Abstract
Plasmacytoid dendritic cells (pDCs) are the major producers of IFN-α, an antiviral cytokine involved in immunomodulation and control of HIV type 1 replication, whereas Toxoplasma gondii is a life-threatening opportunistic infection in AIDS patients. During infection with HIV type 1, human pDCs decrease in circulation and remaining pDC produce lower amounts of IFN-α in response to viral stimulation. In this study, we investigated the impact of coinfection with T. gondii on the innate virus-directed responses of human pDCs. Using intracellular flow cytometry and fluorescence microscopy, we determined that T. gondii invaded but did not induce IFN-α or TNF-α in human pDC. However, T. gondii inhibited IFN-α and TNF-α produced in response to HSV and HIV, thus functionally inactivating pDC. IFN-α production was inhibited only in cells infected by T. gondii, which inhibited neither uptake of GFP-HSV nor localization of TLR9 in CD71+ endosomes, directing us to investigate downstream events. Using imaging flow cytometry, we found that both T. gondii and IL-10 inhibited virus-induced nuclear translocation, but not phosphorylation, of IFN response factor 7. Blockade of IFN response factor 7 nuclear translocation and inhibition of the IFN-α response was partially reversed by a deficiency in the T. gondii–derived ROP16 kinase, known to directly phosphorylate STAT3, a critical mediator of IL-10’s anti-inflammatory effects. Taken together, our results indicate that T. gondii suppresses pDC activation by mimicking IL-10’s regulatory effects through an ROP16 kinase-dependent mechanism. Our findings further imply a convergent mechanism of inhibition of TLR signaling by T. gondii and IL-10 and suggest potential negative consequences of HIV/T. gondii coinfection.
Introduction
Human plasmacytoid dendritic cells (pDC) are part of the innate immune system with the intrinsic ability to produce large quantities of type I IFN in response to viral nucleic acids. pDC constitutively express endosomal TLR7 and TLR9, which recognize viral RNA and DNA molecules, respectively (reviewed in Ref. 1). Following recognition of nucleic acid ligands by TLR7/9, an intracellular signaling pathway is triggered in pDC that leads to phosphorylation and nuclear translocation of IFN response factor (IRF)-7 (2). Human pDC can mount this effective response through their constitutively high expression of IRF-7 (3, 4). As a result of activation, phosphorylated IRF-7 binds to the promoter of type I IFN genes and initiates production of type I IFN (mostly IFN-α). Secreted IFN-α then controls viral replication and spread by inducing upregulation of various antiviral molecules and by enhancing immune responses. The status of pDC and IFN-α production in pDC from individuals with HIV infection is complex: in individuals with HIV infection, both numbers of pDC and their production of IFN-α in response to virus is diminished (reviewed in Refs. 5, 6), resulting in suppression of normal antiviral immune responses. Decreased pDC and decreased IFN-α production is strongly associated with progression to opportunistic infections (OI) in HIV-infected individuals. Other evidence, however, suggests that chronic stimulation of remaining pDC contributes to chronic immune activation in HIV infection (reviewed in Refs. 5, 6).
Although it is clear that HIV infection leads to enhanced susceptibility to opportunistic pathogens, whether OIs reciprocally and adversely impact the innate immune response to HIV is not known. A common OI in AIDS patients is caused by Toxoplasma gondii, an intracellular parasite that causes fatal encephalitis upon reactivation of the latent cyst stages of the parasite in the CNS (7, 8). Worldwide distribution of T. gondii has been estimated at 14% in the United States and up to 75% in Western Europe and sub-Saharan Africa, making it the leading CNS infection in regions with the highest frequencies of HIV infection (9). Upon active entry into the host cell, T. gondii, through its apical end, injects multiple protein factors from a specialized secretory organelle, known as rhoptries, to modify host cell signaling and immune responses (10). Prominent examples of such parasite-derived secreted factors include ROP18 and ROP16 kinases, which have been implicated in inhibition of IFN-γ–dependent antimicrobial effector responses and TLR-dependent IL-12 and TNF-α production, respectively (11). In mouse cells, the ROP16-mediated negative modulation of TLR responses was reported to be mediated by inhibition of nuclear translocation of NF-κB and by inducing phosphorylation of STAT3 and STAT6 (11–15). Given the global prevalence of T. gondii in the HIV-infected population and the known prowess for immunomodulation by the parasite, we used T. gondii coinfection of human pDCs to investigate (or model) the impact of opportunistic pathogens on the innate immune response by these cells to HIV and HSV, which stimulate pDC through TLR7 and TLR9, respectively. In this study, we report that T. gondii infection inhibits antiviral responses of human pDCs by functional mimicry of IL-10.
