MyD88-Dependent IFN-γ Production by NK Cells Is Key for Control of Legionella pneumophila Infection

Mice, bacteria, and immunizations

A/J, BALB/c, 129/Sv, C57BL/6 (B6, WT-CD45.2), B6.SJL (WT-CD45.1), C3H/HeN (C3H), or C3H/HeJ (tlr4^0/0) mice were purchased from Janvier Elevage or Harlan. All mice, including 129A (ifn-i r^−/−-129/Sv) (29), 129G (ifn-ii r^−/−-129/Sv) (30), myd88^−/− (myd88^−/−-CD45.2) (31), tlr2^−/−-B6 (32), tlr9^−/−-B6 (33), tlr2/4^−/−-B6/129/Sv (mixed background), IL-6^−/−-B6 (34), il-8r^−/−-B6 (35), tnfr1^−/−-B6 (36), and il-1r^−/− (37), were used at 7–16 wk of age (sex, age, and background matched within experiments; backcrossed over at least nine generations to the indicated strain). All animal experiments were performed in accordance with institutional policies and have been reviewed by the cantonal veterinary office.

The Lpn strains used in this study were JR32 (wild-type Philadelphia-1) (38), JR32-GFP (constitutively expressing GFP from plasmid pMMB207-Km14-GFPc) (39), GS3011 (icmT deletion mutant lacking a functional Icm/Dot T4SS; ΔT) (5), and Corby (wild-type or flaA deletion mutant) (40). L. pneumophila was grown for 3 days on charcoal yeast extract agar plates. To maintain plasmids, chloramphenicol (5 μg/ml) was added. Thymidine-auxotroph derivatives of Lpn JR32 were generated as described by Mintz et al. (41).

If not stated otherwise, mice were infected by i.v. injection of 5 × 10⁶ CFU Lpn JR32 suspended in 200 μl of prewarmed PBS. For intranasal (i.n.) infection, 5 × 10⁶ CFU Lpn in 20 μl of prewarmed PBS was applied directly to one nostril of anesthetized mice using a Gilson-type pipette. Where indicated, bacteria were boiled for 5 min at 95°C or UV irradiated for 2 min. Complete inactivation was tested by plating UV-inactivated and heat-killed Lpn.

Antibodies

All Abs and streptavidin were purchased from BD Biosciences, with the exception of the polyclonal rabbit anti asialo-GM1 preparation used for in vivo NK cell depletion, which was obtained from Wako Chemicals.

FACS analysis

Single-splenocyte suspensions were prepared by forcing whole organs through a metal mesh using syringe plungers. Alternatively, lungs were digested for 30 min at 37°C with collagenase V and then processed as described above. For intracellular staining of IFN-γ, 1 × 10⁷ nucleated splenocytes were cultured in 24-well plates for 4 h in 1 ml of RPMI 1680 supplemented with 10% heat-inactivated FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM glutamine, and 2 μM monensin A. Cells were harvested and washed in ice-cold FACS buffer (PBS, 2% heat-inactivated FCS, 5 mM EDTA, and 0.02% sodium azide).

Cells were resuspended in FACS buffer and stained on the surface with fluorescence- or biotin-labeled Abs for 20 min on ice, followed by fluorescence-labeled streptavidin if applicable. For intracellular staining of IFN-γ, cells were washed once and fixed/permeabilized for 10 min at room temperature using 500 μl of FIX/perm solution (FACSLyse; BD Biosciences; diluted to 2× concentration in distilled water and 0.05% Tween 20). Cells were washed once and stained with directly conjugated Abs against IFN-γ. Cells were then washed again and resuspended in PBS containing 1% paraformaldehyde. Where applicable, TO-PRO3, a DNA-binding dye for live vs dead discrimination (Molecular Probes), was added to the samples immediately before data acquisition. Data were collected on a FACSCalibur flow cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star).

Cytokine analysis in serum (cytometric bead array)

The concentrations of murine inflammatory cytokines (TNF-α, IL-12p70, IFN-γ, IL-10, and IL-6) in sera from Lpn-infected mice and control animals were determined using the BD Cytometric Bead Array (BD Biosciences) according to the manufacturer’s instructions.

