P2X₇ Receptor-Mediated Killing of an Intracellular Parasite, Toxoplasma gondii, by Human and Murine Macrophages Free

The purinergic P2X₇R functions as a proinflammatory receptor in cells of the monocyte/macrophage lineage. P2X₇R cell surface membrane expression is upregulated by IFN-γ, and the receptor is activated by extracellular ATP released from a variety of cellular sources, including platelets and damaged cells (1). Activation of monocyte and macrophage P2X₇R has been shown to kill intracellular Mycobacterium (2–11), Chlamydia (12–14), and Leishmania species (15). The P2X₇R gene (P2RX7) is highly polymorphic, and a number of non-synonymous single nucleotide polymorphisms (SNPs) have been described that alter receptor function. The majority of SNPs, including 946G>A (rs28360457, NM_002562.4:c.920G>A, Arg³⁰⁷ to Gln), 1096C>G (rs2230911, NM_002562.4:c.1070C>G, Thr³⁵⁷ to Ser), Gln Arg 1513A>C (rs3751143, NM_002562.4:c.1487A>C, Glu⁴⁹⁶ to Ala), and 1729T>A (rs1653624, NM_002562.4:c.1703A>T, Ile⁵⁶⁸ to Asn) confer a loss-of-function phenotype. Significantly, the loss-of-function 1513A>C SNP reduces the in vitro ability of human macrophages to control Mycobacterium tuberculosis (10). Furthermore, inheritance of the 1513A>C SNP has been associated with susceptibility to extrapulmonary tuberculosis in humans (11, 16).

Similar to Mycobacterium, Chlamydia, and Leishmania spp., the apicomplexan parasite, Toxoplasma gondii, is able to infect and survive in cells of the monocyte/macrophage lineage. T. gondii is an obligate intracellular protozoa that infects approximately one-third of humans worldwide (17). Human infection occurs after ingestion of either tissue cysts in raw/undercooked meat or oocysts from infected cat feces, and can be acquired congenitally following primary maternal infection. Although severe disease can occur in the immunocompetent human host, infection is usually asymptomatic or a mild illness characterized by malaise, lympha-denopathy, fever, and headache (17). Considerable morbidity and mortality in immunosuppressed individuals, in particular toxo-plasmic encephalitis, is caused by reactivation of chronic in-fection, and severe fetal abnormalities can occur in association with primary maternal infection (17).

In this study, we describe three immunocompetent people with toxoplasmosis who prompted further investigation of P2X₇R as a factor influencing host response to T. gondii infection. We show, in studies using human macrophages and P2X₇R knockout (P2X₇R^−/−) mice, that P2X₇R activation by extracellular ATP kills T. gondii parasites in infected cells. P2X₇R-mediated T. gondii death occurs in parallel with host–cell apoptosis and is independent of NO production.

Subjects from the Nepean Hospital (Penrith, New South Wales, Australia) provided informed consent for study of their PBMCs. The experimental protocol was approved by the Sydney West Area Health Service, the University of Sydney Human Ethics Committees, and the University of Technology, Sydney, Human Research Ethics Committee, with approval code UTS HREC 2004-077A. Peripheral blood was collected, mononuclear cells were separated, and macrophages were generated and cultured, as described previously (9). Ethidium influx after activation of cells with 1 or 3 mM ATP was measured by flow cytometry (7), and P2RX7 genotyping was per-formed, as described previously (9). Peripheral blood samples provided the DNA obtained from participants in the National Collaborative Chicago-Based Congenital Toxoplasmosis Study (NCCCTS) (18). Geno-typing of these samples was performed using TaqMan technology (19) for P2RX7 SNPs at rs28360457, rs1718119 (NM_002562.4:c.1042G>A), rs2230911, rs2230912 (NM_002562.4:c.1379A>G), rs3751143, rs1653624, and rs1621388 (NM_002562.4:c.1746G>A) (National Center for Biotechnology Information Entrez SNP, www.ncbi.nlm.nih.gov/sites/entrez). Ethical approval for the NCCCTS was obtained from the Institutional Review Boards of the University of Chicago and Michael Reese Hospital and Medical Center, and oversight was provided by an Internal Data Safety Monitoring Committee, the Data Safety Monitoring Board, and the National Institutes of Health (Bethesda, MD).

