Toll-Like Receptor 3–TRIF Pathway Activation by Neospora caninum RNA Enhances Infection Control in Mice

Neospora caninum is a protozoan parasite closely related to Toxoplasma gondii and has been studied for causing neuromuscular disease in dogs and abortions in cattle. It is recognized as one of the main transmissible causes of reproductive failure in cattle and consequent economic losses to the sector.

ABSTRACT

Neospora caninum is a protozoan parasite closely related to Toxoplasma gondii and has been studied for causing neuromuscular disease in dogs and abortions in cattle. It is recognized as one of the main transmissible causes of reproductive failure in cattle and consequent economic losses to the sector. In that sense, this study aimed to evaluate the role of Toll-like receptor 3 (TLR3)–TRIF-dependent resistance against N. caninum infection in mice. We observed that TLR3^−/− and TRIF^−/− mice presented higher parasite burdens, increased inflammatory lesions, and reduced production of interleukin 12p40 (IL-12p40), tumor necrosis factor (TNF), gamma interferon (IFN-γ), and nitric oxide (NO). Unlike those of T. gondii, N. caninum tachyzoites and RNA recruited TLR3 to the parasitophorous vacuole (PV) and translocated interferon response factor 3 (IRF3) to the nucleus. We also observed that N. caninum upregulated the expression of TRIF in murine macrophages, which in turn upregulated IFN-α and IFN-β in the presence of the parasite. Furthermore, TRIF^−/− infected macrophages produced lower levels of IL-12p40, while exogenous IFN-α replacement was able to completely restore the production of this key cytokine. Our results show that the TLR3-TRIF signaling pathway enhances resistance against N. caninum infection in mice, since it improves Th1 immune responses that result in controlled parasitism and reduced tissue inflammation, which are hallmarks of the disease.

INTRODUCTION

Neospora caninum is an obligate intracellular parasite that belongs to the phylum Apicomplexa and is the causative agent of neosporosis. Canids have been described as its definitive hosts, while cattle, sheep, and other warm-blooded species act as intermediate hosts (1, 2), which can be infected by oral or transplacental routes. N. caninum is closely related to Toxoplasma gondii and has been studied in recent decades as a major cause of neuromuscular disease in dogs and repeated abortions in cattle; it is recognized as one of the most important identifiable causes of economic losses in the beef and dairy industries (3).

The innate immune response triggered by N. caninum plays an important role in protection of the host, leading to the activation of adaptive responses that restrict parasite proliferation. Innate cells express pattern recognition receptors (PRRs), like the Toll-like receptors (TLRs), that recognize pathogen-associated molecular patterns (PAMPs) (4–7). This recognition requires intracellular signaling through adapter molecules, mainly myeloid differentiation factor 88 (MyD88) or molecule-inducing beta interferon (IFN-β) with a TIR domain (TRIF) (8). Most TLRs share MyD88 as their adapter protein, with the exception of TLR3, which signals only through TRIF, and TLR4, which signals through both pathways, depending on the stimulus (9). The MyD88-dependent signaling pathway leads to the activation of mitogen-activated protein kinases (MAPK) and transcription factors like nuclear factor κB (NF-κB), which in turn activates the transcription of classic proinflammatory mediators. On the other hand, the TRIF-dependent pathway activates interferon-regulatory factors (IRFs), resulting in the production of type I interferon (IFN-β and IFN-α), which may also potentiate the production of proinflammatory factors, such as tumor necrosis factor (TNF) (8, 10, 11).

The MyD88-dependent pathways are known to be activated during N. caninum infection, acting as a relevant host resistance factor (12, 13). In this context, TLR2 and TLR11 participate in parasite recognition, leading to the activation of antigen-presenting cells and polarization of immune responses to a Th1 profile (14, 15). The participation of MyD88-dependent endosomal receptors, such as TLR7, -8, and -9, has also been suggested during infection in different animal species (16–18). Although there is little information about the specific involvement of TRIF in N. caninum infection in mice, it is known that, unlike T. gondii, TLR3-TRIF recognizes N. caninum RNA and induces type I IFN responses (19).

In that light, this study aimed to evaluate the role of the TLR3-TRIF-dependent signaling pathway in mouse resistance to N. caninum infection, using in vivo and in vitro approaches to observe phenotypes and to elucidate the molecular mechanisms involved in the interaction.

RESULTS

TRIF is required for proper resistance during infection by N. caninum.