Materials and Methods
Isolation and culture of PBMC and pDC
Heparinized peripheral blood from healthy donors was collected with informed consent and PBMC were separated using Lymphocyte Separation Media (Mediatech). PBMC were cultured in RPMI 1640 media containing 10% FCS, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, and 25 mM HEPES (RPMI 1640 10%) at a final concentration of 2 × 106 cells per ml at 37°C in 5% CO2. For experiments requiring highly enriched pDC, a negative selection kit (Miltenyi Biotec) was used according to the manufacturer’s protocol. Average purity of enriched pDC using this kit was 93%, with an average yield of 0.16% of total PBMC. For ImageStream experiments, partial enrichment of pDC was conducted using the Miltenyi negative enrichment kit with half the recommended Ab and bead mixture, and purity of 60–80% was obtained.
T. gondii
Parasite strains RH, GFP-RH were obtained from the American Type Culture Collection; mCherry-RH was kindly provided by Marc-Jan Gubbels (Boston College, Boston, MA). RHΔku80 (RH) and RHΔku80Δrop16 (RHΔrop16) knockout strains were kindly provided by D. Bzik (Dartmouth Medical School, Hanover, NH) (13). All T. gondii strains were grown in a monolayer of human foreskin fibroblasts for 48 h prior to harvest. Then 1 h before infection of pDC, parasites were collected from human foreskin fibroblasts by passage through a 25 gauge needle. Parasites were then washed at 2000 × g and resuspended in RPMI 1640 10%. Infection of 1 million PBMC or 40,000 purified pDC was conducted at multiplicities of infection (MOI) as described in the figure legends. PBMC or pDC with added T. gondii were then spun down at 100 × g for 3 min to enhance contact between parasites and cells. Parasite cultures were tested for Mycoplasma contamination using a PCR primer set (Agilent Technologies). Parasite and PBMC cultures were given a 1 h head-start at 37°C in 5% CO2 prior to stimulation of cells with viruses.
HSV type 1 and HIV type 1
Vero monkey kidney cells (American Type Culture Collection) were infected with HSV type 1 (HSV-1) strain 2931 (originally obtained from Dr. C. Lopez, then of Sloan-Kettering Institute, New York, NY) until development of cytopathic effect. Cells were then lysed using a sonicator, debris was centrifuged down, and virus aliquots were stored at −80°C. HSV-1 titer was assessed using plaque formation assay in Vero cells, and virus was used at an MO1 of 1. HSV K26-GFP was originally provided by P. Desai (Johns Hopkins University, Baltimore, MD), grown in Vero cells, and concentrated using a sucrose density gradient. AT-2 inactivated HIVMN was kindly provided by Dr. J. Lifson (Frederick National Laboratory for Cancer Research, Frederick, MD).
Pretreatment of pDC with IL-10
PBMC were pretreated with recombinant human IL-10 (PeproTech), 50 ng/ml, for 1 h prior to addition of the virus.
Flow cytometry
For detection of pDC in PBMC populations, Fc receptors were blocked using 2% heat-inactivated pooled human serum for 5 min, and cells were stained using CD123-APC (clone 6H6; BioLegend) and CD303-PE (BDCA-2, clone AC144; Miltenyi Biotec), at 4°C in 0.1% BSA in PBS buffer for 15 min. Following washes, cells were fixed overnight in 2% paraformaldehyde (PFA) in PBS. Intracellular staining for IFN-α as conducted by permeabilizing cells using 0.5% saponin in 2% FCS PBS followed by incubation with biotinylated anti–IFN-α Ab (PBL), which was then detected with streptavidin-PeCy7 (BioLegend); at the same time, cells were stained with anti–TNF-α–Pacific Blue (clone MAb-11; BioLegend). A BD LSR II was used to collect 300,000 events and analysis was performed using FlowJo software (Tree Star).
Determination of effect of T. gondii on annexin V/7AAD staining
Annexin V/7AAD (BD Biosciences) staining was carried out in purified pDC after 18 h exposure to GFP–T. gondii. Surface staining was carried out as described above and annexin V–biotin followed by streptavidin-Alexa680 on cells gated for T. gondii–GFP.
