Critical coordination of innate immune defense against Toxoplasma gondii by dendritic cells responding via their Toll-like receptors

Abstract

Toll-like receptors (TLRs) play an important role in host defense against a variety of microbial pathogens. We addressed the mechanism by which TLRs contribute to host defense against the lethal parasite Toxoplasma gondii by using mice with targeted inactivation of the TLR adaptor protein myeloid differentiation primary response gene 88 (MyD88) in different innate cell types. Lack of MyD88 in dendritic cells (DCs), but not in macrophages or neutrophils, resulted in high susceptibility to the T. gondii infection. In the mice deficient in MyD88 in DCs, the early IL-12 response by DCs was ablated, the IFN-γ response by natural killer cells was delayed, and the recruited inflammatory monocytes were incapable of killing the T. gondii parasites. The T-cell response, although attenuated in these mice, was sufficient to eradicate the parasite during the chronic stage, provided that defects in DC activation were compensated by IL-12 treatment early after infection. These results demonstrate a central role of DCs in orchestrating the innate immune response to an intracellular pathogen and establish that defects in pathogen recognition by DCs can predetermine sensitivity to infection.

The protozoan parasite Toxoplasma gondii is a highly successful intracellular pathogen that has evolved sophisticated mechanisms for its evasion of immune killing (1). Worldwide, it is estimated that 30–70% of the human population is infected with this parasite, and essentially the entire human population is at risk for infection (2). Type I cell-mediated immunity plays a key role in the host resistance to T. gondii infection. Most attention has been focused on the essential role of IFN-γ–producing T helper type 1 (Th1) effector T cells for the resolution of acute infections and for control of latent, chronic infections (3). Recently, evidence has begun to accumulate also implicating innate immune cell types in early defense against T. gondii infection. For example, depletion of natural killer (NK) cells or dendritic cells (DCs) rendered mice susceptible to T. gondii infection (4, 5). In both cases, the cause still may be related to defective activation of CD4⁺ T cells. Interestingly, another study has found that failure to recruit Gr-1⁺ inflammatory monocytes to T. gondii-infected sites also resulted in increased susceptibility despite apparently normal development of the Th1 response (6), suggesting that the innate arm of host immunity also plays a unique role in addition to activating the T-cell response. It is unclear how these different immune cell types interact to promote protection against T. gondii.

Toll-like receptors (TLRs) are critical innate receptors for recognition of ligands expressed by T. gondii, including profilin, which in the mouse stimulates TLR11 (7) (however, the human Tlr11 gene appears to contain a stop codon that would prevent expression of the protein T. gondii. Glycans and diacylglycerols also can stimulate TLR2 and TLR4 (8, 9). Although mice deficient in individual TLRs show only a partial defect in controlling T. gondii infection (7, 9–11), mice deficient for MyD88, a key adaptor for signaling by almost all TLRs, are highly susceptible to acute T. gondii infection. Mice deficient in MyD88 exhibit a defective Th1 response (12–14), but, because TLRs are expressed by many immune cell types and perform many functions (15), it is not clear how TLR signaling regulates different immune cells and contributes to host defense against this pathogen.

To investigate these questions, we took advantage of a recent mouse model in which specific ablation of the mouse MyD88 gene in certain cell types can be engineered using Cre-lox technology (16). We report here that MyD88 signaling in DCs, but not in macrophages or neutrophils, is required for innate immune control of parasite replication during acute T. gondii infection of the peritoneum before the involvement of Th1 effectors in the defense. Mice deficient in MyD88 in DCs and infected with T. gondii exhibited a nearly complete defect in IL-12 production, a delay in IFN-γ production by NK cells, and a failure to activate killing mechanisms of inflammatory monocytes, resulting in uncontrolled parasite growth. Although the Th1 response to T. gondii was slightly attenuated in the mice lacking MyD88 selectively in DCs, it was sufficient to maintain protection if the innate defect was corrected by injection of IL-12. Therefore, our results establish a critical function of MyD88 in DCs in activating type 1 innate immunity against this intracellular parasite and also provide evidence that a defect in early innate immunity may predetermine the outcome of the infection even in the presence of a largely normal adaptive immune response.

Results

DC-Myd88^−/− Mice Are Highly Susceptible to T. gondii Infection.