The immortalized mouse macrophage-like cell line, RAW 264.7, was cultured, as described previously (20). All animal research was approved and conducted in accordance with the University of Technology, Sydney, Animal Care and Ethics Committee, with approval code UTS ACEC 2008-30. Murine bone marrow was isolated from BALB/c (Animal Resource Centre, Murdoch, Western Australia, Australia), C57BL/6J (Animal Resource Centre), and mice on a C57BL/6J background with the P2RX7 gene deleted (P2X₇R^−/− mice, originally obtained from Pfizer, Ann Arbor, MI, and subsequently bred at the Ernst Facility, University of Technology, Sydney, Australia), and bone marrow macrophages were cultured, as described previously (21).

Three assays were used to assess viability of T. gondii in vitro, ensuring a robust assessment of effect of activation of cells by ATP (1 or 3 mM) on survival and replication of tachyzoites.

YFP-YFP RH T. gondii (a gift of B. Striepen, University of Georgia, Athens, GA) replication in cultured monocytes/macrophages was mon-itored, as described previously (22), at the University of Chicago Cellular Screening Centre using an F3 robot and Acumen eX3 microplate cytometer. Hourly measurements of fluorescent parasites were conducted for a total period of 24 h, with three initial measurements taken prior to the addition of ATP to half of the cell samples for each person immediately prior to the fourth measurement.

In addition, a flow cytometry-based viability assay for intracellular T. gondii tachyzoites was established that allowed rapid assessment of viability of thousands of intracellular T. gondii tachyzoites. Cells were infected with T. gondii tachyzoites and incubated overnight, and ex-tracellular tachyzoites were removed. Intracellular tachyzoites were re-leased by mechanically lysing host cells and parasite viability assessed by flow cytometry. Prior to experiments, anti-p30 T. gondii mAb was first conjugated to Alexa Fluor 647 using the Alexa Fluor 647 protein labeling kit, according to the manufacturer’s instructions. Ab was combined with Alexa Fluor 647 dye, incubated at room temperature, and separated from unbound dye by size exclusion chromatography. Tachyzoite viability was assessed using flow cytometry by staining the cell suspension with anti-p30 T. gondii-Alexa Fluor 647-conjugated mAb diluted 1/100 and Sytox Green diluted 1/10,000. Previously prepared 97 ml tachyzoite suspensions were combined with 1 ml anti-p30 T. gondii-Alexa Fluor 647-conjugated mAb and 2 ml Sytox Green working solution (diluted 1/200) and incubated at room temperature for 30 min protected from light, mixing the suspension gently after 15 min. Following incubation, 500 ml PBS plus 1% BSA and 0.05% NaN₃ was added, and the suspension was transferred to a FACS tube for analysis on a BD FACSCalibur flow cytometer. Tachyzoites were identified by log increases in anti-p30 T. gondii-Alexa Fluor 647-conjugated mAb fluorescence, measured on the FL-4 detector. Tachyzoite viability was analyzed based on Sytox Green uptake, measured on the FL-1 detector. Following the acquisition of 5000 gated tachyzoite events, listmode data files were removed from the FACSCalibur Macintosh computer using a portable USB storage device and transferred to a PC for analysis with the free flow cytometry analysis software, WinMDI, version 2.9 (http://facs.scripps.edu/software.html). Tachyzoites were gated for viability analysis, with the resulting bimodal populations corresponding to viable (Sytox Green nonfluorescent) and nonviable (Sytox Green fluorescent) tachyzoites. Percentage of viable T. gondii tachyzoites was plotted using GraphPad Prism, version 5.00 (GraphPad, San Diego, CA; www.graphpad.com).