Immunofluorescence microscopy
Highly enriched pDC were surface stained and fixed in 4% PFA for 10 min at room temperature, and stained and permeabilized using the same protocol as for flow cytometry. Following the last wash and fixation, cells were cytospun using a Shandon Cytospin 2 at 1000 rpm for 6 min. ProLong Antifade Gold (Invitrogen) was used to mount coverslips.
Phospho-flow analyses
Freshly isolated PBMC were allowed to rest for 30 min at 37°C in 5% CO2. Cells were then stimulated as described in the figure legends and fixed with an equal volume of 4% PFA at room temperature for 15 min. Next, cells were washed, surface stained with HLA-DR–Pacific Blue (clone G46-2.6; BD Biosciences) and CD123-APC (clone 6H6; BioLegend), followed by washing and another 10 min 4% PFA fixation. Permeabilization was carried out with 70% ethanol in H2O at −20°C, which was added dropwise to tubes incubated on ice; cells were then stored at −20°C overnight. Permeabilized cells were washed twice with PBS containing 2% FCS and stained using anti-phospho IRF-7–PE (pS477/pS479, clone K47-671; BD Biosciences) or anti-phospho STAT3-PerCpCy5.5 (pY705, clone 4/P-STAT3; BD Biosciences). T. gondii infection was measured using anti-SAG1 mAb (clone p30, IgG2a; GenWay Biotech), which was labeled with Zenon Kit-488 (Invitrogen). Matched isotypes at the same concentration as the phospho-flow Abs were used in each experiment.
Virus internalization
Enriched pDC were preinfected with T. gondii mCherry-RH for 1 h followed by addition of HSV-GFP at an MOI of 10 and cells were cocultured for 20, 40, or 60 min. Cells were then washed with cold PBS buffer containing 5 mM EDTA to remove surface-attached virus and were fixed with 2% PFA overnight. Samples were acquired using an ImageStream IS100 or ImageStreamX Mark II (Amnis) with an extended depth of field element. Uptake of virus was analyzed using a combination of algorithms to quantify internalization of GFP particles. First, single cells were gated using a singlet gate based on the Aspect Ratio to the Area of Brightfield, followed by creation of an eroded cell morphology mask to gate on GFP-positive events within the cell area. Percentage of pDC positive for HSV-GFP was calculated by dividing the number of internalized GFP events by the total number of collected pDC.
Determination of cytokine levels
In total, 40,000 highly enriched pDC were incubated in wells of a 96-well round-bottom plate with a combination of T. gondii and HSV or HIV as described above, in a final volume of 200 μl. Culture supernatants were collected 18 h post-addition of virus and analyzed using the BenderMed Systems human IFN-α module set and the OptEIA Human TNF-α Set. Absorbance was recorded at 450 nm using a TECAN GENios reader and concentration of cytokines was calculated using Microsoft Excel from a linear curve fit. For determination of IL-10 levels, samples were sent to Myriad RBM and analyzed for expression of inflammatory cytokines using Human InflammationMAP v.1.0.
Nuclear translocation of IRF-7
In total, 100,000 enriched pDC were preinfected with T. gondii RH as described above. Surface staining was conducted at 4°C in 0.1% BSA in PBS using a combination of BDCA2-PE and BDCA4-PE (Miltenyi) Abs for detection of pDC. Cells were then fixed overnight using 2% PFA in PBS, and permeabilized using 0.1% TritonX in 2% FCS PBS. Intracellular IRF-7 staining was carried out at room temperature using a mAb (clone G-8; Santa Cruz) labeled with Zenon–Alexa Fluor 488 (Invitrogen), and the parasite’s parasitophorous vacuole membrane (PVM) was visualized using rabbit anti-GRA7 antiserum followed by anti-rabbit Alexa 610–R-PE (Invitrogen); the rabbit anti-GRA7 antiserum was produced by the Yap laboratory by immunization of a rabbit with full-length GRA7 protein. Following washing, cells were resuspended in 50 μl 1% PFA in PBS. All staining was performed in siliconized Eppendorf tubes (Sigma) and tubes were spun in a swinging rotor centrifuge to minimize cell loss. DRAQ5 (1:500 final dilution; Cell Signaling Technology) was added immediately before acquiring cell images to visualize nuclei. Cells were acquired with the ImageStream IS100 or ImageStreamX Mark II (Amnis). T. gondii infection and IRF-7 translocation was analyzed using Amnis IDEAS application 3.0. First, single cells were gated based on an aspect ratio intensity of DRAQ5 (Ch6) versus the area of DRAQ (Ch6). A dot plot of IRF-7 intensity (Ch3) versus BDCA2 intensity (Ch4) was used to gate pDC positive for IRF-7 and BDCA2 signals. Further refinement of single cells was conducted by exclusion of noncircular pDC events using a circularity feature of BDCA2 events. The similarity object function between default nuclear mask (Ch6) and IRF-7 channel was used to calculate the scores for IRF-7 nuclear translocation. The translocation gate was then defined based on a mock-stimulated sample and confirmed by visual inspection of corresponding images. All images shown were collected using a 40× camera.