In order to verify whether the TLR3-TRIF signaling pathway is required for host survival against a severe infection model, we monitored wild-type (WT) and genetically TLR3-, TRIF-, and MyD88-deficient (TLR3^−/−, TRIF^−/−, and MyD88^−/−) mice for 30 days after inoculation with N. caninum tachyzoites (1 × 10⁷ parasites/mouse). MyD88^−/− mice were used as controls, since it has been shown that MyD88-induced signaling is crucial for host resistance to neosporosis (12). WT mice survived the infection, whereas all the MyD88^−/− mice succumbed 18 days after inoculation (P < 0.05). For the TLR3^−/− and TRIF^−/− mice, we found reductions in the survival rate of 50% (P < 0.05) and 33%, respectively, demonstrating that this pathway participates in host resistance to infection (Fig. 1).

FIG 1.

Survival curves of WT, MyD88^−/−, TLR3^−/−, and TRIF^−/− mice (n = 5/group) monitored for up to 30 days after infection with 1 × 10⁷ N. caninum tachyzoites. Differences between groups were compared by Kaplan-Meier survival analysis through the log-rank Mantel-Cox test, with P values representing comparison to the infected WT group. The values are representative of three independent experiments. Differences were considered statistically significant at a P value of <0.05.

Then, in order to evaluate the specific participation of TRIF in the control of parasite growth in vivo, we infected WT and TRIF^−/− mice with a sublethal dose of N. caninum tachyzoites (5 × 10⁶ parasites/mouse) for the quantification of parasite genomic-DNA copies at distinct phases of infection (Fig. 2). We observed that TRIF^−/− mice presented increased parasite burdens in peritoneal cells after 24 h and 7 days of infection, in the lungs after 7 days of infection, and in the brain after 30 days of infection compared to the infected WT group. No significant differences in parasite loads were observed in the livers of the analyzed groups.

FIG 2.

TRIF restricts parasite growth in vivo. WT and TRIF^−/− mice (n = 5/group) were infected with 5 × 10⁶ tachyzoites of N. caninum and euthanized for collection of peritoneal cells and lung, liver, and brain tissue in the acute and chronic phases of infection. The data were obtained from qPCR by amplification of the Nc5 gene. Values are expressed as means and standard errors of the mean (SEM). The data were analyzed using the Mann-Whitney test. Differences were considered statistically significant at a P value of <0.05 (*, P <0.05). The values are representative of the results of two independent experiments.

Based on these findings, we evaluated the role of TRIF in controlling tissue injury and inflammatory lesions in the different tissues of the infected mice. Histological sections revealed pronounced lesions in the lungs and livers of the infected mice at 7 days postinfection, which notably worsened in the absence of TRIF and presented significant loss of tissue structure induced by extensive regions of mononuclear cell infiltrates and, in the most acute cases, necrosis. Inflammatory foci were also observed in brain samples from WT and TRIF^−/− mice; however, no differences in the inflammatory profile were noted between the analyzed groups (Fig. 3).

FIG 3.

TRIF limits pulmonary and hepatic inflammation during acute N. caninum infection. Shown are representative photomicrographs (A) and inflammatory scores (B) of histological sections obtained from WT and TRIF^−/− (n = 5) mice infected with 5 × 10⁶ tachyzoites of N. caninum. Pulmonary and hepatic tissue samples were collected after 7 days of infection and brain tissue after 30 days of infection. The arrows indicate areas of inflammatory cellular infiltrate. The slides were stained with hematoxylin-eosin and analyzed under an optical microscope (magnification, ×10). The values are representative of the results of two independent experiments. Differences were considered statistically significant at a P value of <0.05 (*, P <0.05).

TRIF supports the induction of appropriated Th1 immune profile responses against N. caninum.

As TRIF^−/− mice presented impaired resistance to infection, as shown by decreased survival rates, along with higher parasite burdens and greater tissue injury, we investigated whether the stimulation of adaptive immune responses, based on the production of a proper profile of chemokines, cytokines, and nitric oxide (NO), would also be compromised in TRIF^−/− mice during the acute phase of infection by N. caninum. For this purpose, we mapped important soluble factors involved in the activation, regulation, and chemoattraction of monocytes, lymphocytes, and macrophages in lysed spleen cells obtained from infected WT and TRIF^−/− mice during the acute phase (7 days of infection). The results of the array assay showed that TRIF^−/− mice produced reduced concentrations of several chemokines, such as IP-10, JE, CXCL9, and CCL5, compared to the WT (Fig. 4). In addition, we found that the concentration of major proinflammatory cytokines, such as interleukin 12p40 (IL-12p40) (Fig. 5A), IFN-γ (Fig. 5B), and TNF (Fig. 5C), were significantly reduced in the peritoneal fluids and lung homogenates of TRIF^−/− mice during acute infection. As expected, in vivo nitrate/nitrite production was also restricted in the absence of TRIF (Fig. 5D). These results indicate that this adapter molecule has a relevant role in the development of appropriate effector immune responses against N. caninum infection.