To investigate cell-specific functions of MyD88 during experimental toxoplasmosis, we used the well-established model of i.p. infection with T. gondii (17, 18). Because tissue macrophages are a major innate immune cell type in the peritoneum, and DCs and neutrophils previously were found to play a crucial role in host resistance to T. gondii, we focused on the role of MyD88 signaling in these cell types in the defense against T. gondii infection. Mice lacking MyD88 selectively in DCs (DC-Myd88^−/− mice) were generated by crossing the CD11c-Cre transgene into mice with two loxP sites flanking the essential exon 3 of the Myd88 gene (Myd88^fl allele), as has been reported previously (16). Furthermore, using a quantitative PCR assay as previously reported (16), we did not detect substantial deletion of Myd88^fl allele in the resident peritoneal macrophage of these mice (Table S1). To generate mice lacking MyD88 in macrophages, mice with the Myd88^fl allele were crossed to animals carrying the LysM-Cre transgene, which specifically deletes the targeted gene in tissue, macrophages, and neutrophils (19). In these mice, deletion of the Myd88^fl allele occurred in more than 90% of resident peritoneal macrophages (henceforth referred to as “MN-Myd88^−/−”) (Table S1).

The survival of these conditional MyD88-deficient mice and their littermate control mice was followed after i.p. injection of different numbers of Me49 T. gondii brain cysts. Surprisingly, there was no apparent survival deficit in T. gondii infection of the mice lacking MyD88 in peritoneal macrophages and neutrophils (Fig. 1A). In contrast, the DC-Myd88^−/− mice were highly susceptible to T. gondii infection. About 40% of these mice died within 2–3 wk when infected with 40 brain cysts, and increasing the infecting dose of T. gondii organisms to 200 cysts resulted in almost 100% mortality of the infected mice during the acute phase of the infection (Fig. 1 A and B). These results revealed that MyD88 signaling in DCs plays a major role in protective immunity during acute T. gondii infection. Although all Myd88^−/− mice died during the acute stage, some DC-Myd88^−/− mice survived the lower dose of acute infection but died between 4 and 6 wk after infection (Fig. 1A). The difference observed between Myd88^−/− and DC-Myd88^−/− mice indicates a role for MyD88 signaling in DCs as well as in other cell types (e.g., T cells) (20).

Fig. 1.

Role of MyD88 signaling in DCs for survival of mice after T. gondii infection. (A and B) Female mice 7–9 wk old were infected i.p. with 40 (A) or 200 (B) cysts of the avirulent Me49 strain, and their survival was monitored daily up to 8 wk after infection. (A) Results of WT Myd88^fl/− mice (n = 17), DC-Myd88^−/− mice (n = 14), MN-Myd88^−/− mice (n = 8), and Myd88^−/− mice (n = 6) from two separate experiments were pooled. (B) Results of WT mice (n = 9), DC-Myd88^−/− mice (n = 9), and Myd88^−/− mice (n = 4) from two separate experiments were pooled. Significant differences were found between WT mice and the DC-Myd88^−/− mice in both low and high infecting doses (log-rank test; P < 0.001), and between DC-Myd88^−/− mice and Myd88^−/− mice at low infection dose (P < 0.05). (C) The amount of T. gondii genomic DNA was measured by quantitative PCR after infection with 40 cysts of the Me49 strain. *P < 0.05; **P < 0.01. (D) The amount of T. gondii in the lavaged peritoneal cells of Tcrα^−/− mice, Rag1^−/− mice, and WT mice at day 5 after infection. Data are expressed as mean plus SD of three or four mice and are representative of two to four separate experiments. Statistical comparison is between the DC-Myd88^−/− mice and the WT control mice.

Early Defect in Controlling T. gondii Expansion in DC-Myd88^−/− Mice.