For the RAW 264.7 line, cells were seeded into Labtek II eight-well chamber slides (Nunc, Naperville, IL) at a density of 5 × 10⁴/well and left to grow overnight. The following day, 1 × 10⁵ freshly lysed T. gondii tachyzoites were added to appropriate wells and left to invade for 2 h at 37°C in 5% CO₂. Uninvaded parasites were removed, and then the infected cells were treated with ATP (pH 7.4) for 1 h at 37°C in 5% CO₂. The media were carefully removed, and the chamber was discarded. Determination of viability was undertaken using acridine orange and ethidium bromide, as described previously (23–25). Cells and parasites were viewed on an Olympus BX51 fluorescent microscope with excitation filter 470/20 nm. A minimum of 300 cells or intracellular parasites was counted per sample, and the assay was repeated in duplicate on at least three separate occasions. Images were taken using an Olympus DP70 digital camera at ×1000 magnification.

Eight-week-old male mice were infected by i.p. injection of 500 tachyzoites of T. gondii RH or ME49 strain. Splenic parasite burdens were determined for individual mice using a modification of the method described previously (26). Briefly, spleens were removed and placed into RPMI 1640 containing 5% FCS. Spleens were weighed and single-cell suspensions were made by passing spleens through a 70-μm sieve. Cells were pelleted at 1500 × g, and then resuspended in 4 ml RPMI 1640 containing 5% FCS. One hundred microliters was added to the first well of a 96-well plate, and splenocytes were serially diluted 1/2 across the plate. Plates were incubated at 37°C in 5% CO₂ for 8 d before wells were examined for the presence of parasites. Parasite burden was determined as the last well in which a single parasite was visible, and the number of parasites per gram was calculated as follows: (mean of reciprocal titers from each duplicate/weight of homogenized spleen) × 400, where 400 is the reciprocal fraction of the homogenized spleen inoculated into the first sample well. Serum was collected from the same mice and assayed for NO using the Griess assay, as described previously (27).

RAW 264.7 cells were seeded into six-well plates containing drop-in Teflon cups (Savillex, Minnetonka, MN) at a density of 1 × 10⁶ cells/well. T. gondii tachyzoites (1 × 10⁶) were infected into each well overnight, prior to the addition of ATP to activate P2X₇ receptors. Apoptosis was assessed after addition of ATP using the annexin V FITC apoptosis detection kit (Calbiochem, Merck, Darmstadt, Germany), according to the manufacturer’s instructions. Annexin V FITC and propidium iodide fluorescence were assessed using a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA).

Subject 1 (S1), a 14-y-old male, presented to the Nepean Hospital with a 2-y history of fatigue, lethargy, and generalized painless lymphadenopathy. There was no resolution of symptoms after multiple courses of antibiotic therapy. T. gondii serology was positive for IgM and negative for IgG. A biopsy of an enlarged left axillary lymph node showed lymphadenitis consistent with toxoplasmosis. Symptoms resolved after treatment with sulfadiazine and pyrimethamine, and convalescent serology was positive for Toxoplasma IgG and negative for IgM. Subject 2 (S2), a 20-y-old female, presented with an enlarged submandibular lymph node 5 wk following a dental extraction. Excision lymph node biopsy and histology were consistent with T. gondii lymphadenitis, and T. gondii serology was positive for IgM and IgG. Repeat serology 2 y later was positive for IgG and negative for IgM. Subject 3 (S3), a 24-y-old pregnant female, had a routine fetal ultrasound scan at 18 wk gestation, which showed borderline fetal cerebral ventriculomegaly. A repeat scan at 22 wk gestation showed prominent cerebral ventriculomegaly. A diagnostic amniocentesis was performed, and PCR for T. gondii DNA was positive.

In all three subjects, there was no evidence of any major underlying immunodeficiency: HIV serology was negative; absolute lymphocyte numbers and lymphocyte subsets were normal; absolute neutrophil counts were normal; and serum Ig levels were normal (Fig. 1A). However, all three people displayed impaired macrophage P2X₇R function measured by ATP-induced ethidium bromide uptake that was either reduced to approximately half (S1) or completely absent (S2 and S3) compared with control people (Fig. 1B). Moreover, P2RX7 genotyping showed S1 was heterozygous for 1513A>C, S2 was heterozygous for both 946G>A and 1096C>G, and S3 was heterozygous for both 946G>A and 1513A>C, loss-of-function SNPs (Fig. 1C).