FIG 4.

The absence of TRIF impairs the production of key cytokines and chemokines during in vivo infection. Shown is a proteome profiler array of lysed spleen cells obtained from WT and TRIF^−/− mice 5 days after infection by N. caninum (5 × 10⁶ tachyzoites), showing the expression of different cytokines and chemokines produced by innate and adaptive immune responses. The boxed numbers indicate significantly altered expression of the immune targets, with each protein identified on the left.

FIG 5.

TRIF is relevant for production of proinflammatory cytokines and nitric oxide in vivo. Shown are quantifications of cytokine (IL-12p40 [A], IFN-γ [B], and TNF [C]) production and nitric oxide (D) in peritoneal fluids and lungs from WT and TRIF^−/− mice (n = 5/group) during the acute phase of infection by N. caninum (7 days postinfection [dpi]; 5 × 10⁶ tachyzoites/mouse). The results are expressed as means and standard errors and were analyzed using the two-way ANOVA test followed by Bonferroni’s posttest. Differences were considered statistically significant at a P value of <0.05 (*, P <0.05). The values are representative of the results of two independent experiments.

Unlike T. gondii, TLR3 and interferon response factor 3 (IRF3) are recruited to the N. caninum PV and nucleus, respectively.

After analyzing the relevance of the TLR3-TRIF-dependent signaling pathway in host resistance to N. caninum during in vivo infections, the next step was to elucidate aspects of the interaction between TLR3 and the parasite, also comparing it to T. gondii, which has been previously described as not recognizable by TLR3 (19). First, in order to verify the recruitment of TLR3 in response to N. caninum and T. gondii, we infected immortalized bone marrow-derived macrophages (iBMDMs), transfected with a TLR3-green fluorescent protein (GFP) construct, with tachyzoite forms (Fig. 6) or total RNA extracted from both parasites (Fig. 7). We observed that the recruitment of TLR3-GFP was induced in N. caninum-infected macrophages, which colocalized with the parasitophorous vacuole (PV) after 24 h of infection. On the other hand, T. gondii did not induce notable expression and focus formation of the receptor. In agreement with this, the exposure of iBMDMs to N. caninum total RNA induced the TLR3-GFP construct originally dispersed in the cytoplasm to form agglomerates inside the cell, while T. gondii RNA was again not able to induce a similar phenomenon. When a transfection reagent was mixed with RNA to force its uptake by iBMDMs, no formation of TLR3-GFP foci was observed; however, it was notable that TLR3 was abundantly expressed in cells exposed to N. caninum RNA, while GFP expression in cells stimulated with T. gondii RNA remained basically unaltered.

FIG 6.

Differential induction of TLR3 recruitment to the PV after infection with N. caninum and T. gondii. (A) Immortalized murine macrophages transfected with a TLR3-GFP construct after 24 h of infection with N. caninum (NcLiv_mCherry) and T. gondii (RH_RFP). (B) Quantification of TLR3 recruitment to the PV after 24 h of infection with both parasites. The images are representative of the results of three independent experiments. Differences were considered statistically significant at a P value of <0.05 (*, P <0.05).

FIG 7.

N. caninum RNA triggers TLR3 expression. (A) Immortalized murine macrophages transfected with a TLR3-GFP construct were plated and stimulated with total RNA (1 μg) extracted from N. caninum or T. gondii, with or without transfecting reagent (lipofectamine [Lipo]), for 24 h. (B) Quantification of TLR3 recruitment after stimulus with total RNA from both parasites, with or without transfecting reagent (Lipo). The images are representative of three independent experiments. Differences were considered statistically significant at a P value of <0.05 (*, RNA versus control; **, RNA plus Lipo versus Lipo control).

In the same manner, and still trying to comprehend the mechanisms involved in the recruitment of pathway components, we transfected macrophages with IRF3-GFP constructs, a precocious transcription factor responsible for the induction of IFN-α and IFN-β. These cells were infected with tachyzoites of N. caninum and T. gondii for 3 h and evaluated by fluorescence microscopy (Fig. 8). The results showed that infection by N. caninum strongly induced nuclear translocation of IRF3. Conversely, T. gondii was not able to activate IRF3, corroborating the results previously presented in relation to TLR3 recruitment.

FIG 8.