Purification and adoptive transfer of NK cells

NK cells were purified from freshly isolated donor spleens after Liberase CI/DNase I digestion (Roche) for 30 min at 37°C, followed by immunomagnetic sorting using the NK purification kit from Miltenyi Biotec, according to the manufacturer’s instructions. Routinely, 80% purity was achieved, as measured by FACS (DX5⁺ depletion markers^negative; data not shown). Cell preparations were washed and resuspended in ice-cold PBS. Purified NK cells (1 × 10⁶) were adoptively transferred into naive recipient mice by i.v. injection in 200 μl of PBS.

Production of inflammatory cytokines after Lpninfection

Cytokines are absolutely indispensable for most, if not all, aspects of immune responses as well as the development and maintenance of cells and organs of the immune system. Particularly important for the immediate control of bacterial infections are cytokines produced by activated cells of the innate immune system. Among those are TNF-α, IFN-γ, IL-6, IL-12, and IL-10, to name a few. As shown in Fig. 1⇓A, all these cytokines, but IL-10 are efficiently induced upon i.n. infection of A/J mice with Lpn 12–16 h after infection. To assess whether this strong innate inflammatory response was confined to i.n. infection, we also infected A/J mice with the same dose of Lpn i.v. As for i.n. infection, robust levels of TNF-α, IFN-γ, IL-6, and IL-12, but no IL-10, were induced, which were readily measurable in the serum 12 h after infection. To identify and compare the cell populations responsible for in vivo bacterial uptake after i.v. and i.n. Lpn infection, we infected A/J mice via these two routes with GFP-Lpn (wild-type Lpn JR32 and icm/dot-deficient Lpn ΔT). Regardless of the infection route, neutrophil granulocytes, macrophages, and DCs were the prominent populations that phagocytosed Lpn in vivo (Fig. 1⇓B). GFP⁺ phagocytes had indeed taken up Lpn, because they were not stainable with an anti-Lpn polyclonal serum unless they were previously permeabilized (not shown). As expected from in vitro results (3), the uptake of Lpn ΔT was significantly decreased in macrophages (but not in neutrophils and DCs) compared with that of Lpn JR32.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Inflammatory cytokine response to Lpn infection. A, A/J mice (n = 3) were immunized either i.n. or i.v. with Lpn JR32. Twelve to 16 h after infection, sera were taken and assayed for cytokine content as described in Materials and Methods. B, A/J mice (n = 3) were immunized either i.n. (upper panels) or i.v. (lower panels) with 5 × 10⁸ CFU Lpn JR32-GFP (black solid lines) or ΔT-GFP (gray solid lines). Four hours after infection, lungs or spleens, respectively, were removed, and uptake of Lpn by phagocytes was assessed by FACS analysis. GFP histograms of cells gated on Gr-1^highTO-PRO3⁻ (left panels), F4/80⁺TO-PRO3⁻ (middle panels), and CD11c^highTO-PRO3⁻ are shown. Dashed lines represent the corresponding samples from unimmunized control animals. Results are representative of two independent experiments. C, A/J, B6, 129/Sv, or C3H mice (n = 3) were injected i.v. with Lpn JR32, and their sera were assayed for the indicated cytokines 12–16 h after infection. D, A/J mice (n = 3) were injected i.v. with Lpn JR32 (live, boiled, or UV irradiated; as indicated) or ΔT. Twelve to 16 h later, serum cytokine levels were assayed. Similar results were obtained for additional time points between 4 and 40 h after inoculation (data not shown). E, As in C, but A/J mice immunized with wild-type or thymidine auxotroph Lpn JR32. Results are representative of three (A) or two (C and D) independent experiments.

Based on the finding that the cellular components responsible for in vivo bacterial uptake were similar after i.v. and i.n. infection and that the serum cytokine levels in i.v. infected mice were severalfold higher than those in i.n. infected mice, we concluded that i.v. infection represents a more sensitive system for the assessment of early inflammatory innate immune responses. Although there are slight differences between i.n. and i.v. Lpn immunization in terms of kinetics and magnitude of the response, these limitations do not influence the experimental outcome qualitatively. Subsequent analysis was thus performed using the i.v. infection route.

