The IL-12 Response of Primary Human Dendritic Cells and Monocytes to Toxoplasma gondii Is Stimulated by Phagocytosis of Live Parasites Rather Than Host Cell Invasion Free

Toxoplasma gondii is a highly successful obligate intracellular protozoan parasite capable of infecting virtually any warm-blooded species and an important pathogen of humans. Toxoplasma is typically acquired by consuming undercooked meat of infected livestock or by ingesting food or water contaminated with oocysts shed from felines, the definitive hosts of the parasite (1). Although most acute infections are asymptomatic, a fraction of infected immunocompetent individuals develops significant morbidities such as ocular or congenital toxoplasmosis, whereas life-threatening conditions such as cerebral toxoplasmosis can occur in immunocompromised patients (2). Nevertheless, humans are thought to be accidental, end-stage hosts, and the natural transmission cycle is believed to depend largely on predation of small animals, in particular rodents, by cats (3).

Because mice are natural intermediate hosts of T. gondii, they provide a biologically relevant and convenient model for elucidating the cellular and molecular immune mechanisms governing the host–pathogen interaction. Early studies established key roles for IFN-γ (4) and IL-12 (5) in murine host resistance to T. gondii and identified dendritic cells (DC) as a major source of the latter cytokine. CD8α⁺ DC were characterized as the principal IL-12–producing subset in the spleen (6, 7), although CD11b⁺ CD8α⁻ DC were later demonstrated to produce IL-12 transiently at the site of infection (8). IL-12, in turn, was shown to be critical for the generation of IFN-γ–producing NK cells as well as the CD4 and CD8 T cells required for adaptive immunity to the parasite (5). Interestingly, parasites belonging to the type I and II T. gondii genetic lineages, which in mice are virulent and avirulent, respectively, differed in their capacity to trigger IL-12 production in vivo (9).

A major question has concerned the mechanism by which infective T. gondii tachyzoites stimulate the production of IL-12 as well as other proinflammatory cytokines by myeloid cells such as DC. Toxoplasma tachyzoites actively invade host cells though a process dependent on the parasite’s own actin cytoskeleton where it establishes and resides in a nonfusogenic parasitophorous vacuole (PV) (10). This process enables the parasite to evade the normally toxic host cell phagocytic machinery and propagate. Nevertheless, some parasites are taken up via phagocytosis, where they are rapidly destroyed. In the mouse, recent studies have established that neither host cell invasion nor phagocytosis of the parasites is critical for the IL-12 response (8, 11). Instead, a parasite product present in a soluble extract of tachyzoites (STAg) and later identified as T. gondii profilin (TgPRF) was shown to be sufficient for stimulating high levels of the cytokine from DC (12). Additional studies demonstrated that IL-12 production by T. gondii or TgPRF-stimulated DC is dependent on the TLR/IL-1R adaptor protein MyD88 (13). This requirement in turn was shown to reflect the recognition of profilin by TLR11/12 (12, 14). Consistent with the latter observation, TLR11/12^−/− mice rapidly succumb to T. gondii infection (14). Importantly, however, humans lack functional TLR11 and the entire TLR12 gene (15), yet are still capable of mounting a protective immune response. Thus, the above studies leave open the major question of how human cells detect and respond to T. gondii in the absence of these two receptors as well as which myeloid cell types are involved.

In this study, we have systematically investigated the cytokine response of human myeloid cells to T. gondii using elutriated leukocytes from healthy blood donors as the cellular source. Monocytes and DC are important cytokine-producing cells in human peripheral blood that respond to microbial stimulation. Human monocytes are heterogeneous and consist of three major subsets. The largest subset consists of the classical monocytes, which make up 85% of the total population, express CD14⁺, and are CD16^neg. This subset is the human equivalent of murine inflammatory monocytes (Ly-6C⁺) (16), both of which are known to produce effector cytokines and chemokines and enter into inflamed tissues. The two minor subsets are the intermediate monocytes (IM) and nonclassical monocytes (NC), which both express CD16 and are CD14⁺ and CD14^dim, respectively. The NC subset is the human equivalent of the murine patrolling monocyte subset (Ly-6C^low) that responds to viral pathogen-associated molecular patterns (PAMPs), recruits neutrophils to sites of inflammation, and clears cellular debris from the luminal side of blood vessels (17). The myeloid DC (mDC) in human peripheral blood can also be divided into two major subsets, the CD1c⁺ mDC1 and the CD141 mDC2, which have unique functions and are the human equivalent of murine CD8α⁻ CD11b⁺ DC (18) and CD8α⁺ DC (17), respectively.