To investigate the cause of death of DC-Myd88^−/− mice during the acute phase of infection, we quantified T. gondii organisms after infection by measuring the amount of T. gondii genomic DNA in the lavaged peritoneal cells using quantitative PCR. We found that in control mice the number of tachyzoites increased slightly from day 1 to day 5 and then declined. In contrast, T. gondii tachyzoites expanded rapidly in the DC-Myd88^−/− mice and Myd88^−/− mice, with significantly more tachyzoites present in the peritoneum of the DC-Myd88^−/− mice as early as day 3 after infection, and with >100-fold more tachyzoites present in these mice on day 8 (Fig. 1C). The DC-Myd88^−/− mice contained only slightly fewer tachyzoites than the Myd88^−/− mice, demonstrating a major role for DCs in the MyD88-mediated innate immune response to T. gondii infection (Fig. 1C). Substantially more tachyzoites also were detected in other organs of the DC-Myd88^−/− mice (Fig. S1A), indicating an accelerated dissemination from the original locus of infection that may have caused death during the acute phase of the infection. Strikingly, this early defect in controlling T. gondii expansion was not seen in Tcrα^−/− mice lacking αβ T cells or in Rag1^−/− mice lacking both T and B cells (Fig. 1D), indicating that insufficient activation of the adaptive immune system was not the cause of the parasite expansion and dissemination seen in DC-Myd88^−/− mice. Consistent with the survival data (Fig. 1), MN-Myd88^−/− mice were able to control parasite expansion at a level comparable to WT mice (Fig. S1B). These results revealed that a deficiency of MyD88 in macrophages and neutrophils had little effect on innate defense against T. gondii.

Recruitment of Inflammatory Cells in Acute T. gondii Infection.

To understand defects in the immune response to T. gondii infection caused by loss of MyD88 in DCs, we first characterized the inflammatory cells that were recruited rapidly into the peritoneal cavity after infection with the parasite. We observed that among the recruited cells, the Ly6C⁺ inflammatory monocytes were the major cell type (Fig. 2 A–D). Compared with WT mice, infected DC-Myd88^−/− mice often exhibited a small deficit in recruitment of inflammatory monocytes early after infection, but this deficit became less obvious as early as day 3 after infection (Fig. 2A and Fig. S2). Conversely, DC-Myd88^−/− mice generally recruited more eosinophils, as did the Myd88^−/− mice (Fig. S2C). At later time points, the infected DC-Myd88^−/− mice recruited considerably more cells than WT mice, probably in response to increased numbers of T. gondii. In contrast, substantially fewer monocytes and neutrophils were recruited into the peritoneum of Myd88^−/− mice than DC-Myd88^−/− mice (Fig. 2C and Fig. S2B), although there were more tachyzoites in Myd88^−/− mice (Fig. 1C). These results suggest that MyD88 signaling in DCs is not essential for the recruitment of these inflammatory cell types.

Fig. 2.

Role of MyD88 in DCs for the recruitment of inflammatory cells in acute T. gondii infection. (A–D) Peritoneal cells were enumerated and analyzed by flow cytometry after infection with 40 cysts of the Me49 strain. The numbers of (A) total infiltrating cells (peritoneal exudate cells, PECs), (B) DCs (CD11c⁺, MHC II⁺), (C) inflammatory monocytes (Mo) (F4/80^int, Ly6C⁺, Ly6G⁻), and (D) NK cells (NK1.1⁺, TCRβ⁻, MHCII⁻, F4/80⁻) are shown. Data are expressed as mean plus SD of three or four mice and are representative of two or three separate experiments. (E) Numbers of infected peritoneal cells (mean plus SD of four mice) on day 5 after infection with 100 tachyzoites of the GFP-expressing PruΔhtp strain. (F) Median fluorescence intensity (MFI) of surface MHC II on inflammatory monocytes. Statistical comparison is between the DC-Myd88^−/− mice and the WT control mice. **P < 0.001. (G) Intracellular iNOS expression in gated inflammatory monocytes (Ly6G⁻, Ly6C⁺, F4/80⁺) on day 3 after infection as analyzed by flow cytometry.

Defective Activation of Inflammatory Monocytes in DC-Myd88^−/− Mice.

Enhanced T. gondii replication in the presence of a nearly normal attraction of inflammatory cells in the DC-Myd88^−/− mice implied a defect in the killing of intracellular parasites in the absence of MyD88 signaling in DCs. To identify the cell types that were infected by T. gondii, we infected mice with another type II T. gondii strain that expresses GFP. Defects in controlling parasite expansion and in surviving the infection observed with this strain were similar to those seen in the Me49 strain (Fig. S3). The peritoneal cells were lavaged and analyzed by flow cytometry 5 d after infection. It was evident that all immune cell types can be infected by T. gondii; however, inflammatory monocytes constituted the majority of the infected cells (Fig. 2E and Fig. S4).

