Cutting Edge: CD36 mediates phagocyte tropism and avirulence of Toxoplasma gondii

Abstract

Resistance and tolerance are vital for survivability of the host-pathogen relationship. Virulence during Toxoplasma infection in mice is mediated by parasite-kinase dependent antagonism of IFN-γ induced host resistance. Whether avirulence requires expression of parasite factors that induce host tolerance mechanisms or is a default status reflecting the absence of resistance-interfering factors is not known. Here, we present evidence that avirulence in Toxoplasma requires parasite engagement of the scavenger receptor CD36. CD36 promotes macrophage tropism but is dispensable for development of resistance mechanisms. Instead CD36 is critical for re-establishing tissue homeostasis and survival following the acute phase of infection. The CD36-binding capacity of T. gondii strains is negatively controlled by the virulence factor, ROP18. Thus, the absence of resistance-interfering virulence factors and the presence of tolerance-inducing avirulence factors are both required for long term host-pathogen survival.

Introduction:

Two equally important host mechanisms, namely resistance and tolerance operate to ensure mutual survivorship in a microbial pathogen-host relationship (1). Resistance, typically mediated by effector functions of the innate and acquired immune response, operates to limit pathogen burden through mechanisms that block invasion or directly attack and eliminate invading microorganisms. Tolerance is defined as the ability of the host to withstand the tissue damage directly inflicted by the pathogenic organism or indirectly by the host’s own immune or altered physiological response to infection. A third critical factor that determines host-pathogen survival is the virulence of the invading pathogen. In the case of toxoplasmosis, virulence factors ROP18 and ROP16 secreted from specialized rhoptry organelles phosphorylate immunity related IRGs and Th2-associated STATs (STAT6 and STAT3), resulting in evasion of immune resistance (2). However, it is not known whether compromised host pathogen survivability might result from the pathogen’s failure to engage host mechanisms that promote disease tolerance. In this context, avirulence can be viewed as an affirmatively acquired status, rather than a default fate that results from the absence of virulence factor(s) required to evade or antagonize resistance mechanisms.

In 2014, we reported that avirulent strains of T. gondii exhibit an enhanced macrophage tropism by selectively adhering to and engaging with this host cell’s phagocytic machinery(3). Importantly, the parasite dashes demise and establishes productive infection by escaping the phagosome to establish its own vacuolar niche. We have previously hypothesized that avirulent T. gondii strains have the ability to engage an unidentified macrophage phagocytic receptor, a feature which the hypervirulent RH strain lacks. Here, we reveal the identity of this avirulence-associated phagocytic receptor and assess its impact on the development of host resistance and tolerance mechanisms during T. gondii infection.

Materials and Methods:

Experimental animals, cell lines and parasite strains

Wild type (WT) mice and CD36^−/− (4) C57BL/6 mice were procured from the Jackson Laboratory. Irgm3^−/− mice were bred in house. Mouse macrophage-like cell line J774A.1 was a gift from Xiaojing Ma (Weill Cornell Medicine, New York). Chinese hamster ovary cell line CHO745 (xylosyltransferase deficient) and CHO745-CD36 expressing human CD36 were generous gifts from Joseph Smith (University of Washington, Seattle, WA). T. gondii strains of GFP-PTG, GFP-RH and ME49 tachyzoites were purchased from ATCC. mCherry-RH was obtained from Marc-Jan Gubbels (Boston College, Boston, MA). The derivation of ROP18- and ROP16 deficient RH parasites were previously published. CTG strain parasites expressing wildtype and kinase-inactive ROP-18 were kind gifts from David Sibley (Washington University, St. Louis, MO)(5).

Parasite adherence and phagocytosis assays

Confluent BMDM, J774A.1, RAW264.7 or HFF cultures seeded in 24-well plates were first incubated with 2 million GFP-expressing parasites for 15 minutes at 12℃. The plate was then transferred into 37℃ incubator to initiate infection. At the indicated time points, infected cells were fixed in 4% paraformaldehyde and were subjected to surface α-SAG1 (P30/3) and phalloidin-Alexa Fluor 350 (Thermofisher Scientific) staining. For enumeration of vacuoles, infected cells were permeabilized in cold acetone and stained with rabbit α-GRA7.