Next, we addressed the question of whether the early Lpn-induced inflammatory response was mouse strain dependent; in particular, whether the response was biased toward the A/J mouse strain, which is permissive for in vitro Lpn replication in macrophages (7, 8). Different mouse strains (C57BL/6 (B6), 129/Sv, BALB/c, and C3H mice), whose macrophages do not efficiently support Lpn replication in vitro, were infected i.v. with Lpn, and 12–16 h later, their serum cytokine levels were determined. As controls, A/J mice were treated identically. As shown in Fig. 1⇑C, TNF-α, IL-12p70, and IFN-γ in the sera of non-A/J mice reached similar, if not higher, levels (TNF-α in B6, and IFN-γ in 129/Sv) compared with those in A/J mice, demonstrating that induction of inflammatory responses was not dependent on the genetic background of the host animal.

Next, we analyzed whether induction of an inflammatory response was dependent on infection with live bacteria. Thus, A/J mice were infected i.v. with live, heat-killed, or UV-inactivated wild-type Lpn, and cytokine production was assessed between 4 and 40 h after infection. We found at all time points that only live Lpn were able to induce early inflammatory cytokine responses (Fig. 1⇑D). We also tested whether a functional Icm/Dot T4SS was required for innate cytokine production. A/J mice were infected i.v. with comparable amounts of wild-type Lpn (JR32) and Icm/Dot-deficient Lpn (ΔT; Fig. 1⇑D). Interestingly, only Icm/Dot-proficient bacteria were able to induce high levels of inflammatory cytokines in vivo (comparable results were obtained after i.n. infection; data not shown). Furthermore, infection with wild-type Lpn (JR32) induced much more pronounced neutrophil recruitment compared with Icm/Dot-deficient Lpn (ΔT), and only wild-type Lpn (JR32) induced activation (CD62L down-regulation) of neutrophil granulocytes (data not shown). Because the major difference between wild-type Lpn (JR32) and Icm/Dot-deficient Lpn (ΔT) is the ability of the former to replicate intracellularly, we determined next whether in vivo replication of Lpn was required at all for induction of inflammatory cytokines. To this end, we infected A/J mice with thymidine auxotroph Lpn, which are Icm/Dot proficient, but depend on an exogenous thymidine source for replication. Twelve to 16 h later, serum was taken and assayed for cytokine content (Fig. 1⇑E). Concentrations of IL-12p70 (left panel), TNF-α (middle panel), and IFN-γ (right panel) in these sera were comparable to those derived from control animals infected with thymidine prototroph Lpn. Thus, although the Icm/Dot T4SS is indispensable, in vivo replication of Lpn is not a prerequisite for induction of efficient early inflammatory responses.

It is worth mentioning that even in A/J mice immunized with prototroph Lpn, the bacterial burden increased only 5-fold within 24 h of infection, and bacterial burdens were therefore only marginally increased compared with those in mice infected with thymidine auxotroph wild-type or ΔT Lpn. These findings corroborate the idea of a limited, if any, in vivo role of Lpn replication in the induction of an inflammatory cytokine response.

Taken together, these results show that experimental infection of various mouse strains via the i.v. route proves useful and reliable to study early Lpn-dependent inflammatory cytokine induction. The independence of cytokine induction on Lpn replication and the fact that comparable cytokine titers accumulate in the various mouse strains tested in this study allowed us to use non-A/J mice for analysis of early inflammatory immune responses. Furthermore, potential differences in the bacterial burden between mice of different genetic backgrounds due to small differences in in vivo replication seem negligible at the inoculation dose and at the early time points used in this study.