Statistical analyses
Experiments were analyzed using one-way ANOVA with Tukey or Dunnet posttest with GraphPad Prism. In cases where only two experimental groups were compared, a two-tailed Student t test was used. All bar graphs represent mean ± SEM of at least three independent experiments from different donors; *p < 0.05, ** p < 0.01, ***p < 0.001; NS, not significant.
Results
Infection of human pDC by T. gondii
To determine whether human pDC are permissive to infection by T. gondii, we used intracellular flow cytometry. pDC were infected with T. gondii expressing GFP and were stained for intracellular GRA7, to mark the formation of the PVM inside the host’s cytoplasm (Fig. 1A). The majority of T. gondii GFP-positive pDC were also positive for GRA7 (61%, upper right quadrant) by 6 h of infection, indicating active parasitization of the human pDC cytoplasm. Next, we used purified pDC to visualize establishment of PVM by immunofluorescence microscopy (Fig. 1B). pDC can be seen either in the process of being infected by T. gondii (GFP, green) or infected by T. gondii with established PVM (GRA7-positive, pink).
Results from these experiments were confirmed using Amnis ImageStream, which allowed us to visualize and quantify the infection of BDCA2-positive pDC by GFP-tagged T. gondii. Using IDEAS software, we quantified the infection of pDC while excluding extracellular parasites through the application of an internalization algorithm. In the representative experiment shown, 46% of the pDC were infected with T. gondii (Fig. 1C, lower panel), with representative images of T. gondii–infected cells shown in Fig. 1C, upper panel. To identify the optimal MOI of T. gondii for infection of pDC that would allow us to analyze both infected and uninfected pDC, we used traditional flow cytometry to gate on T. gondii–GFP positive pDC within the PBMC population. An MOI of 4–8 yielded infection rates of 40–60% (Fig. 1D). Despite low abundance of pDC in the PBMC population, T. gondii efficiently infected these cells of the innate immune system that are crucial for recognition and clearance of viral infections. Taken together, these data identify that an active infection of human pDC and establishment of PVM can be quantified and visualized using ImageStream and flow cytometry. Moreover, we determined that subsets of T. gondii–infected and uninfected pDC can be identified and investigated independently.
T. gondii infection inhibits pDC responses to HSV-1 and HIV type 1
Having found that T. gondii infects human pDC, we next investigated the effects of T. gondii infection on virus-induced production of IFN-α in pDC. Pepper et al. (16) reported that murine pDC produced IFN-α in response to T. gondii; in our hands, T. gondii by itself did not induce production of IFN-α in human pDC (Fig. 2A), nor did stimulation of these cells with profilin (data not shown), which is a T. gondii–encoded TLR11 ligand that has been reported to induce IL-12 production in murine pDC (16). These data indicate that human pDC are either unable to recognize T. gondii pathogen-associated molecular patterns resulting in an IFN-α response or that T. gondii actively inhibits IFN-α production. To address the latter hypothesis, PBMC were preinfected with T. gondii at an increasing MOI, followed by stimulation of the cells with HSV-1. With an increasing ratio of T. gondii to PBMC, there was an increasing inhibition of IFN-α production in response to HSV (Fig. 2A). This loss of IFN-α production could not be explained by additional cell toxicity postinfection with T. gondii as determined by annexin V/7AAD staining (Fig. 2B, upper right for T. gondii versus mock in upper right panel); as expected given the culture-labile nature of pDC, there was loss of viability in the mock cultured pDC, and a similar level of cell viability was seen in the T. gondii–infected pDC. For the combination treatment of T. gondii and HSV, there was actually improved survival of the pDC at 18 h postinfection compared with the other three groups (Fig. 2B, lower right).