Differential induction of IRF3 nuclear translocation after infection with N. caninum and T. gondii. (A) Immortalized murine macrophages transfected with an IRF3-GFP construct after 3 h of infection with N. caninum (NcLiv_mCherry) or T. gondii (RH_RFP) (MOI =1). (B) Quantification of IRF3 nuclear translocation after 3 h of infection with both parasites. The images are representative of the results of three independent experiments. Differences were considered statistically significant at a P value of <0.05 (*, P <0.05).

TRIF is responsible for the induction of type I IFNs in macrophages infected with N. caninum, which have been shown to be relevant modulators of key IL-12 production.

We have demonstrated that the protozoan N. caninum is able to activate the signaling pathway initiated by TLR3, inducing the recruitment of the receptor and a related transcription factor to the PV. We also know from the literature that this signaling pathway is TRIF dependent and results in type I IFN (IFN-α and IFN-β) production, a possibility that we then investigated during infection of fresh BMDMs. Initially, to assess if N. caninum was able to modulate the expression of TRIF during infection, WT BMDMs were exposed to live N. caninum tachyzoites or stimulated with lipopolysaccharide (LPS) as a positive control, and the mRNA encoding the adaptor protein was quantified after 6 h of infection. We observed in this experiment that TRIF expression increased over 10-fold in N. caninum-infected BMDMs compared to naive cells (Fig. 9A). A similar increase in transcription was observed in the follow-up experiments, involving the quantification of IFN-α and IFN-β expression in infected BMDMs derived from WT mice. However, in the absence of TRIF, this expression was severely compromised. IFN-α expression was reduced approximately 40% in infected TRIF^−/− BMDMs (Fig. 9B), while differences in IFN-β transcript levels were even greater, with TRIF^−/− BMDMs inducing over 4-fold-reduced expression of the cytokine compared to WT BMDMs in the presence of the parasite (Fig. 9C).

FIG 9.

TRIF is upregulated in macrophages after infection and coordinates the gene expression of type I IFNs. (A) Quantitative TRIF expression in bone marrow-derived macrophages from WT mice infected with N. caninum (MOI = 0.5) or stimulated with LPS (1 μg/ml). (B and C) IFN-α (B) and IFN-β (C) gene expression in bone marrow-derived macrophages from WT and TRIF^−/− mice infected with N. caninum tachyzoites (MOI = 0.5) or stimulated with LPS. Data analysis was performed considering the expression results obtained by uninfected macrophages as reference values and are shown as fold change (relative increase of the final value in relation to the reference value). The results are expressed as means and standard errors and were analyzed using a two-way ANOVA test followed by Bonferroni's posttest; they are representative of the results of three independent experiments. Differences were considered statistically significant at a P value of <0.05 (*, P <0.05).

Our last step was to evaluate whether the TLR3-TRIF pathway and its known products (type I IFNs) would affect Th1 induction by macrophages. As previously described by Mineo and colleagues (12), MyD88 is a crucial adapter molecule in inducing an effective immune response against the parasite due to its pivotal role in IL-12 production, which in turn is the main IFN-γ-inducing element in the immune system, by polarized Th1 lymphocytes. With that intent, we performed experiments to elucidate whether the TLR3-TRIF pathway would directly affect this essential mechanism in host protection during N. caninum infection. For this, we analyzed the production of IL-12p40 by WT, TLR3^−/−, TRIF^−/−, and MyD88^−/− infected macrophages. In this set of experiments, we observed that TLR3^−/− and TRIF^−/− BMDMs failed to produce regular concentrations of IL12p40 compared to WT cells. As expected, in the absence of MyD88, no production of the analyzed cytokine was detected (Fig. 10A).

FIG 10.

The production of IL-12 is enhanced by the TLR3–TRIF–IFN-α pathway during N. caninum infection. (A) Quantification of IL-12p40 ELISA in the culture supernatants of BMDMs obtained from WT, TLR3^−/−, TRIF^−/−, and MyD88^−/− mice infected with N. caninum tachyzoites (MOI = 0.5). (B) Quantification of IL-12p40 production in the culture supernatants of BMDMs derived from WT, TLR3^−/−, TRIF^−/−, and MyD88^−/− mice treated or not with the recombinant cytokines IFN-α and IFN-β (10 U/ml) and infected with N. caninum tachyzoites (MOI = 0.5). Cytokine concentrations were determined after 24 h of infection. Values are expressed as means and standard errors (*, P <0.05). The values are representative of the results of three independent experiments.