Flow cytometric analysis revealed reduced surface expression of major MHC II molecules by inflammatory monocytes (Fig. 2F), suggesting a defect of monocyte activation in DC-Myd88^−/− mice. To examine the killing machinery by these cells directly, we measured the induction of inducible nitric-oxide synthase (iNOS), a crucial factor for monocyte killing of intracellular pathogens, by staining intracellular iNOS protein in inflammatory monocytes isolated from the peritoneum on day 3 and day 5 after infection. Consistent with the results of MHC II expression, a severely delayed induction of iNOS was seen in the DC-Myd88^−/− mice (Fig. 2G). By quantitative PCR, we did not detect deletion of Myd88^fl in purified inflammatory monocyte (Table S1), ruling out the possibility that the functional defect was caused by deletion of MyD88 by CD11c-Cre in these cells. Together, these results revealed an important interaction between DCs and inflammatory monocytes; that is, MyD88 in DCs is essential for the activation of but not for the recruitment of inflammatory monocytes in T. gondii-infected mice.

Delayed Induction of IFN-γ in T. gondii-Infected DC-Myd88^−/− Mice.

IFN-γ is a key factor inducing MHC II and iNOS expression in macrophages and monocytes. Moreover, it is known that IFN-γ plays a critical role in stimulating T. gondii killing in both immune and nonimmune cells (21, 22). Indeed, we observed that monocyte activation, as measured by surface expression of MHC II and intracellular iNOS, also was severely defective in IFN-γ^−/− or in IFNGR1^−/− mice infected with T. gondii (Fig. S5). To examine whether lack of MyD88 in DCs affected the production of IFN-γ in response to T. gondii infection, the induction of IFN-γ mRNA in peritoneal cells was measured by quantitative PCR. There was a severe reduction in the production of IFN-γ in the DC-Myd88^−/− mice on day 2 postinfection (Fig. 3A). Substantial induction of IFN-γ was detected in the DC-Myd88^−/− mice beginning on day 3 postinfection (Fig. 3A), but Myd88^−/− mice exhibited an even greater delay in IFN-γ production. A similar deficit in the serum levels of IFN-γ also was found in the DC-Myd88^−/− and Myd88^−/− mice (Fig. S6). These results suggested that a timely induction of IFN-γ is crucial for host defense against T. gondii infection and that TLR/MyD88 signaling in DCs is required for inducing the early IFN-γ expression after T. gondii infection.

Fig. 3.

Induction of IFN-γ in the T. gondii-infected DC-Myd88^−/− mice. (A) Induction of IFN-γ mRNA in total peritoneal cells as measured by quantitative RT-PCR. Data are expressed as mean plus SD of the normalized threshold cycle number (−ΔCt) of three or four mice and are representative of two or three separate experiments. (B–D) Intracellular IFN-γ as analyzed by flow cytometry. (B) Numbers of IFN-γ⁺ NK cells (mean plus SD of three or four mice) in the peritoneum after infection. (C) Numbers of IFN-γ⁺ T cells in the peritoneum after infection. Statistical comparison is between the DC-Myd88^−/− mice and the WT control mice. *P < 0.05; **P < 0.01. (D) Representative FACS plots of intracellular IFN-γ staining on gated NK cells (NK1.1⁺, TCRβ⁻) on day 2 and day 3 after infection.

To identify the cellular source of IFN-γ in the T. gondii-infected mice, the peritoneal cells were stained for intracellular IFN-γ. We found that NK cells dominated the early (days 2–5) IFN-γ response (Figs. 3 B and C), and αβ T cells also contributed to some extent. Remarkably, although a large number of NK cells in WT mice expressed intracellular IFN-γ as early as day 2 after infection, this response was ablated almost completely in the DC-Myd88^−/− mice and became obvious only on day 3 (Fig. 3D), demonstrating a critical crosstalk between DCs and NK cells that accounted for the rapid IFN-γ response in the early defense against T. gondii infection. The dialogue between DC and NK cells also was observed in cell-culture experiments in which the addition of live T. gondii parasites to cultures of NK cells, DCs, and macrophages led to NO production by the macrophages that required DCs, NK cells, and IFN-γ (Fig. S7).

The considerable difference in the levels of IFN-γ produced by NK cells in the DC-Myd88^−/− and Myd88^−/− mice probably is explained by the inability of NK cells in the Myd88^−/− mice to respond to IL-18, which enhances IFN-γ production (23, 24). Indeed, anti–IL-18 treatment of T. gondii-infected DC-Myd88^−/− mice reduced IFN-γ production in these mice to the levels observed in the Myd88^−/− animals (Fig. S8).