Genotyping of the NCCCTS cohort (Fig. 2A) identified people with the most common loss-of-function polymorphism in P2RX7, the 1513A>C polymorphism (28), as well as those with wild-type P2RX7 or polymorphisms in linkage disequilibrium (1068G>A and 1772G>A) that do not reduce the function of the receptor (28). A reduction in parasite burden was observed following ATP treatment of monocyte-derived macrophages from people with wild-type P2RX7 or with no loss-of-function polymorphisms (Fig. 2B). Conversely, ATP treatment of monocyte-derived macrophages cultured from people who are homozygous for the 1513A>C loss-of-function polymorphism had minimal effect on the number of intracellular T. gondii tachyzoites (Fig. 2B).

The number of samples and the quantity of cells from people in the NCCCTS cohort with polymorphisms in their P2X₇R were relatively limited. Therefore, P2X₇R^−/− mice on a C57BL/6J background were also used to more definitively assess the ability of the P2X₇R to mediate killing of T. gondii tachyzoites by macrophages. C57BL/6J mice possess a proline to leucine polymorphism at aa 451 in the C-terminal tail of the P2X₇R (29). This polymorphism has variable effects on P2X₇R function (29–34), so we included comparative analyses of BALB/c cells in our investigations because this strain of mouse is known to possess fully functional P2X₇R (29–34) and is also known to be more resistant to T. gondii than C57BL/6J mice (35). P2X₇R function of macrophages from BALB/c, C57BL/6J, and P2X₇R^−/− mice was determined, confirming 100%, 50%, and zero P2X₇R-dependent pore opening, respectively (Supplemental Fig. 1). ATP treatment of macrophages from BALB/c and C57BL/6J mice resulted in a marked reduction in viability of T. gondii RH strain. In contrast, ATP treatment of P2X₇R^−/− murine macrophages produced no significant loss of T. gondii viability (Fig. 3). Very similar results were seen with the relatively avirulent, type 2 ME49 strain of T. gondii; exposure of BALB/c macrophages to ATP reduced parasite viability from 81 ± 4% to 49 ± 12%, in C57BL/6J macrophages, from 83 ± 2% to 49 ± 12%, but, in P2X₇R KO mice, parasite viability remained unchanged at 82 ± 4% (versus 83 ± 3% in nonactivated cells; results are mean ± SE, n = 3).

There were significant differences in the number of parasites recovered from the spleens of ME49-infected mice: on day 12 postinfection, for BALB/c mice, 1693 ± 238 parasites per gram spleen were recovered versus 2353 ± 500 for C57BL/6J mice and 2892 ± 298 for P2X₇R^−/− mice (results are means ± SE, n = 6, and there were significantly more parasites recovered from the P2X₇R^−/− mice, p < 0.05, Mann-Whitney nonparametric test).

To test whether killing of T. gondii after P2X₇R activation is dependent on NO production, we compared serum NO levels from P2X₇R^−/− mice with BALB/c and C57BL/6J mice postinfection by i.p. injection of T. gondii RH strain tachyzoites. There was a trend toward higher serum NO levels in the P2X₇R-deficient infected mice, although none of the differences achieved statistical signi-ficance; thus, control of T. gondii RH strain via the P2X₇R is independent of NO production (Fig. 4).

RAW 264.7 cells express functional P2X₇R (Supplemental Fig. 2). Exposure to ATP induced rapid apoptosis of T. gondii-infected RAW 264.7 cells; more than 70% of cells were positive for annexin V by 1 h (Fig. 5A). At 5 h postexposure to ATP, a significant percentage (>30%) of apoptotic cells had progressed toward lysis (Fig. 5A). Similar levels of ATP-induced apoptosis and necrosis were seen in uninfected RAW 264.7 cells (data not shown).

In situ staining of RAW 264.7 cells with acridine orange and ethidium bromide confirmed the annexin V/propidium iodide flow cytometry results. Acridine orange is able to penetrate all cells and stains nucleic acids green, whereas ethidium bromide, which binds to nucleic acids and fluoresces orange, is only able to enter cells once the integrity of the cell membrane is compromised. Thus, live cells have normal nuclear staining with green chromatin apparent in organized nuclear structures, whereas nonviable cells exhibit orange staining of nuclear material and/or condensed or fragmented chromatin (23–25). Cells were viewed and counted on a fluorescence microscope, and the viability of RAW 264.7 cells was assessed according to these criteria (Fig. 5C). Thus, 54% of RAW 264.7 cells were nonviable after 1 h and 86% nonviable within 2 h of exposure to ATP, compared with 4 and 7%, respectively, for untreated cells. Furthermore, at 2 h postexposure to ATP, intensely stained, condensed, fragmented areas of chromatin were evident within the nuclei of host cells (Fig. 5B, panel 3). At later time points, microscopic examination revealed extensive cell lysis.

