GM-CSF, IFN-γ, and VEGF production was independent of TLR2 signaling. IL-6, IL-12p40/p70, KC, MIG, and IP-10 were partially dependent on TLR2 signaling. The decreased frequency of neutrophils in the BALF of TLR2 KO mice (Figure 6B) is likely a consequence of less KC (CXCL1) as KC is a chemoattractant for neutrophils (39). TLR2 KO mice produced cytokines that have been shown to be important during LVS infection such as IFN-γ, IL-6, and IL-12 (29-31, 40, 41). The ability of TLR2 KO mice to produce these cytokines, even if at reduced levels compared to B6 mice, indicates that infection with wild-type LVS inhibits other immune signaling pathways in addition to TLR2. The identities of these host sensor(s) are currently unknown.
Overall, we have demonstrated a differential requirement for TLR2 signaling during the innate and adaptive immune response to LVS clpB. The T-cell response was similar in B6 and TLR2 KO mice indicating that the adaptive immune response during LVS clpB infection does not require TLR2 signaling. TLR2 KO mice also survived a secondary lethal challenge with LVS when first infected with LVS clpB indicating that TLR2 is dispensable during the secondary response. Importantly, some cytokines and chemokines were produced in TLR2 KO mice indicating that other signaling pathways are also inhibited during wild-type Francisella infection in addition to TLR2. The identities of these pathways are currently unknown, but are a focus of our ongoing research. Together, we have demonstrated that TLR2 is critical during the innate immune response to LVS clpB but is not required during the primary or secondary adaptive immune response.

ACKNOWLEDGMENTS
Analysis of 20 cytokines/chemokines in tissue homogenates was performed by Kristina Riebe at the Duke Human Vaccine Institute/Regional Biocontainment Laboratory (RBL) Immunology Unit (Durham, NC, USA). We also thank Paula Campbell of the University of Arizona Flow Cytometry Core Facility. This work was supported by National Institutes of Health grant R01 AI078345 and National Institute of Allergy and Infectious Diseases Southeast Regional Center for Excellence for Emerging Infections and Biodefense grant U54 AI057157. Select studies were performed at Duke RBL, which received partial support for construction from National Institute of Allergy and Infectious Diseases, National Institutes of Health grant UC6 AI058607. The contents of this work are solely responsibility of the authors and do not necessarily represent the official views of NIH.