Role of IFNs in the control of Lpn infection

After having established the validity of the experimental system using i.v. immunization of A/J mice as well as other mouse strains to measure the induction of an innate immune response to Lpn, we were interested in the roles of individual cytokines in the control of in vivo Lpn infection. We first addressed the role of the IFN response. To this end, IFN-I-R-deficient mice (A129), IFN-II-R-deficient mice (G129), and control 129/Sv mice were infected i.v. with Lpn. Spleen and serum were taken 16 or 36 h after infection and analyzed for bacterial burden and cytokine content, respectively. Although cytokine induction and control of bacterial titers were comparable in A129 and control 129/Sv mice, G129 mice failed to control Lpn replication despite similar levels of cytokine production (Fig. 2⇓A) and developed very high bacterial titers within 16 h, succumbing to the overwhelming infection within 36 h (Fig. 2⇓B). These data demonstrate the vital role of IFN-γ signaling in the early clearance of Lpn infection, whereas the type I IFN system is largely dispensable.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Role of IFNs in the response to Lpn infection. A, Wild-type 129/Sv (WT), ifn-i r^−/−, or ifn-ii r^−/− mice (n = 2, 3, and 3, respectively) were immunized i.v. with Lpn JR32. Twelve to 16 h after infection, sera were assayed for the indicated cytokines. B, The Lpn CFU load of the spleens was determined by plating out homogenized organs of infected 129/Sv (□), ifn-i r^−/− (▦), or ifn-ii r^−/− (▪) mice at the time points indicated. Results are representative of two independent experiments.

We also assessed the in vivo role of additional inflammatory cytokines and chemokines by Lpn infection of tnf-r1^−/−, il-6^−/−, and il-8^−/− mice and the corresponding control mice and found no significant difference in serum cytokine levels or bacterial burdens in the spleens (data not shown).

NK cells are the principal producers of IFN-γ early in Lpn infection

Not all the above-listed cytokines are produced exclusively by cells of the innate immune system. For example, IFN-γ and TNF-α are also produced by T lymphocytes. However, all these cytokines become detectable in the serum of Lpn-infected mice at a time after infection when naive lymphocytes are not yet capable of producing them (Fig. 1⇑A and data not shown), indicating that innate cells or nonhemopoietic cells are the sources of these cytokines. Due to its prominent role in the innate response to Lpn infection, the source of early IFN-γ was of particular interest. Splenocyte suspensions of wild-type mice infected i.v. with Lpn 12 h previously were cultured for 4 h in the presence of monensin (but without additional stimulation) to allow accumulation of cytokines in activated leukocytes. The cells were then stained for the presence of intracellular IFN-γ and for surface markers, which allowed identification of leukocyte subsets. No significant staining for intracellular IFN-γ could be detected in macrophages, DC, neutrophil granulocytes, or γδT cells (data not shown). In contrast, a very large fraction (37%) of NK cells stained positively for this cytokine (Fig. 3⇓A). As a control, the samples stained for NK cells were colabeled with Abs against CD3ε to distinguish between classical NK cells (DX5⁺CD3⁻) and NKT cells (DX5⁺CD3⁺). No staining for IFN-γ was detectable in NKT cells (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

NK cells are the principal source of IFN-γ early after Lpn infection. A, A/J mice (n = 3) were immunized with Lpn JR32 i.v. Twelve to 16 h after infection, IFN-γ production by splenocytes was assessed by FACS analysis as described in Materials and Methods. Dot plots of stainings for IFN-γ and DX5 on splenocytes falling into a leukocyte scatter gate are shown. B and C, Anti-asialo-GM1 Ab (150 μg) or solvent was injected into the peritoneum of A/J mice (n = 3) 1 day before the animals were immunized i.v. with Lpn JR32. Sixteen hours after infection, sera were assayed for cytokine content (B), and the Lpn CFU load of the spleens (C) was determined by plating out homogenized organs. Results are representative of four (A) and two (B and C) independent experiments, respectively. 