In situ staining with acridine orange and ethidium bromide also showed that loss of viability of intracellular tachyzoites of T. gondii occurred in parallel to host cell apoptosis. Thus, 5 min postexposure to ATP, almost all tachyzoites were stained green with acridine orange (Fig. 5B, panel 1) and, after 1 h, the parasites were still predominantly (73%) acridine orange positive and ethidium bromide negative (Fig. 5B, panel 2; Fig. 5C). However, by 2 h postexposure to ATP, 43% of the parasites were stained orange with ethidium bromide (Fig. 5B, panel 3; Fig. 5C). Meanwhile, tachyzoites exposed to ATP in cell-free media remained 100% viable, even after 2 h (data not shown). This demonstrates that a direct toxic effect of exogenous ATP on T. gondii is unlikely; rather, ATP acts via effects on the host cell. By 24 h postexposure to 1 mM ATP, flow cytometric assessment of parasite viability showed that only 25 ± 5% (mean ± SE, n = 4) of T. gondii tachyzoites were viable, a result comparable with the effect of 3 mM sodium nitroprusside (28 ± 5% viable, mean ± SE, n = 4), a known toxin for T. gondii tachyzoites (36). This viability was statistically significantly different (p < 0.001, general linear model two-factor ANOVA with Tukey’s post hoc test) from controls at the same time point where 70 ± 8% (mean ± SE, n = 4) of tachyzoites in control cultures were viable, a level not statistically different from the number of viable parasites after 2, 8, or 16 h of normal in vitro culture (data not shown). The effects on intracellular T. gondii of exposure to ATP could be totally reversed by pretreatment of RAW264.7 cells with oxidized ATP (data not shown), a potent antagonist of the murine P2X₇R (37), further aiding in confirmation of the specificity of the effect.

P2X₇R genotyping and relative P2X₇R activity of three acute toxoplasmosis patients from Nepean Hospital (Fig. 1) are consistent with previous studies showing 50% reduction in monocyte P2X₇R function in 1513A>C heterozygotes and complete loss of function for compound heterozygotes (8, 9, 11). To follow up the surprising association between P2X₇R function and acute symptoms in these patients, and to study the effect of polymorphisms at P2RX7 on function in more detail, we accessed cryopreserved cells from NCCCTS individuals of known genotypes/haplotypes (20). We demonstrated that ATP-dependent killing of tachyzoites of T. gondii in vitro was virtually non-existent in people with the 1513A>C polymorphism, but readily observable in macrophages from people with wild-type receptors or in cells from people with polymorphisms that do not cause loss of receptor function (Fig. 2).

The number of cells available from NCCCTS subjects was limited and did not allow us to perform additional, confirmatory studies with antagonists of the P2X₇R. Therefore, we compared parasite killing using macrophages from P2X₇R^−/− mice with BALB/c mice and C57BL/6J mice to confirm that reduced parasite killing is P2X₇R specific; cells from the knockout mice were totally unable to affect the viability of T. gondii upon activation with ATP (Fig. 3). These data confirm that ATP-dependent macrophage killing of T. gondii is via P2X₇R and does not involve other P2X or P2Y receptors, which are also activated by extracellular ATP (37). Parasite burden in vivo in BALB/c, C57BL/6J, and P2×₇R^−/− mice appears to at least partially confirm that the P2X₇R plays a role in controlling T. gondii. Thus, there were differences in the number of parasites recovered from the spleens of ME49-infected mice, in proportions consistent with their relative P2X₇R function.