Defective IL-12 Induction in T. gondii-Infected DC-Myd88^−/− Mice.

Because mice lacking IL-12, a potent inducer of IFN-γ, are highly susceptible to T. gondii infection (25, 26), and MyD88 signaling in DCs is believed to be required for IL-12 induction by TLR ligands, we explored the effect of deletion of MyD88 in DCs on the expression of IL-12 and its relationship with the induction of IFN-γ during T. gondii infection. Initial analysis revealed substantially reduced the levels of the p40 subunit of IL-12 (IL-12p40) in the peritoneal lavage supernatant taken from the infected DC-Myd88^−/− mice compared with that from infected WT mice (Fig. 4A). Moreover, intracellular staining of IL-12p40 and quantifying the mRNA level of IL12p40 in purified peritoneal cells revealed that most IL-12–producing cells were conventional DCs and not inflammatory monocytes or resident peritoneal macrophages, because the cytokine-positive cells were CD11c⁺ and MHC II⁺ and were CD11b^low–neg, F4/80⁻, and Ly6C⁻ (Fig. 4B and Fig. S9).

Fig. 4.

Induction of IL-12 in the T. gondii-infected DC-Myd88^−/− mice. (A–D) Induction of IL-12 in peritoneal cells was analyzed after infection with 40 cysts of the Me49 strain. (A) Total amount of IL-12p40 (mean plus SD of three or four mice) in the supernatant of peritoneal lavage measured by ELISA. (B) Representative flow cytometry plots of intracellular IL-12p40 staining of gated DCs (CD11c⁺, MHC II⁺) at day 3 after infection. (C) Correlation analysis of the amount of IL-12p40 and the induction of IFN-γ mRNA in the peritoneum at day 2 and day 3 after infection. The amount of IL-12p40 was determined as in A, and the level of IFN-γ mRNA was determined as in Fig. 3A. ◇, WT mice; ◆, DC-Myd88^−/− mice. (D–F) The effects of IL-12 treatment of mice infected with 40 cysts of the Me49 strain. (D) Representative flow cytometry plots of intracellular IFN-γ in gated NK cells 1 d after supplementation with IL-12 or with PBS control on day 1 after infection. (E) Representative flow cytometry plots of intracellular iNOS in gated inflammatory monocytes on day 3 after infection. Mice in this experiment received supplementation with IL-12 or with PBS control on day 1 and 2 after infection. (F) Total amount of T. gondii in the peritoneal cells at day 7 after infection. Four mice of each strain were treated twice with PBS or 150 ng IL-12 daily on day 1 and day 2. Statistical comparison is between the DC-Myd88^−/− mice treated with or without IL-12. **P < 0.01.

When the relationship between the amount of IL-12p40 in the peritoneum and the induction of IFN-γ mRNA in each individual mouse was compared, a strong positive correlation was found between these values on day 2 (Fig. 4C, Upper), supporting the conclusion that IL-12 plays a critical role in the induction of IFN-γ in mice infected with T. gondii. Surprisingly, the close correlation of these cytokines disappeared only 1 d later, when IFN-γ was greatly induced in the DC-Myd88^−/− mice although the amount of IL-12 was severely lower in these mice (Fig. 4C, Lower), suggesting that by day 3 other cytokines produced in infected DC-Myd88^−/− mice were capable of promoting secretion of IFN-γ by NK cells.

To test whether the loss of IL-12 induction and the transient delay in IFN-γ production in the DC-Myd88^−/− mice could account for the severely defective immune control of T. gondii infection, recombinant IL-12 was administrated into the peritoneum of these mice. After a single treatment on day 1 after infection, NK cells in the DC-Myd88^−/− mice expressed substantially higher intracellular IFN-γ on day 2 than did the NK cells in the untreated mice, demonstrating that restoration of IL-12 could correct the defect in the induction of IFN-γ by the NK cells of these mice (Fig. 4D). Moreover, injection of the DC-Myd88^−/− mice with IL-12 on the first 2 d after infection also corrected the defective induction of iNOS in inflammatory monocytes in these mice (Fig. 4E). Strikingly, this short treatment dramatically alleviated the moribund appearance of these mice and resulted in a >10-fold decrease in the number of tachyzoites present in the peritoneum of these mice on day 7 as compared with the vehicle-treated mice (Fig. 4F). These results demonstrate that the early induction of IL-12 by DCs plays a major role in the defense against T. gondii infection.