To determine whether pDC function is directly or indirectly downmodulated by T. gondii, we examined IFN-α production under conditions where only a subset of the pDC was infected. Suppression of IFN-α was observed only in T. gondii–infected cells (green) whereas uninfected cells from the same sample responded normally to HSV stimulation (Fig. 2C, upper panels, black). Similar results were found for IFN-α production in response to AT2-inactivated HIV, which induces IFN-α in pDC through TLR7 signaling (Fig. 2C, lower panels).
To further investigate the inhibitory effects of T. gondii infection on production of cytokines, highly enriched pDC were stimulated with HSV. pDC produced neither IFN-α nor TNF-α in response to T. gondii; however, strong inhibition of IFN-α production (55% inhibition), and of TNF-α (64% inhibition) was observed in T. gondii–infected pDC stimulated with HSV (Fig. 2D), suggesting that the effect of T. gondii on pDC was not dependent on the presence of other cell types. Because recognition of HSV-1 by pDC occurs through TLR9 (17), we next assessed whether T. gondii also inhibits HIV type 1 (HIV-1)–induced IFN-α production in pDC, which is mediated through TLR7 (18). Similar to what we observed with HSV-1, T. gondii strongly inhibited production of both IFN-α and TNF-α by purified pDC in response to AT2-inactivated HIVMN (Fig. 2E). To further evaluate whether the effect of T. gondii on pDC requires infection of the pDC by the parasite, we carried out experiments in purified pDC infected with a suboptimal MOI of T. gondii to yield populations that contained both infected and uninfected pDC (Fig. 2F). Similar to what we observed in pDC within the whole PBMC population (Fig. 2C), significant inhibition of IFN-α occurred only in actively parasitized pDC, which were positive for PVM marker GRA7 (Fig. 2F); pDC from the population exposed to T. gondii but not actively infected by the parasite produced IFN-α at the same level as those stimulated with HSV in the absence of T. gondii. Heat-inactivated T. gondii, which is unable to actively invade cells, failed to inhibit virus-induced IFN-α production by pDC (Fig. 2F).
T. gondii infection of pDC affects neither uptake of GFP-tagged HSV nor the endosomal localization of TLR9
To determine whether failure to take up the virus accounted for the deficiency in IFN-α and TNF-α production in T. gondii–infected pDC, we evaluated the uptake of GFP-tagged HSV by these cells using imaging flow cytometry. T. gondii infection did not reduce the capacity or alter the kinetics of pDC for virus uptake (Fig. 3A, left panel), and cells were clearly visible that had taken up both HSV and T. gondii (Fig. 3A, right panel). To further characterize the effect of T. gondii infection on the TLR9 pathway, we quantified the colocalization of TLR9 with CD71, a marker of early and recycling endosomes, in pDC either infected or uninfected with T. gondii and in the presence or absence of HSV-1 stimulation. Unstimulated pDC expressed high levels of TLR9 in CD71-positive endosomes as quantified by high constitutive colocalization of these two molecules. Similarity scores quantify the colocalization of the two markers, with scores higher than 1 indicating colocalization of TLR9 and CD71 as confirmed by visual inspection of representative events on the right in Fig. 3B. Neither HSV stimulation nor T. gondii infection affected the median similarity score of 1.9 across all experimental conditions (Fig. 3B). Moreover, T. gondii infection did not decrease cytoplasmic levels of TLR9 (as measured by TLR9 mean fluorescent intensity), indicating that TLR9 in T. gondii–infected pDC remains accessible for engagement by HSV (data not shown). Together, these data strongly suggest that T. gondii infection neither prevents uptake of virus nor affects the endosomal localization of TLR9, indicating that the target of inhibition is located downstream of TLR9 engagement.
T. gondii infection does not inhibit phosphorylation of IRF-7 but abrogates IRF-7 nuclear translocation in virus-stimulated pDC
Having ruled out inhibition of initial steps in viral uptake and TLR9 cellular localization in CD71-positive endosomes as being responsible for the inhibitory effects of T. gondii infection on pDC, we next investigated downstream events of TLR9-mediated recognition of HSV-1. We have previously shown that pDC express high constitutive levels of IRF-7 (3), the major transcription factor required for IFN-α production by pDC (2). First, we measured the phosphorylation of IRF-7 in response to virus, which is crucial for production of IFN-α in pDC. Phospho–flow cytometry results indicate that by 3 h postexposure of pDC to HSV, IRF-7 is phosphorylated on serine 477/479. T. gondii infection of pDC did not inhibit the HSV-induced phosphorylation of IRF-7. Interestingly, T. gondii–infected pDC (in the absence of virus stimulation) showed slightly higher pIRF-7 mean fluorescent intensities than did the mock stimulated cells (Fig. 4A). Even though phosphorylation of IRF-7 was not inhibited by T. gondii, infection by the parasite resulted in slightly decreased cytoplasmic levels of IRF-7 and T. gondii prevented HSV-1 induced upregulation of this transcription factor (Fig. 4A, data not shown).