In order to establish the mechanism by which TLR3-TRIF contribute to IL-12 production in macrophages, TRIF^−/− and WT BMDMs were treated with recombinant IFN-α and IFN-β and infected with N. caninum tachyzoites for subsequent determination of IL-12 levels. According to the results, treatment with IFN-α was able to completely reestablish IL-12p40 production in TLR3^−/− and TRIF^−/− BMDMs, while IFN-β had an overall inhibitory effect on the production of IL-12p40 in all macrophage lineages tested (Fig. 10B). None of the treatments with recombinant cytokines were able to induce IL-12p40 in MyD88^−/− BMDMs.

DISCUSSION

Neospora caninum is the obligatory intracellular parasite that causes neosporosis and has medical-veterinary importance mainly for causing neuromuscular paralysis in dogs and abortion in cattle, with significant economic impacts on countries that produce meat and dairy products. In order to prevent the spread of the infection and its clinical consequences, an understanding of the innate immune pathways triggered against the parasite, as well as the participation of parasite molecules that actively interfere with host resistance to the infection, is crucial (3).

In this work, we investigated how the TLR3-TRIF innate recognition pathway assists the host response against infection by N. caninum in mice. Initially, we aimed to evaluate the importance of this pathway in the survival of N. caninum-infected mice, using MyD88^−/− animals as a comparison, since the molecule is central to survival of the host during neosporosis (12). In our study, TLR3^−/− and TRIF^−/− mice presented decreased survival rates after infection, although portions of the experimental groups survived, unlike MyD88^−/− mouse groups, which fully succumbed to the N. caninum challenge. These experiments demonstrate that the TLR3-TRIF pathway participates indirectly in host resistance to infection, in a secondary/complementary role to MyD88.

As observed during in vivo experiments, our results showed that TRIF is also related to the regulation of effector host immune responses, since TRIF^−/− mice presented higher parasite burdens associated with low concentrations of IFN-γ and NO, which are essential for parasite elimination (14). Nonetheless, intense tissue inflammation was observed in the lungs of TRIF-deficient mice with significant loss of structure by pulmonary alveoli, characterizing the severity of the infection in the absence of key immune factors. Hsia and collaborators demonstrated that TRIF^−/− mice presented higher numbers of lung injuries due to impaired ability of lung dendritic cells to support the polarization of the immune responses to a Th1 profile (20). In addition, we observed in our study that in the absence of TRIF there was hepatic injury in the acute phase of the infection, with a high number of inflammatory foci, followed by tissue injury. In the case of viral infections involving the liver, such as infection by hepatitis B virus, it has been shown that there is a significant reduction in the expression of TRIF in hepatocytes as a form of viral evasion of the immune response, and the overexpression of this molecule inhibits viral replication and expression of viral proteins (21).

Previous work indicated that inducible nitric oxide synthase (iNOS) (nitric oxide synthase 2) is absent and NO synthesis is totally impaired in TRIF-deficient murine peritoneal cells stimulated with poly(I·C) or LPS (22). There is also evidence that TRIF participates, not only in NO production, but also in the induction of reactive oxygen species (ROS) by macrophages via TLR3 (23). In addition, we have observed that the absence of this adapter molecule also leads to impaired production of IL-12p40, TNF, and important chemokines, indicating that the TRIF pathway acts in immune control of the acute phase of disease through the potentiation of specific Th1 responses, coordinated by MyD88.

Based on these results, we decided to explore the mechanism by which N. caninum interacts with the TLR3-TRIF pathway. Despite its phylogenetic similarity to T. gondii, N. caninum has many distinct biological characteristics, especially regarding its interaction with the host cell. Our results demonstrate that one of these differences is the ability of N. caninum to induce the innate recognition pathway dependent on the TLR3 receptor, which is well characterized for recognizing viral nucleic acids. We have shown that the RNA of the protozoan N. caninum triggers the induction of TLR3, recruiting the receptor dispersed in the cytoplasm for the formation of an active agglomerate around the PV. Similar results were obtained by Beiting, demonstrating that protozoan parasites are capable of inducing type I IFN responses and that N. caninum performs this activation at a higher magnitude than T. gondii (24). We also evaluated the behavior of the transcription factor IRF3 in response to N. caninum and observed that N. caninum was able to induce IRF3 nuclear translocation, unlike T. gondii. Beiting demonstrated that N. caninum is also able to induce increased expression of IRF7, another transcription factor involved in antiviral responses (24).