It perhaps needs to be noted that ATP-induced killing of T. gondii was comparable in BALB/c and C57BL/6J mice even though the C57BL/6 strain is known to possess a proline to leucine polymorphism at aa 451 that reduces ATP-dependent pore formation by 50% (Supplemental Fig. 1). This could be because of the higher concentration of ATP used in the parasite-killing assay versus the pore-opening assay (3 mM versus 1 mM). However, it is also important to recognize that this particular polymorphism has quite variable effects on P2X₇R function depending on the cell type and activity being examined. Thus, for example, it reduces ATP-dependent pore formation (29), impairs cell death in thymocytes (30), inhibits ATP-induced IL-2 production by splenocytes (31), and affects intercellular calcium waves in astrocytes (32). In contrast, this polymorphism has no effect on phospholipase D activation (30), no effect on expression of the P2X₇R or sensitivity to ATP-induced cell death in splenic CD4⁺CD25⁺ T cells (33), and no effect on P2X₇R-mediated calcium influx in bone marrow-derived macrophages or splenocytes (34).

Activation of the P2X₇R initiates a cascade of intracellular events, including activation of NF-κB (38), phospholipase D (39), and metalloproteases (40); release of reactive oxygen and nitrogen intermediates (41); and stimulation of caspases, leading to apoptosis (42). Loss-of-function polymorphisms in the P2X₇R are also recognized (43, 44) to have a negative effect on the activation of the inflammasome, a complex of cytosolic proteins that regulates caspase-1 activation and, therefore, the processing of IL-1β and IL-18 from inactive to active forms. It is known that killing of mycobacteria and Leishmania amazonensis via the P2X₇R is independent of NO (3, 6, 15); rather, it is associated with apoptosis of host cells (2, 3, 7, 8, 10, 11, 15), as is P2X₇R-mediated killing of Chlamydia (12). We confirmed that P2X₇R-mediated killing of T. gondii is also not associated with NO levels (Fig. 4). To test whether T. gondii killing was associated with apoptosis, we studied in vitro parasite death in RAW 264.7 cells after induction of apoptosis with ATP. Both flow cytometry and in situ staining with acridine orange and ethidium bromide enabled us to document that loss of viability of intracellular tachyzoites of T. gondii occurred in parallel to host cell apoptosis (Fig. 5).

The vulnerability of T. gondii to ATP-induced, P2X₇R-mediated killing shares much in common with similar phenomena observed with Mycobacteria (2–11), Chlamydia (12–14), and Leishmania (15) species, not least the association with apoptosis and independence from NO generation. However, there are likely to be differences too; P2X₇R-dependent killing of Mycobacteria and Chlamydia species is actually dependent on phospholipase D, which is associated with both apoptosis and phagosome-lysosome fusion (4–6, 13). This scenario seems less likely for T. gondii as the parasite inhibits phagosome-lysosome fusion after active penetration of the host cell and modification of phagosomal membrane proteins (45–47). A role for caspase-1 and subsequent release of mature IL-1β, which can cause apoptosis in surrounding cells (37), can also be ruled out because RAW 264.7 cells lack the key adaptor protein, apoptosis-associated specklike protein containing a C-terminal caspase-activating recruiting domain, to form the inflammasome (48). Whereas the limitations of the correlative link between host cell apoptosis and parasite killing must be acknowledged, we believe that, on balance, this effect is likely to be P2X₇R specific and does not imply that just any proapoptotic agent will cause killing of this parasite. This is because of the extremely well-demonstrated ability of T. gondii to inhibit apoptosis induced by a remarkable spectrum of extrinsic and intrinsic proapoptotic stimuli in a variety of host cell types (49–55). It is also worth noting that apoptosis induced by CD95 ligation or by H₂O₂ or necrosis induced by complement does not directly influence the survival of mycobacteria within macrophages (3). Thus, although we have shown that host cell apoptosis is a potential effector mechanism of parasite killing via P2X₇R activation, the precise mechanism underpinning this remains to be definitively determined.

Our observations suggest that inheritance of SNPs that reduce P2X₇R function might cause a defective response to T. gondii infection, higher risk of reactivation in immunocompromised people, and more severe congenital toxoplasmosis. Genotyping P2RX7 in these patient groups and identifying those at risk would allow for more intensive monitoring for reactivation of infection and, if required, suppressive treatment using anti-Toxoplasma therapy.

Disclosures The authors have no financial conflicts of interest.