Next, we assessed whether HSV-triggered nuclear translocation of IRF-7 is affected by T. gondii infection. In agreement with our previous results (4, 19), HSV treatment of pDC resulted in strong nuclear translocation of IRF-7 (Fig. 4B, 4C). T. gondii alone did not result in IRF-7 translocation (Fig. 4C, black square, at 6 h). However, T. gondii infection of pDC blocked HSV-induced IRF-7 nuclear translocation, reaching statistical significance at 6 h (Fig. 4B, 4C). Taken together, these data convincingly show that T. gondii infection of pDC inhibited nuclear translocation of IRF-7 in response to stimulation with HSV-1 without attenuating phosphorylation of IRF-7.
T. gondii mimics IL-10 anti-inflammatory signaling
To understand the molecular basis of T. gondii–mediated inhibition of IRF-7 nuclear translocation, we investigated the hypothesis that the parasite mimics IL-10 anti-inflammatory signaling in pDC without signaling through IL-10, as previously suggested by Butcher et al. (11) in a murine macrophage system. Using purified pDC, we observed that pDC did not produce IL-10 in response to HSV-1, HIV-1, or T. gondii following overnight stimulation with these pathogens (data not shown). Common to IL-10 signaling and T. gondii infection is the phosphorylation of STAT3 (11, 20). Similar to previous findings in other cell types, T. gondii infection of pDC resulted in strong phosphorylation of STAT3 on tyrosine 705 by 5 h (Fig. 5A, upper right quadrant), whereas T. gondii–uninfected pDC (lower left quadrant) did not exhibit STAT3 phosphorylation. These findings indicate that STAT3 phosphorylation is mediated in cis, only in T. gondii–infected cells, ruling out the effect of soluble extracellular factors in STAT3 phosphorylation, such as IL-10 or IL-6. As expected, treatment of pDC with recombinant human IL-10 resulted in rapid phosphorylation of STAT3 in the entire population of pDC (Fig. 5B).
We next directly compared the inhibitory effect of IL-10 treatment to the effect of T. gondii infection on HSV-induced IFN-α production. T. gondii infection and IL-10 treatment inhibited IFN-α to a similar extent (Fig. 5C). This experimental system confirms our previous published findings that IL-10 inhibits HSV-induced IFN-α production (21). Furthermore, to address the question of how IL-10 mediates inhibition of IFN-α, we focused on phosphorylation and nuclear translocation of IRF-7. As was seen with T. gondii coinfection, IL-10 stimulation of pDC along with HSV stimulation did not inhibit phosphorylation of IRF-7 (Fig. 5D) but did inhibit subsequent HSV-1–induced nuclear translocation of IRF-7 (Fig. 5E). These data thus reveal a previously unreported mechanism of inhibition of TLR signaling that is shared by IL-10 and T. gondii.