Regarding the ability of the TLR3-TRIF pathway to induce IFN-α and IFN-β, here, we have confirmed this mechanism in the Neospora model. Many examples of this induction may be found in the literature: in a study that used mesothelial cells, which act as a protective barrier against invasive pathogens, it was observed that stimulation with poly(I·C), a TLR3 agonist, induced the expression of IFN-β mRNA in WT cells; however, this induction was significantly reduced in TRIF-deficient cells (22). In other infection models, such as the bacteria Listeria monocytogenes and Chlamydia muridarum, it has been observed that macrophages deficient in TRIF also present impaired IFN-β expression, demonstrating again the strict relationship of the adapter molecule to the cytokine (25, 26). On the other hand, it is well known that the development of a suitable Th1 immune response is considered of primary importance for protection against intracellular parasites such as N. caninum. The information that was new to us was the link between type I IFN production and the establishment of a robust and classical Th1 profile, mediated through the partly TLR3-TRIF-dependent production of IL-12. This Th1 induction pathway is not usually assessed by the scientific community, but it has been previously shown that LPS stimulation in the absence of TRIF significantly compromises the production of proinflammatory cytokines like IL-12p40 and TNF (27). Furthermore, we also showed the direct effect of type I IFN within this context, as IFN-α completely rescued IL-12 production in TRIF^−/− macrophages. Similarly, it was demonstrated that IFN-α treatment induced high levels of IL-12 secreted by these cells (28). In addition, it is known that the actions of IFN-α and IL-12 together can activate macrophages and NK cells and elevate the levels of other proinflammatory cytokines, generating a positive-feedback loop (29, 30).

Together, our results show that, unlike T. gondii, N. caninum is able to induce the activation of the TLR3–TRIF–IRF3–IFN-α/β signaling pathway, classically known to act on viral infections. In addition, we have demonstrated that TRIF is required for proper resistance against infection induced by N. caninum, enhancing Th1 immune responses against the parasite, which confers increased survival, reduced parasite burden, and decreased inflammatory lesions on the host. Based on these findings, future work should be directed to determining the actual TLR3 agonist contained within the N. caninum RNA, which may be used for new prophylactic and/or therapeutic approaches aimed at neosporosis or other conditions that require Th1 boosting.

MATERIALS AND METHODS

Ethics statement.

All experiments with mice were approved by the animal research ethics committee at Universidade Federal de Uberlândia (UFU) (Comitê de Ética na Utilização de Animais da Universidade Federal de Uberlândia [CEUA/UFU]; protocol number 109/16) and were carried out in accordance with the recommendations in the International Guiding Principles for Biomedical Research Involving Animals of the International Council for Laboratory Animal Science (ICLAS), countersigned by the Conselho Nacional de Controle de Experimentação Animal (CONCEA) (Brazilian National Council for the Control of Animal Experimentation) (https://olaw.nih.gov/sites/default/files/Guiding_Principles_2012.pdf). The institution’s animal facility (Centro de Bioterismo e Experimentação Animal [CBEA/UFU]) is accredited by the National Commissions in Animal Experimentation (CONCEA, CIAEP) (01.0105.2014) and Biosecurity (CTNBio [CQB]) (163/02).

Parasites.

N. caninum tachyzoites (NcLiv isolate) and T. gondii (strain RH) were maintained in confluent monolayers of HeLa cells (CCL-2; ATCC) at 37°C with 5% CO₂ in RPMI 1640 medium supplemented with glutamine (2 mM) and antibiotic/antimycotic solution (Thermo Scientific, Wilmington, MA). Following cellular disruption, the extracellular parasites were obtained by centrifugation of the supernatant at 800 × g for 10 min at 4°C. The pellet was resuspended in RPMI 1640, and the tachyzoite numbers were estimated in a Neubauer chamber for in vivo and in vitro experiments. For in vitro colocalization assays, parasite strains constitutively expressing red fluorescent protein (RFP) were used. These modified strains were kindly supplied by Peter Bradley (NcLiv_mCherry) and Vern Carruthers (RH_RFP) and maintained under conditions similar to those for wild-type parasites.

Animals and experimental infections.

WT C57BL/6 mice, along with genetically TLR3-, TRIF-, and MyD88-deficient (TLR3^−/−, TRIF^−/−, and MyD88^−/−) littermates, 6 to 8 weeks old, were obtained and were maintained at CBEA/UFU in individually ventilated cages under controlled conditions of temperature (22 to 25°C) and light (12 h/12 h) without restriction of water and food and subjected to acute (24-h and 7-day) and chronic (30-day) infection protocols, as well as analysis of survival after infection with N. caninum tachyzoites. Each experiment was conducted with 5 animals per group, with two or three independent experiments each.

Quantification of the parasite burden.