Partial Defect in Th1 Priming in T. gondii-Infected DC-Myd88^−/− Mice.

Because T-cell–mediated immunity is known to be essential for surviving an acute T. gondii infection, we next examined the activation of effector T cells after infection. Eight days after infection, the numbers of effector T cells producing IFN-γ were reduced in DC-Myd88^−/− mice compared with WT mice, indicating that DC-Myd88^−/− mice also had a less robust T-cell response to T. gondii (Fig. 5 A and B). However, compared with Myd88^−/− mice, considerably greater numbers of IFN-γ⁺ CD4⁺ and CD8⁺ T cells were found in the DC-Myd88^−/− mice, demonstrating that effector T cells can develop despite a deficiency of MyD88 in DCs. The partial activation of T cells may have rescued some of the T. gondii-infected DC-Myd88^−/− mice from acute death, although all these mice died within 5 wk (Fig. 1).

Fig. 5.

Activation of T cells in the DC-Myd88^−/− mice. Mice were infected with 40 cysts (A–C) or 100 cysts (D) of the Me49 strain. (A and B) Splenocytes were analyzed for intracellular IFN-γ and TNF 8 d later after in vitro stimulation with anti-CD3 for 4 h in the presence of brefeldin A. (A) The total number of IFN-γ⁺ CD4⁺ and CD8⁺ T cells in the spleens (mean plus SD of four or five mice). (B) Representative flow cytometry plots of intracellular IFN-γ and TNF staining of gated CD4⁺ and CD8⁺ T cells. (C and D) Mice were treated with either PBS or 150 ng IL-12 daily from day 1–3. T. gondii tissue cysts were enumerated in brain homogenates of surviving mice at day 28 after infection. Statistical comparison is between WT mice and the DC-Myd88^−/− mice or the DC-Myd88^−/− mice treated with or without IL-12. *P < 0.05. **P < 0.01. (D) The survival 7- to 8-wk-old female mice was monitored daily up to 6 wk after infection. Results of PBS-treated WT mice (n = 12), DC-Myd88^−/− mice (n = 10), IL-12-treated WT mice (n = 5), DC-Myd88^−/− mice (n = 5), and Myd88^−/− mice (n = 4) are shown. A significant difference was found between DC-Myd88^−/− mice treated with or without IL-12 (log-rank test; P < 0.01).

Examination of the brains of infected mice 4 wk after infection revealed that the surviving DC-Myd88^−/− mice contained 5- to 10-fold more T. gondii tissue cysts than did the WT mice (Fig. 5C), indicating that they probably experienced more severe toxoplasmic encephalitis. To test whether the failure to control dissemination from the peritoneum and formation of tissue cysts was caused by the defective activation of effector T cells or by increased parasite burden resulting from the defective innate control of parasite replication, we treated the mice with IL-12 for a short time after infection to restrict the expansion of T. gondii tachyzoites. We found that this short-term treatment did not correct the number of IFN-γ–producing T cells in the DC-Myd88^−/− mice (Fig. S10). This result is in agreement with a previous observation indicating that maintenance of Th1 effector function in chronic toxoplasmosis requires persistent IL-12 production (27). To our surprise, the short IL-12 treatment nonetheless substantially reduced the numbers of T. gondii tissue cysts in the brains of DC-Myd88^−/− mice infected with 40 T. gondii brain cysts (Fig. 5C) and even could rescue mice infected with a higher number of T. gondii from acute mortality (Fig. 5D), suggesting that the major defect causing chronic death in T. gondii-infected DC-Myd88^−/− mice lies in the failure to contain the early expansion of tachyzoites via innate immunity.

Discussion

In the work described here, we have examined the role of TLRs on DCs, macrophages, and neutrophils in mediating immunity against T. gondii. These experiments revealed that lack of MyD88 signaling only in DCs largely recapitulated the deficiency in innate immunity against T. gondii infection observed in complete MyD88^−/− mice and also provided considerable insight into the mechanism by which DCs coordinate immune defense against this intracellular pathogen.