Given the crucial role of IFN-γ in the response to Lpn infection and the identification of NK cells as the principal source of IFN-γ early on, we hypothesized that NK cells are important in the control of Lpn infection. To this end, we depleted NK cells in vivo by injection of anti-asialo-GM1 Abs into the peritoneum of A/J mice. One day later, the mice were infected i.v. with Lpn. Sixteen hours after infection, we measured the cytokine content in the serum of NK-depleted and untreated control mice. Although the levels of TNF-α and IL-6 induced by Lpn infection remained almost completely unaffected by the absence of NK cells in the host mice (Fig. 3⇑B and data not shown), IFN-γ was reduced to near-basal levels when NK cells were depleted from mice before Lpn infection. The concomitant severe reduction in serum IL-12p70 was most likely an indirect effect of the absence of IFN-γ on the main IL-12 producers in vivo, the DCs (42). At the same time, we determined the bacterial titers in the spleens of NK cell-depleted mice. Complementing the results from the experiments in ifn-ii r^−/− mice, we found that the bacterial burden in NK cell-depleted mice was massively increased compared with that in untreated mice (Fig. 3⇑C). Similar results were obtained using anti-NK1.1 Ab in Lpn-infected B6 mice (data not shown). In conclusion, these data suggest a crucial role for NK cells in the innate immune response to Lpn infection as the principal source of the indispensable cytokine IFN-γ.

Production of IFN-γ by NK cells upon Lpn infection is dependent on MyD88

Research focusing on the molecules and mechanisms involved in the recognition of pathogens by the mammalian innate immune system is enjoying sustained attention in recent years. Especially, pattern recognition receptors belonging to the family of TLRs have been the focus of much interest, because they have proven to be of major importance in immune responses against many pathogens (for a review, see Ref. 43). Most TLRs signal through the adaptor molecule, MyD88. Accordingly, mice deficient in MyD88 are highly susceptible to infections by several pathogens, especially those of bacterial and parasitic nature (28). Consequently, we tested whether there is a role for MyD88 in the innate response to Lpn infection, in particular in the induction of the early and potent IFN-γ production by NK cells. To test this, myd88^−/− and wild-type control mice were infected i.v. with Lpn, and 16 h later, blood and spleens were taken for analysis. Indeed, the sera from Lpn infected MyD88-deficient mice were characterized by the absence of all tested inflammatory cytokines (Fig. 4⇓A).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Essential role of MyD88 in innate response to Lpn. A, Wild-type B6 (WT) or myd88^−/− mice (n = 3) were immunized i.v. with Lpn JR32. Sixteen hours after infection, cytokine levels in sera were assayed. B, IFN-γ production and CD69 expression by splenocytes were assessed by FACS analysis. Dot plots of stainings for IFN-γ (upper panels), CD69 (lower panels), and DX5 on splenocytes falling into a leukocyte scatter gate are shown. C, The Lpn CFU load of the spleens was determined by plating out homogenized organs 2 days after infection. D, Wild-type B6 (WT) or myd88^−/− mice (n = 3) were immunized i.v. with 2.5 × 10⁷ CFU Lpn JR32-GFP. Eight hours after infection, the uptake of GFP-expressing Lpn by phagocytes was assessed by FACS. Histograms of GFP signal in Gr-1^highTO-PRO3⁻ cells falling into a leukocyte scatter gate are shown. Results are representative of three (A–C), and two (D) independent experiments.

Because IFN-γ detection in serum is a rather insensitive method, this finding did not necessarily indicate that NK cells were completely unable to produce IFN-γ in MyD88-deficient hosts. However, as expected, intracellular IFN-γ was totally absent in NK cells of Lpn-infected myd88^−/− mice (Fig. 4⇑B, upper panel). In fact, NK cells did not show any sign of cellular activation, because neither they nor any other splenic leukocyte up-regulated CD69, a phenotypic marker for early leukocyte activation (Fig. 4⇑B, lower panel). In agreement with our findings in ifn-ii r^−/− and NK cell-depleted mice (Figs. 2⇑ and 3⇑), the bacterial burden in myd88^−/− mice was increased on day 2 after infection (Fig. 4⇑C).

The complete absence of any sign of readily measurable activation of the innate immune system in Lpn-infected myd88^−/− mice was not due to a failure of phagocytes to internalize the bacteria. We observed identical in vivo capacities of phagocytes to take up GFP-labeled Lpn in MyD88-deficient and corresponding wild-type mice (Fig. 4⇑D). We infer from these data that the innate recognition of Lpn and the induction of antibacterial effector mechanisms are entirely MyD88 dependent.