Parasites infection of peritoneal macrophages and intracellular cytokines assays 1 million GFP-PTG or GFP-RH tachyzoites were i.p. injected into mice. Three hours later, peritoneal exudate cells were lavaged, stained for surface markers and analyzed using an LSRFortessa X-20 (BD Biosciences). Yolk sac-derived macrophages are defined as CD11b⁺/F4.80⁺/MHC-II^-/Ly6C^- while bone marrow-derived macrophages are defined as CD11b⁺/ F4.80^-/MHC-II⁺/Ly6C^-. For intracellular IL-12 and IFN-γ staining, the cells were surface stained, fixed and permeabilized with BD Cytofix/Cytoperm followed by intracellular cytokines staining.

Recombinant proteins binding assay to T. gondii tachyzoites

Recombinant human IgG1 Fc fragment (110-HG), mouse SR-B1-Fc chimera (9644-SR), mouse LIMPII/SR-B2-Fc chimera (1888-LM) and mouse CD36/SR-B3-Fc chimera (2519-CD) proteins were purchased from R&D Systems (Minneapolis, MN). 2 million parasites were incubated with 0.1μg recombinant protein in 100μl binding buffer (10mM HEPES, 0.14M NaCl, 2.5mM CaCl2, pH7.4) for 1h at 12℃. After extensive washing, the bound proteins were stained with donkey anti-human IgG-Alexa fluor 647 (Jackson ImmunoResearch Labs).

Parasitemia estimation in mouse tissues by quantitative PCR

Spleens and livers were harvested from wild type, CD36^−/− or Irgm3^−/− mice 7 days post 10 ME49 cysts infection. Parasite load in150ng genomic DNA was measured using QuantiTect SYBR Green PCR kit (Qiagen) targeting Toxoplasma B1 gene.

ELISA assays for IL-12, IFN-γ, GDF-15 and FGF-21 in mouse serum

Serum cytokines levels were measured using BD OptEIA ELISA kit for IL-12 (p40) and IFN-γ (BD Biosciences) and using DuoSet mouse GDF-15 and Quantikine ELISA mouse/rat FGF-21 kit (R&D Systems).

Statistical analysis

Data are represented as Mean ±SEM. Statistical analysis was carried out using Prism (GraphPad Software, San Diego, CA). Statistical significance was analyzed by two tails unpaired Student’s t-test or ANOVA.

Results and Discussion:

Identification of CD36 as the receptor mediating macrophage phagocytic uptake of avirulent T. gondii.

We have previously shown that avirulent PTG strain, but not virulent RH strain of T. gondii preferentially engage RAW264.7 and primary macrophages. To identify which receptor accounts for this avirulence-associated phagocytic activity, we screened additional macrophage cell lines and found that unlike RAW264.7 cells, J774A.1 cells were not adhesive and did not phagocytose avirulent parasites (supplemental Figure 1A). Examination of differential gene expression between these two cell lines using a published dataset (6) indicated that J774A.1 cells have low expression of the scavenger receptor CD36 (supplemental Figure 1B and C), pointing to CD36 as a candidate receptor. Indeed, CD36-deficient BMDMs lost their ability to engage, phagocytose and became less productively infected by avirulent parasites (Figure 1A–C). Expression of human CD36 in Chinese hamster ovary cells increases engagement of avirulent T. gondii, which is inhibited by an anti-CD36 monoclonal antibody (Figure 1 D–E). Taken together, data in Figure 1 identifies CD36 as the macrophage receptor selectively engaged by avirulent T. gondii.

Figure 1.