ROP16 is responsible for inhibition of IRF-7 nuclear translocation and inhibition of IFN-α production
To further test the hypothesis that parasite-mediated phosphorylation of STAT3 is responsible for the inhibitory effect of T. gondii, we used rop16 knockout parasite (13). ROP16 is a T. gondii virulence factor responsible for direct phosphorylation and activation of STAT3 and STAT6 (14, 22). STAT3 is a signal transducer activated by IL-10 that mediates the anti-inflammatory effect of this cytokine (20); therefore, by directly phosphorylating STAT3 in infected cells, T. gondii mimics IL-10 anti-inflammatory signaling. As expected, there was significantly less phosphorylation of STAT3 on tyrosine 705 in the pDC infected with rop16 knockout parasite than in the cells infected with the RH strain (Fig. 6A, 6B). To corroborate the involvement of ROP16 kinase in abrogation of TLR9 signaling, the effects of loss of ROP16 kinase on IRF-7 nuclear translocation and induction of IFN-α were investigated. rop16 knockout RH parasites failed to inhibit IRF-7 nuclear translocation as compared with the RH strain (Fig. 6C). This release of the IRF-7 blockade resulted in significantly higher production of IFN-α in response to HSV stimulation with the rop16 knockout versus RH strain (Fig. 6D). To normalize for the effect of parasite infection rate between the wild-type and knockout parasites, the intracellular expression of IFN-α in pDC infected with T. gondii was quantified using flow cytometry. No difference in infection rate between RH and RHΔrop16 was observed (data not shown). As shown in Fig. 6E, with gating on T. gondii–infected pDC (SAG1 positive, gray bars), rop16 knockout parasites were significantly less efficient than the RH strain in inhibiting IFN-α production. Interestingly, we were not able to fully reverse T. gondii–mediated inhibition of IFN-α, and we did not observe rescue in inhibition of TNF-α production in response to HSV-1 (data not shown). Therefore, these data suggest that ROP16-mediated phosphorylation of STAT3 is an important mechanism for T. gondii inhibition of IFN-α production, but that the parasite uses additional mechanisms for inhibition of TNF-α.
Discussion
Immunocompromised HIV-infected patients suffer from OIs including toxoplasmosis (8). In this investigation, we have shown that infection of human pDC by the opportunistic pathogen, T. gondii, results in the inability of pDC to respond to viral nucleic acids with the production of IFN-α, linking our study to T. gondii/HIV-1 coinfection in AIDS patients. Although T. gondii/HIV coinfections are widely known, to date no investigation has focused on understanding the molecular interaction of these two pathogens in the context of host antiviral responses. Based on our findings that T. gondii infection blocks IFN-α production by pDC without inducing death of pDC, we hypothesize that this mechanism of immune-modulation by reactivated parasites could further result in exacerbation of an immunocompromised state in HIV-infected patients.
T. gondii infects multiple cell types of the innate immune system including monocytes/macrophages, DC, and pDC (16, 23, 24). Pepper et al. (16) previously reported that mouse pDC became activated in response to T. gondii and responded with production of both IFN-α and IL-12. These intriguing findings led us to investigate the effect of T. gondii infection on human pDC in the context of viral coinfection and innate recognition. In our investigation, an IFN-α response was not detected in human PBMC population infected with T. gondii or in purified human pDC infected by T. gondii. This difference in host response versus the report of Pepper et al. most likely reflects interspecific variability in the immune responsiveness of mouse and human pDCs. A major difference between murine and human pDC relevant to this study is that human pDC, unlike mouse pDC, generally do not produce IL-12 (25). Tosh et al. (24) described human monocytes and CD1c+ DC as capable of producing IL-12 and TNF-α in response to T. gondii, but did not assess the pDC subset of DC. Moreover, Pepper et al. (16) reported that murine pDC recognized parasite ligand profilin through TLR11, a TLR that is not present in humans (26). Moreover, human pDC, unlike their mouse counterparts, express CD4, CCR5, and CXCR4, and are themselves susceptible to HIV-1 infection (27).
T. gondii, a protozoan eukaryotic pathogen, has been reported to use multiple mechanisms for evasion and modulation of immune responses. There are numerous reports of inhibition of NF-κB and MAPK signaling that result in their downmodulation of TNF-α and IL-12 production by macrophages (28, 29). Additional reports implicate induction of SOCS-1 and modulation of chromatin rearrangement as a mechanism for inhibiting TLR4 signaling in response to LPS (30–32). Because TLR4 signaling can use both MyD88-dependent or independent pathways, it was not clear whether the T. gondii inhibition of LPS effects were restricted to a specific pathway. The human pDC system used in this study allowed us to investigate the inhibition of purely MyD88-dependent, TLR7-, and TLR9-mediated signaling response to HIV-1 and HSV-1, respectively, with IFN-α as our major readout. With this system, we were able to identify that T. gondii targets inhibition of MyD88-dependent signaling through blocking the nuclear translocation of IRF-7, an inhibitory mechanism that has not been previously reported for T. gondii. This inhibition was not due to T. gondii toxicity on the pDC, as determined by annexin V/7AAD binding. Further evidence of a lack of general toxicity of the T. gondii on the pDC was found in that there was normal phosphorylation of IRF-7 and STAT1, as discussed below, in pDC infected with T. gondii. Although it is established that IL-10 acts as an anti-inflammatory cytokine through STAT3 signaling and induction of the suppressor of cytokine signaling (33), precisely how it mediates immunomodulation of pDC innate responses remains unclear. Our novel finding in Fig. 5 showing that IL-10 inhibited IRF-7 nuclear translocation but not phosphorylation provides a mechanism for our previously reported observation that IL-10 potently inhibits IFN-α production by human pDCs (21). In agreement with our previous observations (21), purified pDC failed to produce IL-10 in response to viral stimulation, nor did they produce IL-10 in response to T. gondii, thus indicating that the inhibition by T. gondii functionally mimicked the IL-10 pathway without being dependent on IL-10. Lack of a soluble mediator responsible for the effects of T. gondii on pDC IFN-α production was further indicated by experiments showing that within mixed populations of pDC, which were infected or uninfected with T. gondii, only the actively parasitized pDC were deficient in IFN-α production. Future investigations should address the presumed STAT3 requirement for IL-10’s effect on IRF-7 translocation and whether interference with IRF-7 signaling requires STAT3-driven de novo synthesis of new factors or STAT3/protein interactions in the cytosol. Moreover, the ability of T. gondii to induce STAT3 phosphorylation in pDC did not extend to STAT1 phosphorylation: T. gondii did not induce STAT1 phosphorylation, nor did it inhibit STAT1 phosphorylation in pDC in response to IFN-α (data not shown).