To better characterize the specific requirement for MyD88 in the crucial activation of NK cells, we adoptively transferred NK cells, purified from the spleens of wild-type (CD45.2⁺) or myd88^−/− (CD45.2⁺) mice, into congenic wild-type CD45.1⁺ recipient mice. One day later, these mice were infected i.v. with Lpn. Sixteen hours after infection, IFN-γ production by endogenous (CD45.1⁺) and grafted (CD45.2⁺) NK cells was determined by FACS. The endogenous NK cells served as an internal control in these experiments. As expected, IFN-γ was readily detectable in endogenous NK cells from both experimental groups, because they were MyD88 proficient, as well as in the adoptively transferred cells derived from wild-type animals (Fig. 5⇓A). In contrast, NK cells derived from myd88^−/− mice that were grafted into wild-type mice completely failed to produce intracellular IFN-γ (Fig. 5⇓A, bottom right panel). These results demonstrate a strict requirement of MyD88 expression in NK cells for Lpn-induced IFN-γ production. These findings, however, do not exclude the potential need for a MyD88-dependent activation of accessory cells that might subsequently trigger NK cells. To test this possibility, the inverse experiment was performed. Purified wild-type NK cells (CD45.1⁺) were transferred into either wild-type (CD45.2⁺) or myd88^−/− (CD45.2⁺) host mice, which were infected with Lpn 1 day later. Splenic NK cells were assayed for IFN-γ production 16 h later. The endogenous NK (CD45.2⁺) cells from the wild-type hosts produced IFN-γ, whereas the corresponding endogenous cells in the myd88^−/− hosts did not stain for the cytokine (Fig. 5⇓B, upper and lower panels, respectively). Interestingly, however, the CD45.1⁺ wild-type NK cells produced IFN-γ regardless of whether they were in a wild-type or myd88^−/− environment. Together, these findings suggest that MyD88 expression in NK cells is necessary and sufficient for the induction of IFN-γ production upon in vivo Lpn infection.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Requirement for MyD88 NK cells for Lpn-dependent IFN-γ production. A, NK cells purified from WT-CD45.2 or myd88^−/−CD45.2 spleens were adoptively transferred into congenic naive WT-CD45.1 mice. One day later, the hosts were injected with Lpn JR32 or PBS. Sixteen hours after immunization, IFN-γ production by CD45.2⁺ and CD45.2⁻ NK cells was assessed by FACS analysis. Dot plots of stainings for IFN-γ and CD45.2 on DX5⁺ splenocytes falling into a leukocyte scatter gate are shown. B, NK cells purified from WT-CD45.1 spleens were adoptively transferred into naive WT-CD45.2 or myd88^−/−-CD45.2 mice. Hosts were then immunized and analyzed as described in A. Dot plots of stainings for IFN-γ and CD45.1 on DX5⁺ splenocytes falling into a leukocyte scatter gate are shown. Results are representative of two experiments (A and B).

Redundant role of TLRs in induction of the early IFN-γ response upon Lpn infection