The parasite burdens of peritoneal cells and other sampled tissues were determined by quantitative PCR (qPCR) normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH), as previously described (31). Briefly, DNA was extracted from 50 mg of cells or tissue using SDS and proteinase K, and its content was estimated by 260/280 ratio (NanoDrop; Thermo Scientific). The concentration was adjusted to 40 ng/μl for qPCR assays, which used specific primers designed for the Nc5 sequence of N. caninum and GAPDH listed in Table 1. The parasite burden was estimated by extrapolation of the number of Nc5 copies in the samples (three technical replicates per sample) compared to predetermined standard curves with DNA obtained from known counts of N. caninum tachyzoites, using dedicated equipment and its associated software (Step One Plus; Thermo Scientific).

TABLE 1.

Oligonucleotides used in this study for the transfection of GFP constructs and quantification of gene expression or parasite burden

Primer	Oligonucleotide (5′–3′)	Application
TLR3 forward	ATGAAAGGGTGTTCCTCTTATC	Transfection assay
TLR3 reverse	ATGTGCTGAATTCCGAGATCC	Transfection assay
IRF3 forward	ATGGAAACCCCGAAACCG	Transfection assay
IRF3 reverse	GATATTTCCAGTGGCCTGG	Transfection assay
IFN-α forward	TGTCTGATGCAGCAGGTGG	Gene expression
IFN-α reverse	AAGACAGGGCTCTCCAGAC	Gene expression
IFN-β forward	AAGAGTTACACTGCCTTTGCCATC	Gene expression
IFN-β reverse	CACTGTCTGCTGGTGGAGTTCATC	Gene expression
GAPDH forward	CTCGTCCCGTAGACAAAATGG	qPCR
GAPDH reverse	AATCTCCACTTTGCCACTGCA	qPCR
Nc5 forward	GCTGAACACCGTATGTCGTAAA	qPCR
Nc5 reverse	AGAGGAATGCCACATAGAAGC	qPCR

Histological analysis.

Samples of lungs, livers, and brains were collected from WT and TRIF^−/− mice at different stages of infection and analyzed for inflammatory infiltrates and tissue damage. For this, the samples were fixed in buffered formalin and maintained in 70% alcohol before being subjected to paraffin embedding, ultrathin sectioning, and subsequent staining with hematoxylin and eosin (32), which were performed with at least 6 sections/organ/mouse. Tissue injury scores were based on scales of zero to three (including partial scores), according to a procedure published previously (33). Image acquisition was performed using an automated inverted microscope (FSX100; Olympus, Japan).

Cytokine measurement.

For the measurement of cytokine profiles and chemokine production, spleens from WT and TRIF^−/− mice infected by N. caninum (5 × 10⁶ tachyzoites) were collected after 5 days of infection in RPMI medium. The cell suspensions were washed in medium and treated with lysis buffer (0.16 M NH₄Cl and 0.17 M Tris-HCl, pH 7.5) and resuspended in complete RPMI medium containing 10% fetal bovine serum (FBS). Pooled viable cells were adjusted to 1 × 10⁷ for the cytokine array by Western blotting (Proteome Profiler Array; R&D Systems, Minneapolis, MN, USA) according to the manufacturer's instructions. In addition, production of the main proinflammatory cytokines (IL-12p40, IFN-γ, and TNF) was also measured individually by enzyme-linked immunosorbent assay (ELISA) in peritoneal fluids and lung homogenates from WT and TRIF^−/− mice (two technical replicates per sample), infected or not with 5 × 10⁶ N. caninum tachyzoites for 7 days.

For in vitro experiments, WT, TLR3^−/−, TRIF^−/−, and MyD88^−/− macrophages were plated in 96-well plates (2 × 10⁵ cells/well; 3 to 5 replicates per stimulus) and incubated for 18 h at 37°C and 5% CO₂. After incubation, the BMDMs were treated or not with the recombinant cytokines IFN-α and IFN-β (R&D Systems) at 10-U/ml concentration for 6 h, using LPS (1 μg/ml) as a positive control. The cells were infected with viable N. caninum at a multiplicity of infection (MOI) of 0.5 tachyzoite (i.e., 1 parasite/2 cells) and incubated for 24 h. The cell culture supernatant was collected for further quantification of cytokines by ELISA.

Measurement was performed using commercial ELISA kits (R&D Systems, BD Biosciences, San Diego, CA) according to the protocols recommended by the manufacturer, with two technical replicates per sample. The final concentrations of the cytokines were determined according to standard curves with known concentrations of recombinant proteins, and the results were expressed as picograms per milliliter, observing the respective detection limits for each assay: IL-12p40, 15.6 pg/ml; IFN-γ, 4.1 pg/ml; and TNF, 3.7 pg/ml.