We have identified CD16⁺ monocytes and CD1c⁺ DC as the major myeloid cells in human peripheral blood that produce IL-12 and TNF-α in response to stimulation with T. gondii tachyzoites. In addition, we demonstrate that, in contrast to murine DC, human monocytes as well as DC fail to secrete these important innate cytokines when exposed to soluble tachyzoite products. Instead, cytokine production depends on the uptake of live parasites by the host cells. Interestingly, we show that the cytokine response does not require active tachyzoite invasion, but rather depends on parasite phagocytosis and endosomal acidification. Together, these findings reveal fundamental differences in the mechanism by which T. gondii triggers innate cytokine production in the human versus murine host.

Peripheral blood monocytes were obtained from healthy volunteers by counterflow centrifugal elutriation at the National Institutes of Health Blood Bank under Institutional Review Board–approved protocols of both the National Institute of Allergy and Infectious Diseases and the Department of Transfusion Medicine. Of the 49 donors, 79% were males and 21% females; 63% were Caucasian, 26% African American, 8% Hispanic, and 2% Asian. Donor age ranged from 21 to 70 with a median value of 41, whereas seropositivity for CMV and toxoplasma was 42 and 35%, respectively.

Elutriated monocytes were further purified by CD14 MicroBead (Miltenyi Biotec) positive selection to minimize the donor-to-donor variability by contaminating cell populations. Monocyte purity was >95% as assessed by flow cytometry (data not shown).

Tachyzoites of the RH-88 type I, including the mCherry⁺ (19), GFP⁺ and conditional T. gondii profilin KO (TgPRF KO) parasites (20), and ME-49 type II strains were propagated in the human foreskin fibroblast cell line Hs27 (ATCC CRL-1634). Confluent fibroblast cultured in 75-cm² flasks was infected with 10⁶ tachyzoites and incubated for 24 h before changing the media to remove free parasites. After an additional 48- to 72-h culture period, fibroblasts were detached by scrapping and passed through a 25G needle to release tachyzoites. The conditional TgPRF KO RH88 parasites were collected from 72-h fibroblast cultures performed in the presence or absence of 1 μg/ml anhydrotetracycline. The deletion of profilin in TgPRF KO parasites from cultures containing anhydrotetracycline was confirmed by plaque assay (data not shown), reflecting the requirement for this protein in parasite invasion.

The parasites were washed in PBS, filtered through a polycarbonate 3-μm membrane (Whatman), and counted. Their viability, determined by either Trypan Blue Exclusion or by fluorescein diacetate (Sigma-Aldrich), was >98%. In some experiments, tachyzoites were killed either by fixation in a 2% paraformaldehyde (PFA) solution or by heating at 65°C for 15 min.

A soluble tachyzoite fraction (STAg) was prepared from RH88 parasites, as previously described (21). A total extract containing both the soluble and insoluble membrane fractions of the RH88 tachyzoite (referred to as Extract) was prepared by harvesting 2 × 10⁹ tachyzoites in PBS with 5 mM EDTA, which were then subjected to 10 rounds of rapid freeze/thaw/sonicate cycles. For each round, parasites were snap frozen in liquid nitrogen, rapidly thawed in a 56°C water bath, and sonicated in a water bath sonicator (Sonicator Ultrasonic Processor XL, Misonix) at 60% using 0.1-s bursts for a total of 30 s. An aliquot was inspected by light microscopy at original magnification ×40 to ensure that all parasites were completely lysed. Stock preparations of STAg and Extract were endotoxin free based on Endpoint Chromogenic Limulus amebocyte lysate assay (Lonza).

CD14⁺ column-purified monocytes were isolated from elutriated monocytes and differentiated into macrophages or DC, as previously described (22). Briefly, monocyte cultures from a same donor were treated in parallel with human rGM-CSF (800 U/ml; PeproTech) and rIL-4 (1000 U/ml; PeproTech) or with human rM-CSF (500 U/ml; PeproTech) on days 0, 3, and 5, for generation of DC and macrophages, respectively. On day 7, cells from both cultures were harvested and stimulated, as indicated.

Monocyte and DC subsets were FACSort purified after staining elutriated monocyte with Alexa Fluor 488 anti–HLA-DR (L243), PE anti-CD14 (TüK4), allophycocyanin anti-CD16 (3G8), PerCP Cy5.5 anti-CD141 (M80), and allophycocyanin-Cy7 anti-CD1c (L161) on a BD FACSAriaII (BD Biosciences). Neutrophils were excluded from analysis by gating out side scatter (SSC)^high HLA-DR^neg CD16⁺ cells. Monocytes were gated as SSC^low HLA-DR⁺ cells, and subsets were defined by the expression of CD14 and CD16. DC were gated as SSC^low HLA-DR⁺ CD14^negCD16^neg, and subsets were distinguished by the expression of either CD141 or CD1c. Postsort analysis confirmed purities of >98% for each subset (Supplemental Fig. 2).