To date, the TLR family in the mouse comprises 11 members, which are differentially expressed in a variety of tissues and cells; for many of them, knockout or naturally occurring mutant mice are available. Based on our finding that the induction of early inflammatory cytokine production upon Lpn infection was entirely MyD88 dependent, it was very likely that one or more TLRs were involved. Candidates include TLR2 (recognizing peptidoglycans) (32), TLR4 (LPS from Gram-negative bacteria) (44), and TLR9 (hypomethylated CpG motifs) (33). Surprisingly, when we tested the ability of tlr2^−/−, tlr4 mutant, and tlr9^−/− mice to 1) induce an inflammatory response to Lpn infection (IFN-γ in NK cells) and 2) control the bacterial load, we found neither a significant decrease in the frequency of IFN-γ⁺ NK cells (Fig. 6⇓A), nor a considerably higher bacterial load in the spleen (Fig. 6⇓B). Given that bacteria express many different TLR ligands, it is likely that there is a certain degree of redundancy in recognition of Lpn by TLRs. We addressed this possibility by infecting tlr2/tlr4 double-knockout mice with Lpn. However, as in the single knockout animals, there was no significant difference in the percentage of NK cells producing IFN-γ, nor did we measure a significantly increased Lpn bacterial load in these mice (Fig. 6⇓B). Another prominent candidate that could mediate Lpn recognition is TLR5 (45). We addressed the relevance of TLR5 (recognizing flagellin) (46) by infection of A/J mice with flagellated (Corby) or aflagellated (Corby flaA⁻) Lpn. Similar to the results obtained with wild-type Lpn in TLR2-, TLR4-, and TLR9-deficient mice, infection of A/J mice with aflagellated Lpn produced an induction of IFN-γ production by NK cells comparable with that after infection with the corresponding flagellin-expressing Lpn 16 h after infection (Fig. 6C). This result indicates no exclusive activation of innate immune responses by Lpn flagella. Finally, we addressed the combined roles of TLR2, TLR4, and TLR5 activation in vivo by infection of tlr2/4 double-knockout mice with flagellated (Corby) and aflagellated (Corby flaA⁻) Lpn. Again, aflagellated Lpn induced only marginally reduced IFN-γ production in NK cells compared with flagellated Lpn, and bacterial titers were also only slightly increased 16 h after infection (Fig. 6⇓D). MyD88 not only serves as an adaptor downstream of TLRs, but is also involved in certain cytokine signaling pathways. We assessed the role of one of these cytokines, IL-1, in Lpn-induced IFN-γ production by NK cells by infecting IL-1R-deficient mice. Similar to the situation in the TLR-deficient mice tested above, IFN-γ secretion by NK cells upon Lpn infection was only slightly impaired in il-1r^−/− mice compared with control mice (Fig. 6⇓E), suggesting that the absolute MyD88 dependence for IFN-γ production by NK cells was not solely due to IL-1R signaling. Taken together, the results suggest a rather redundant pattern of TLR activation in vivo by Lpn infection, such that no individual TLR or even a combination of TLR2, TLR4, and TLR5 could be identified as exclusive TLR targets of Lpn for induction of IFN-γ production by NK cells in vivo. Alternatively, our findings can be interpreted as evidence that TLRs do not play a role in the activation of innate immune responses upon Lpn infection in vivo or, in particular, in NK cell activation. However, despite this apparent redundancy of single or multiple TLRs, induction of innate immune responses was absolutely dependent on MyD88 signaling.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

TLR redundancy in the innate response to Lpn. A, IFN-γ production by splenocytes from wt, tlr2^−/−, tlr4^0/0, tlr9^−/−, or tlr2^−/−× tlr4^−/− mice (n = 2) immunized with Lpn JR32 12–16 h before FACS analysis. Dot plots of stainings for IFN-γ and DX5 on splenocytes falling into a leukocyte scatter gate are shown. B, The Lpn CFU load in spleens from the same experimental mice as those in A was determined by plating out homogenized organs. Only one representative WT control for all strains corresponding to the TLR-deficient animals is shown (A and B). C, A/J mice were immunized with 5 × 10⁶ CFU Lpn Corby or Corby flaA⁻, and IFN-γ production in NK cells was assessed by FACS analysis 16 h later. Dot plots of stainings for IFN-γ and DX5 on splenocytes falling into a leukocyte scatter gate are shown. D, tlr2^−/−× tlr4^−/− mice (n = 3) were immunized with 5 × 10⁶ CFU Lpn Corby or Corby flaA⁻. Sixteen hours after infection, IFN-γ production by splenocytes was assessed by FACS. Dot plots of stainings for IFN-γ and DX5 on splenocytes falling into a leukocyte scatter gate are shown (left and middle panels). From the same experimental mice, the spleen Lpn CFU load was determined (right panel). Results are representative of two independent experiments. E, Wild-type or il-1r^−/− (n = 2) mice were immunized and analyzed as described in A.