Quantification of nitric oxide.

Quantification of NO was carried out in the peritoneal fluids of WT and TRIF^−/− mice infected with 5 × 10⁶ N. caninum tachyzoites. The assay determines the nitric oxide concentration based on the enzymatic conversion of nitrate to nitrite, followed by its colorimetric detection at 540 nm. The nitrate/nitrite concentration for each sample (two technical replicates per sample) was estimated from a standard curve, following the manufacturer’s instructions (R&D Systems; detection limit, 0.78 μmol/liter).

Differentiation of BMDMs, establishment of immortalized macrophage lineages, and functional tests.

BMDMs were generated from WT, TLR3^−/−, TRIF^−/−, and MyD88^−/− mice, using conditioned medium as previously described (34). The immortalization of BMDMs was accomplished through a protocol based on prolonged exposure of the differentiated cells to the Cre-J2 retrovirus (35).

For the generation of murine BMDMs expressing fluorescent TLR3 and IRF3, total RNA was extracted with TRIzol from brains of wild-type mice, and the cDNA was synthesized using a commercial kit (GoScript reverse transcription system; Promega, Madison, WI, USA), according to the manufacturer's instructions. Gene amplification was performed by PCR using specific primers for TLR3 and IRF3, as listed in Table 1. The resulting PCR products were cloned in the mammalian expression vector pcDNA 3.1(+) encoding GFP (Invitrogen-Life Technologies). The plasmid was linearized with ScaI-HF for stable transfection in 1 × 10⁶ immortalized murine macrophages placed in 24-well plates, using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA). The transfection efficiency was monitored by fluorescence microscopy after 72 h, and the GFP-positive cells were drug selected (Geneticin, G418; HyClone Laboratories) and subsequently cloned by limiting dilution prior to functional tests. The selected clones were placed in 24-well plates (1 × 10⁶/well) and infected with N. caninum (NcLiv_mCherry) or T. gondii (RH_RFP) tachyzoites (MOI = 1). In other sets of experiments, macrophages were stimulated with 1 μg of total RNA extracted from both parasites, accompanied or not by Lipofectamine (1.5 μl, according to the manufacturer's guidelines), for 24 h. The quantification of the assays was based on the frequency of colocalization of TLR3-GFP with the PV and IRF3-GFP with the nucleus in infected cells, as well as the percentage of positivity of macrophages expressing TLR3-GFP after stimulation with N. caninum or T. gondii RNA. Briefly, the quantification was performed in 3 independent experiments that comprised 3 to 5 technical replicates each. For the analysis, 10 merged images of each condition, containing between 50 and 100 cells, were examined for the number of cells infected with N. caninum or T. gondii tachyzoites (presence of red fluorescent proteins) that showed green fluorescence in the PV (for TLR3) or nucleus (for IRF3). In relation to the stimulation of macrophages with RNA, the cells with a perceptible increase in green fluorescence were counted, with basal GFP expression discounted.

Quantification of gene expression in macrophages.

After differentiation, WT and TRIF^−/− BMDMs were plated in 24-well plates (1 × 10⁶ cells/well) and left at 37°C and 5% CO₂ for 18 h. The cells were then infected or not with N. caninum tachyzoites (MOI = 0.5) and left for 6 h before RNA extraction, cDNA synthesis, and subsequent analysis of the expression of TRIF, IFN-α, and IFN-β by qPCR (3 to 5 replicates for each condition in three independent experiments). RNA extraction was performed with the TRIzol (Thermo Scientific) protocol. After this procedure, the purity and yield were determined in a spectrophotometer for subsequent cDNA synthesis using a commercial kit (GoScript; Promega). The gene expression assay (qPCR) was performed using SYBR green (Promega) with specific primers listed in Table 1 (three technical replicates/sample). The gathered data were analyzed by relative gene expression, as previously described (36), using the GAPDH gene as the housekeeping gene. The data were displayed as the fold increase in gene expression, using uninfected WT macrophages as the baseline parameter.

Statistical analysis.

Statistical analyses were performed using dedicated software (Prism 6.0; GraphPad Software Inc.). The results were expressed as means and standard errors, and the differences were considered statistically significant when the P value was <0.05, as determined by two-way analysis of variance (ANOVA) with multiple-comparison Bonferroni posttests, t tests, or Mann-Whitney tests. For the survival of infected mice, we applied Kaplan-Meier survival analysis, followed by the log-rank Mantel-Cox test.