After RBC lysis, cells in the lymphocyte-elutriated fraction were stained with PE anti-CD56 (HCD56), PE-Cy7 anti-CD16 (3G8), Pacific Blue anti-CD127 (eBioRDR5), allophycocyanin anti-CRTH2 (BM16), Pacific Orange anti-CD45 (HI30), and Lineage mixture: eFluor450 anti-CD3 (OKT3), eFluor450 anti-CD4 (RPA-T4), eFluor450 anti-CD8 (OKT8), eFluor450 anti-CD14 (61D3), Pacific Blue anti-CD11c (3.9), Pacific Blue anti-CD11b (ICRF44), eFluor450 anti-CD19 (HIB19), Pacific Blue anti-FceR1a (AER-37), and NK cells defined as SSC^low, lineage-negative, CRTH2^neg CD127^negCD56⁺ population, including both CD16⁺ and CD16^dim subsets, were purified on a BD FACSAriaII sorter.

All cultures were performed in RPMI 1640 medium (Life Technologies) supplemented with 10% heat-inactivated FCS, 4 mM l-glutamine, 10 mM HEPES, 1 mM sodium pyruvate, 1× MEM nonessential amino acids, 100 U/ml penicillin, 100 μg/ml streptomycin, and 50 μM 2-ME. Experiments were carried out on the day of collection. Monocytes were plated at 1 × 10⁶/ml in 0.5 ml in 48-well tissue culture plates and stimulated with T. gondii tachyzoites at a multiplicity of infection (MOI) of 1:1, STAg (10 μg/ml), extract (10 μg/ml), or recombinant TgPRF (1 μg/ml) (AdipoGen Life Sciences). As a positive control, the cells were stimulated with LPS (100 ng/ml; Escherichia coli O111:B4; InvivoGen), R848 (300 ng/ml; InvivoGen), or both.

CD14⁺ column-purified monocytes (5 × 10⁵), NK cells (5 × 10⁵), or autologous monocyte and NK cell cocultures (1 × 10⁶) were cultured with RH88 tachyzoites (MOI 1:1). In a second set of experiments, NK cells (1 × 10⁶/ml) preactivated with 10 U/ml rIL-2 (PeproTech) for 30 min were stimulated with monocyte culture supernatants (75%). Supernatants generated by stimulating monocytes with medium or RH88 tachyzoites or R848 300 μg/ml for 24 h were centrifuged at 1000 × g for 5 min, aliquoted, and stored at −80°C. Before adding to NK cell cultures, each supernatant was treated with either 2 μg/ml neutralizing anti-p70 IL-12 (clone MAB219; R&D Systems) or an isotype control Ab (clone 11711; R&D Systems) for 30 min. In both experiments, NK cells were cultured for 24 h and IFN-γ was measured by ELISA.

Cell culture supernatants were collected after 18–24 h of incubation and stored at −80°C. TNF-α, p70 IL-12, p40 IL-12, and IFN-γ were measured by Human TNF-α DuoSet ELISA (R&D Systems), Human IL-12 p70 Quantikine ELISA (R&D Systems), Human p40 IL-12 Direct ELISA kit Novex (Life Technologies), and Human IFN-γ ELISA Ready-SET-Go! (Affymetric eBioscience), respectively. When indicated, supernatants were analyzed for human sCD40L, IFN-γ, IL-10, IL-1β, IL-6, IL-8, TGF-α, TNF-α, vascular endothelial growth factor, epidermal growth factor, Fractaline, G-CSF, GM-CSF, IFN-α2, IL-12p40, IL-12p70, IL-13, IL-1α, IL-1RA, IFN-γ–inducible protein-10 (IP-10), MIP-α, and MIP-1β using a Millliplex MAP Human Cytokine/Chemokine Magnetic Bead Panel Immunoassay (Millipore).

CD14⁺ column-purified monocytes (1 × 10⁶) were cultured in 0.7 ml in 24-well plates with a 0.4-μm pore-size Transwell insert (Corning 3412) containing 1 × 10⁸ RH88 tachyzoites. After 24 h, the Transwell insert was removed, supernatants were collected, and cytokines were measured by ELISA.