CD36 is the receptor mediating macrophage tropism of avirulent Toxoplasma. (A) Total GFP-PTG parasite attachment to bone marrow-derived macrophages (BMDM) from wild type or CD36^−/− mouse. (B) Extracellular GFP-PTG parasite-induced phagocytosis in wild type or CD36^−/− BMDM. (C) GFP-PTG parasite vacuole formation in wild type or CD36^−/− BMDM 30 minutes after infection. (D) Immunofluorescence staining of GFP-PTG parasites binding to CHO745 or CD36-overexpressing CHO745 cells. (E) α-CD36 antibodies blockade of GFP-PTG parasites binding to CHO745-CD36 by clone CRF D2712, HM36 or IgA isotype control at 2μg/ml. Scale bar, 10μm. The statistical significance is analyzed by student t-test. * p < 0.05, ** p < 0.01 and *** p < 0.001. Experiments were repeated at least 3 times.

Differential binding of recombinant CD36 to T. gondii strains and effect of ROP18 virulence factor.

To demonstrate direct engagement of CD36 by T. gondii, we examined binding of recombinant CD36 to extracellular tachyzoites by flow cytometry. As shown in Figure 2A and B, CD36 but not other related scavenger receptor Fc fusion proteins bound to PTG parasites specifically, as binding is inhibited by an antibody to CD36. Importantly, CD36 did not bind to virulent RH tachyzoites (Figure 2C). Mixing of avirulent and virulent strains in the same culture does not confer CD36 binding to virulent RH strains (Figure 2D), indicating strain-intrinsic, rather than host cell/environmental control of scavenger-ligand expression. To address the potential role of T. gondii virulence factors in controlling CD36 binding, we examined the effects of ROP16 or ROP18 deficiency in the virulent RH strain and ROP18 overexpression in the avirulent CTG strain. As shown in Figure 2E and 2F, ROP18 deletion in the RH strain revealed a cryptic ability of this strain to bind CD36, while transgenic expression of the ROP18 virulence factor in the avirulent CTG strain, which does not itself express ROP18, attenuated CD36 binding. Overall, these in vitro binding studies clearly indicate that avirulent strains of T. gondii selectively bind to CD36 and that the virulent strain acquires the ability to express the CD36-binding ligand when ROP18 is absent. Surprisingly, the virulence factor ROP18 attenuates the expression of the putative ligand binding to CD36 in a kinase-independent manner (Figure 2F), suggesting a non-canonical role for ROP18 distinct from its kinase-dependent inactivation of immune effectors. Although unexpected, this is not unprecedented as the ability of ROP-18 to modulate basal antigen-presentation and anti-tumor responses was also previously reported to be kinase-independent(7, 8).

Figure 2.

CD36 differential binding to T. gondii strains. (A) FACS staining of mouse recombinant protein SR-B1-Fc, SR-B2/LIMP-II-Fc and SR-B3/CD36-Fc or human Fc fragment binding to free GFP-PTG parasites. Data were pooled from five independent experiments. (B) Histogram showing recombinant CD36-Fc binding to free GFP-PTG parasites after α-CD36 antibodies clone CRF D-2712 or clone HM36 blockade. (C) CD36-Fc binding to extracellular GFP-PTG or GFP-RH. N=4. (D) Binding of recombinant CD36-Fc to mixed GFP-PTG and mCherry-RH parasites inoculated and lyzed from the same human foreskin fibroblast culture. (E) CD36 binding to parental RH and ROP18 and ROP16 mutant strains. (F) CD36 binding to parental avirulent CTG strain and CTG strains expressing wild-type (V1) and kinase-inactive (DAL1) ROP-18. Experiments were repeated at least 3 times.

Altered T. gondii tropism and heightened early innate responses in CD36−/− mice.