Previous studies have established the essential role of T cells in defense against T. gondii infection, so we initially were surprised that DC-Myd88^−/− mice were highly susceptible to acute T. gondii infection even though the Th1 responses in these mice were attenuated only slightly. Our subsequent results revealed that TLR recognition by DCs is critical for mounting a rapid type 1 innate immune response to T. gondii and that this innate response is essential to prevent acute mortality. Additionally, the innate defect of DC-Myd88^−/− mice predetermined the outcome of chronic infection by affecting the parasite burden and dissemination to other sites such as the brain. The early immune defects in infected DC-Myd88^−/− mice included an ablated early IL-12 response by DCs, a delayed IFN-γ response by NK cells, a defective activation of the recruited inflammatory monocytes, and a failure to kill T. gondii parasites. Short-term treatment with IL-12 early after infection largely corrected these innate defects but did not correct the attenuated Th1 response, indicating that the defective innate response in these mice is responsible for their increased susceptibility and that the primary cause of this susceptibility is the defective IL-12 response by MyD88-deficient DCs. Previous studies depleting DCs (5) or NK cells (4) or examining IL-12–deficient (25, 26), IFN-γ–deficient (21), or MyD88-deficient mice (12–14) failed to distinguish between early innate defect and defects in generating Th1 effector cells, and it generally was assumed that susceptibility to infection largely reflects a defect in adaptive immunity. Instead, our results demonstrate that a more crucial function of TLR/MyD88 signaling in DCs in defense against T. gondii infection is to activate NK cells and inflammatory monocytes. The efficiency of DCs in activating those cells probably arises from their ability to organize the early clustering of those innate immune cells around the loci of infection, as recently observed in mice infected with Listeria monocytogenes (28), another intracellular pathogen. Surprisingly, among peritoneal-resident cells, only DCs could respond to parasite TLR ligand and coordinate the early type 1 immune defense against T. gondii.

We previously found DC-intrinsic MyD88 to be essential in activating Th1 effector cells in response to immunization with soluble T. gondii antigen (29). In the case of infection with live T. gondii parasites, however, there evidently were additional innate stimuli that promoted a substantial, if somewhat attenuated, Th1 effector response. This finding agrees with our previous observations that aggregated TLR ligands can activate TLR/MyD88 signaling in other myeloid cell types and that in this context cytokines can promote the function of DCs and Th1 development (16). In the case of T. gondii infection, we demonstrated that T cells activated by this extrinsic pathway were sufficient to mediate a high level of protection against T. gondii provided the innate immune defect was corrected by injection of IL-12.

In summary, our results demonstrate a critical role of TLR/MyD88 signaling in DCs for initiating IL-12– and IFN-γ–dependent innate immune responses to the intracellular pathogen T. gondii. Although deficiency of MyD88 signaling in DCs somewhat compromised the activation of effector T cells, the most important defect for host defense was an impaired and delayed activation of innate effector cells, including NK cells and monocytes. Because type 1 immunity similarly is essential for defense against other intracellular pathogens, examining how MyD88 signaling in DCs and other innate immune cells contribute to their defense would add to our understanding of the mechanisms used by the immune system to defend against highly evolved intracellular pathogens.

Materials and Methods

Full details of methods and associated references are available in SI Materials and Methods.

Mice.

Myd88^fl mice (B6.129P2-Myd88^tm1Defr) and Myd88^fl/fl CD11c-Cre (DC-Myd88^−/−) mice have been described previously (16). All animals were housed in a specific pathogen-free animal facility, were used at 7–10 wk of age, and were sex and age matched.

T. gondii Infection, Quantitative PCR, and Flow Cytometry.

Two protocols were used to study the immune responses to T. gondii infection. In the first protocol, mice were infected i.p. with tissue cysts of the avirulent Me49 strain. In the second protocol, mice were injected with 100 tachyzoites of the GFP- and firefly luciferase (FLUC)-expressing PruΔhpt strain (30). To quantify T. gondii organisms in vivo, genomic DNA was extracted using the Allprep DNA/RNA kit or the DNeasy Blood & Tissue kit (Qiagen) from peritoneal cells or organs of the infected mice, and the numbers of T. gondii genomes (B1 gene) were quantified by PCR using iTaq SYBR Green Supermix. All flow cytometry data were collected on a LSRII flow cytometer (Becton Dickinson) and were analyzed with FlowJo software (TreeStar).

Statistical Analysis.

Statistical significance was calculated with the log-rank test, the unpaired Student's t test, or the Mann–Whitney U test. All P values ≤0.05 were considered significant.