Tachyzoites (3 × 10⁶) in 0.5 ml PBS were treated with either 100 μM 2,4′-dibromoacetophenone (23) (Sigma-Aldrich), 3 μM mycalolide B (MYB) (24) (Enzo Life Sciences), or DMSO (Sigma-Aldrich) as the vehicle control. After 30-min incubation at room temperature, tachyzoites were thoroughly washed with PBS and resuspended in culture media. An aliquot of each drug-treated parasite preparation was tested for viability using fluorescein diacetate (Sigma-Aldrich), and the tachyzoites were shown to be ≥95% viable in all conditions. To confirm that each drug inhibited parasite invasion, monocytes infected for 24 h were examined by cytospins, and no intracellular PV were observed in the drug-treated conditions, but were clearly visible in vehicle control cultures.

CD14⁺ column-purified monocytes (2 × 10⁶) were cultured in medium alone, or with R848 (300 ng/ml) or mCherry⁺ RH88 tachyzoites (MOI 1:1) pretreated with DMSO or MYB (3 μM) for 12 h with the last 9 h in the presence of brefeldin A (3 μg/ml). Cells were then washed with PBS containing 2 mM EDTA, stained with Fixable Viability eFluor 780 (Affymetrix eBioscience), and fixed with 2% PFA. After an overnight incubation in permeabilization buffer (Affymetrix eBioscience), cells were stained with FITC-labeled mouse anti-human IL-12 p40/p70 (clone C11.5) Ab for 1 h. After extensive washing, samples were acquired on BD LSRFortessa (BD Biosciences), and data were analyzed with FlowJo 10.0.8 software (FlowJo).

Column-purified or FACS-purified monocytes and DC (1 × 10⁶) were cultured with RH88 tachyzoites (MOI 1:1) in 48-well plates in the presence of either 3 μM Bafilomycin A1 or 5 mM NH₄Cl for 24 h. Culture supernatants were collected and analyzed p40 IL-12 and TNF-α analyzed by ELISA. Drug treatment did not affect host cell viability or the rate of parasite infection (data not shown).

Column-purified monocytes (1 × 10⁶) were cultured with GFP-RH88 (provided by M. E. Grigg, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD) at MOI 1:1 for 3 h in 15-ml polypropylene tubes. Cultures were washed with PBS, fixed with 2% PFA, permeabilized with PBS containing 0.1% Saponin and 5 mM EDTA, and stained with Alexa Fluor 700 anti–LAMP-1 (clone H4A3) and mouse anti–T. gondii GRA7 Ab (provided by I. Coppens, Johns Hopkins University, Baltimore, MD) for 1 h on ice. After extensive washing, cells were additionally stained with a Rhodamine Red-X–donkey anti-mouse Ab (Jackson ImmunoResearch Laboratories) for 1 h on ice, washed, and affixed on glass slides by cytospin (Shandon). Cells were mounted using Prolong Gold Antifade with DAPI (Molecular Probes) and allowed to cure for 24 h at room temperature before imaging on the Confocal Laser Scanning Microscope Leica SP8 (Leica MicroSystems).

Statistical differences between conditions were analyzed by matched, nonparametric Friedman test using GraphPad Prism software version 6.0e (GraphPad), unless otherwise indicated. In all cases, p values are indicated as *p < 0.05, **p < 0.01, ***p < 0.001.

To determine whether primary human monocytes are capable of detecting and responding to T. gondii, we exposed CD14⁺ column-purified elutriated monocytes from healthy blood bank donors to live tachyzoites of two different strains of toxoplasma, RH88 and ME49, which, based on infection outcome in mice, are considered virulent or avirulent, respectively. Both RH88 and ME49 induced highly significant levels of p40 IL-12 p40 and TNF-α from of most donors, and these responses approached the levels induced by LPS (Fig. 1A). Although a high degree of donor-to-donor variation was observed, monocytes from the same donors tested >6 mo apart secreted comparable levels of cytokines when cultured with RH88 (Supplemental Fig. 1A), suggesting that this variation reflects donor-intrinsic differences rather than mere experimental variation. However, the observed variation in cytokine responses between donors could not be attributed to donor age or prior toxoplasma or CMV exposure (Supplemental Fig. 1B–D).

To better define the monocyte response to T. gondii, we extended the analysis to 16 different cytokines, chemokines, and growth factors. Monocytes failed to produce detectable amounts of TGF-α, epidermal growth factor, vascular endothelial growth factor, G-CSF, GM-CSF, IFN-α2, or CX₃CL1 (data not shown) when exposed to either strain of T. gondii. However, they did secrete significant amounts of p70 IL-12 as well as several other proinflammatory cytokines and chemokines (Fig. 1B). Interestingly, the T. gondii–induced cytokine profile was qualitatively and quantitatively distinct from that triggered by LPS, which, consistent with previously published studies (25, 26), was found to be dominated by high levels of IL-1, MIP-1α, IL-6, IL-8, and IL-10.