To examine the impact of CD36 on the initial cell tropism and innate immune response, we injected PTG tachyzoites ip into WT and CD36−/− mice and examined peritoneal exudate and spleen cells 3–4 hrs post inoculation. As shown in Figure 3, tracking of T. gondii infection in large yolk-sac derived macrophages versus small bone marrow-derived macrophages present in the peritoneal cavity revealed a selective effect of CD36-deficiency on infection tropism in the bone marrow-derived macrophage compartment (Figure 3A). As expected, this effect of CD36 was not observed when the pulse in vivo infection was performed using the virulent RH strain (Figure 3B). Interestingly, the innate IL-12 response in CD8⁺DCs and the IFN-γ response in NK cells were not deficient and were instead elevated (Figure 3C and D). These results suggest that CD36 regulates both infection tropism and the immediate innate response to T. gondii, but is not required to mount the early innate type 1 response to infection.

Figure 3.

Altered tropism and acute innate immune response in CD36^−/− mouse after injection of T. gondii. (A) GFP-PTG infection of two subsets of peritoneal macrophages in WT or CD36^−/− mouse 3h after i.p. pulse infection. (B) GFP-RH infection of two subsets of peritoneal macrophages in WT or CD36^−/− mouse 3h after i.p. pulse infection. N=3 (C-D) FACS plot showing percent IL-12-expressing CD8α⁺ DCs (C) or IFNγ-expressing NK cells (D) in spleens of wild type or CD36^−/− mouse 4h after GFP-PTG i.p. pulse infection. N=3.

CD36 deficiency results in enhanced mortality despite effective Th1 immunity to T. gondii.

Upon in vivo infection with avirulent ME49 cysts, CD36-deficient mice exhibited a higher rate of mortality relative to WT mice (Figure 4A). Although CD36−/− lost more weight (Figure 4B), mortality never approached 100% and occurred over a protracted period past the acute time points when IL-12, IFN-γ, Irgm3 or IL-10 deficient mice are known to succumb to infection(9–12). Indeed, CD36-deficiency did not lead to uncontrolled parasite replication and did not result in deficient production of the key proinflammatory cytokines IL-12 and IFN-γ required for immune control of the parasite (Figure 4C and D). CD36-deficiency also does not negatively affect the production of IL-10 (Figure 4E), the principal regulatory cytokine required to prevent excessive Th1 immunity and host immunopathology(9).

Figure 4.

Increased susceptibility of CD36^−/− mouse without immunodeficiency and loss of parasite replication control. (A) Survival of wild-type mice (n=10) and CD36-deficient mice (n=8) after intraperitoneal infection with 10 ME49 cysts. Log-rank test x² = 8.470; p=0.0036. (B) Weight loss (presented as percent initial weight) of wild type and CD36-deficient mice after infection in the same experiment as in (A). (C) Parasite loads in spleen and liver of wild type ( n=7), CD36^−/− (n=7) and Irgm3^−/− (n=4) mice 7 days after 10 ME49 cysts infection. (D) ELISA assay of serum IL-12p40 and IFN-γ level at days 0, 7 and 12 post infection of wild type and CD36^−/− mice. n≥5 per group. (E) Il10 mRNA expression (relative to naïve controls) in the spleen of wild type and CD36^−/− mice 7 and 12 days after infection. N=7.

CD36 deficiency causes a protracted tissue stress response immediately following the acute phase T. gondii infection.