Upon infection of the host cell, T. gondii rapidly injects ROP16 into the cytoplasm and translocated into the host cell nucleus (34). ROP16 is a tyrosine kinase, which has been shown to mimic IL-10 and IL-4/IL-13 signaling through phosphorylation of STAT3 and STAT6 (14, 22). These reports of immune evasion through ROP16-mediated phosphorylation of STAT3 prompted us to investigate the effect of T. gondii infection on this purely MyD88-dependent signaling of TLR7 and TLR9 in pDC. Both of these TLRs are crucial for viral recognition and clearance, and MyD88 is crucial for host survival to T. gondii infection in mice. Our investigation of TLR7/9 signaling in the context of viral and T. gondii coinfection resulted in identification of inhibition of IRF-7 nuclear translocation in response to HIV and HSV that was similar to that observed with IL-10 inhibition of HSV-induced IRF-7 translocation. Knockout of rop16 from T. gondii resulted in the release of IRF-7 blockade and partially restored the IFN-α response; although statistically significant, this release of the IRF-7 blockade was not complete, suggesting that an additional mechanism responsible for the residual inhibition of IFN-α remains to be identified. It is possible that the remaining inhibition is also mediated through STAT3 because rop16 knockout parasites retain the ability to phosphorylate STAT3 on serine 727 (13).
Inhibition of type I IFN in pDC by T. gondii may have consequences for viral infections. IFN-α is crucial for control of viral replication through induction of multiple antiviral host response genes. These IFN-stimulated genes include, among many others: IRF-7, Trim5, Mx protein, PKR, and 2′5′ OAS, which inhibit translation, induce apoptosis, sequester viral nucleoproteins, and induce RNA degradation to control viral replication as well as stimulate a variety of antiviral host immune responses (35–38). Therefore, the loss in function in any of these genes can lead to decreased immune responses making the host more susceptible to viral infections like HIV-1. To our knowledge, this is the first report of inhibition of innate immune responses to HIV-1 and HSV-1 by T. gondii through inhibition of IFN-α production. Evasion of immune responses through blockade of IRF-7 nuclear translocation by T. gondii further adds to the complexity of immune-evasive maneuvers employed by this highly adaptable opportunistic pathogen.
Acknowledgements
We thank Marc-Jan Gubbels (Boston College, Boston, MA) for the provision of the mCherry-RH. RHΔku80 (RH) and RHΔku80Δrop16 (RHΔrop16) knockout strains were kindly provided by David Bzik (Dartmouth Medical School, Hanover, NH). The HSV K26-GFP virus was kindly provided by P. Desai (Johns Hopkins University, Baltimore, MD).
Footnotes
This work was supported by grants from the National Institute of Allergy and Infectious Diseases, National Institutes of Health (R01AI062806 and R01AI106125 to P.F.-B. and R56AI124691 and R01AI083405 to G.S.Y.) and National Institute for Research Resources Grant S10ODO18103 to P.F.-B. P.L.P. was supported by a predoctoral fellowship from the New Jersey Commission on Cancer Research.
References
Disclosures
The authors have no financial conflicts of interest.