To search for correlates of enhanced mortality in CD36^−/− mice, we compared pooled serum samples from infected WT and CD36^−/− mice using a commercial Proteome Profiler Mouse XL Cytokine Array purchased from R&D. Two serum markers elevated in CD36^−/− serum were GDF-15 and FGF-21 (data not shown). Both factors have been identified as biomarkers of tissue injury in the heart, skeletal muscle, kidneys and liver and have proven useful in predicting adverse clinical outcomes(13–15). GDF-15 induces anorexia by binding to the GFRAL receptor expressed by neurons in the area postrema and nucleus solitaris regions of the hindbrain (16). FGF21 is a protein hormone that exerts paracrine and endocrine controls of energy homeostasis and adiposity (17). The regulation of FGF21 is complex, but it is often induced in multiple tissues by nutrient starvation and ER-stress (18). We next examined the kinetics of these two tissue stress-cytokines during T. gondii infection. As shown in Figure 5A, both GDF-15 and FGF-21 levels become elevated to the same extent in WT and CD36 ^−/− mice during the day 7 time point. However, on day 12 post infection, these stress-cytokine remain elevated in CD36^−/− mice while levels in WT mice approach normality. The observation of continually elevated stress cytokines in CD36−/− mice suggested a critical role for CD36 in the re-establishment of tissue homeostasis after the acute response to infection. Because CD36^−/− mice exhibited heightened NK cell IFNγ response (Figure 3), we tested the effect of transient IFNγ neutralization on cytokine stress levels. Administration of anti-IFNγ on days 10 and 11 resulted in decreased serum GDF-15 and FGF-21 levels on day 12 and importantly prevented weight loss in T. gondii-infected CD36−/− mice (Figure 5B and C). Overall, these results indicate that CD36 is critical for re-establishing whole body homeostasis immediately following the acute phase of infection, through modulation of the Th1-induced tissue stress response.

Figure 5.

CD36 deficiency results in increased and prolonged serum stress cytokine response, which is blunted by IFN-γ neutralization. (A) GDF15 and FGF21 concentrations in serum of wild type and CD36^−/− mice 0, 7, and 12 days after infection. N≥3 per group; (B) Serum GDF15 and FGF21 concentration in CD36^−/− mice 12 days post infection. (C) Percent weight change at day 12 relative to day 10 (B). N = 4.

In this study, we have identified CD36 as the macrophage-expressed receptor selectively engaged by avirulent strains of T. gondii. CD36 is a multifunctional scavenger receptor playing wide-ranging functions in health and during infectious and non-infectious pathogenesis(19). Direct binding of recombinant CD36 to the extracellular tachyzoite form of the parasite in a cell-free system and the complete absence of binding to the virulent RH strain supports our hypothesis that avirulent strains display a specific ligand for CD36 on the cell surface, which is absent or hidden on the surface of virulent RH parasites. While the molecular identity of the putative CD36 ligand remains unknown, the observation that deletion of the virulence factor ROP18 in the RH background results in acquisition of CD36 binding sites suggests that the RH strain is inherently endowed with the biosynthetic machinery for production of the putative CD36-ligand. Future work is needed to identify the CD36 ligand and to understand how its biosynthesis, assembly and/or surface expression is suppressed by ROP18.

The absence of CD36 has marked impacts that are evident immediately and over a longer period of time following in vivo T. gondii infection. CD36 deficiency decreases infectivity in the bone-marrow derived small macrophage compartment, which might later affect how the parasite disseminates into both lymphoid and non-lymphoid tissues. Rather than causing an immunodeficiency phenotype, the absence of CD36 preserves and sometimes enhances the magnitude of the type 1 response to infection. CD36 deficiency was associated with increased weight loss and decreased host survival, which correlated with the sustained levels of circulating tissue stress hormones following the acute phase of infection. Interestingly, circulating IFNγ levels were not elevated in CD36-deficient mice on day 12, suggesting local tissue regulation of stress hormone production/release by IFNγ. Thus, CD36 is critical for host tolerance mechanisms that operate to re-establish tissue homeostasis. In sum, our study advances new models to explain the metastability of host-pathogen relationships. In two mutually non-exclusive scenarios, a breakdown in host-pathogen survivability can result from virulence factor mediated antagonism of immune effectors (ie., resistance) or from a failure to deploy avirulence factor(s) that engage or induce host tolerance. Interestingly, in the Toxoplasma system, the virulence factor ROP18 performs two distinct functions; 1) disabling host resistance mechanisms and 2) suppressing surface expression of an unidentified avirulence factor that acts as a parasite ligand for CD36.

Highlights:

CD36 selectively binds to avirulent T. gondii and mediates macrophage tropism.

CD36 is not required for Th1 immune resistance, but is vital for tissue tolerance.

The ROP-18 virulence factor negatively regulates CD36 binding to Toxoplasma gondii.