In the murine model of T. gondii infection, p70 IL-12 plays a protective role by inducing IFN-γ production by NK cells and T cells (27). To determine whether the IL-12 produced by infected human monocytes is sufficient to drive NK cell IFN-γ, we tested the IFN-γ response to RH88 tachyzoites of monocytes alone, NK cells alone, or autologous monocyte plus NK cell cocultures. Only when NK cells were cultured with monocytes were significant amounts of IFN-γ detected (Fig. 1C). We next tested the ability of cell-free supernatants collected from cultures of RH88-infected monocytes to induce IFN-γ NK cells in vitro. We found that the monocyte-conditioned media in itself was sufficient for triggering NK cell IFN-γ production and that the response observed was IL-12 dependent as preincubating the conditioned media with neutralizing anti-p70 IL-12 mAb completely inhibited IFN-γ production (Fig. 1D). Taken together, these results demonstrate that primary human monocytes can detect and respond to T. gondii by producing a specific set of proinflammatory cytokines, including IL-12, which can drive NK cell IFN-γ production.

To test whether other human myeloid populations can respond to T. gondii infection, we first compared the cytokine production of donor-matched monocytes and in vitro derived macrophages and DC. When stimulated with RH88 tachyzoites, monocyte-derived DC, but not macrophages, secreted p40 IL-12 (Fig. 2A). Moreover, the quantity of IL-12 produced was comparable between monocytes and DC. Interestingly, neither the in vitro derived macrophages nor DC secreted TNF-α following stimulation with T. gondii, although significant levels of the cytokine were detected in parallel cultures after stimulation with LPS (Fig. 2A). Together these results indicate that monocytes and in vitro generated DC share the ability to respond to T. gondii by producing IL-12.

To determine whether primary DC are similarly responsive, we FACS purified blood mDC and measured cytokine production following exposure to live T. gondii or LPS. We found that the cytokine profile of blood DC to both stimuli closely resembled that of elutriated monocytes, although blood mDC overall displayed lower levels of IL-1β, MIP-1α, IL-8, IL-6, and IL-10 (Fig. 2B versus Fig. 1B). As in the monocytes, the T. gondii–induced response in mDC was dominated by p70 IL-12, TNF-α, IL-1RA, and IFN-γ–inducible protein-10. These findings demonstrate that, despite their different origins and functions, primary human monocytes and mDC share the ability to detect and respond to T. gondii by producing IL-12 and a signature set of proinflammatory cytokines.

To determine which specific subsets within the monocyte and DC populations are responsible for the observed cytokine response, we cultured equal numbers of FACS-purified monocyte and DC subsets (see Supplemental Fig. 2A for gating strategy) with T. gondii or LPS. Surprisingly, we found that the minor CD16⁺ IM and NC monocyte subsets produce the majority of the TNF-α and IL-12, whereas classical monocyte CD16^neg monocytes are largely nonresponsive (Fig. 3A). Even more striking was the clear dichotomy in TNF-α and IL-12 production between the mDC1 and mDC2 subsets following T. gondii exposure (Fig. 3B). Thus, the CD1c⁺ mDC1 subset produced high amounts of p40 IL-12 and TNF-α after stimulation with tachyzoites, whereas secretion of the same cytokines by the CD141⁺ mDC2 cells was near the limit of detection. The same response pattern was observed when p70 IL-12 was measured in culture supernatants from T. gondii–stimulated monocyte and DC subsets (data not shown). Importantly, all subsets produced TNF-α and IL-12 in response to LPS (Supplemental Fig. 2B), confirming their general capacity to respond to PAMPs. Interestingly, on a per-cell basis, the mDC1 produced more p40 IL-12 than the CD16⁺ monocyte subsets, whereas the latter cells produced more TNF-α. In the same series of experiments, sort-purified neutrophils failed to produce either cytokine (data not shown). Together, these data identify CD16⁺ monocytes and CD1c⁺ DC as the major circulating myeloid cells producing TNF-α and IL-12 in response to T. gondii.

Because soluble tachyzoite Ag (STAg), and specifically TgPRF, are potent inducers of IL-12 in murine cells, we tested whether human monocytes and DC are capable of responding to the same parasite products. Importantly, human monocyte failed to produce cytokines upon stimulation with either STAg, a whole membrane-containing tachyzoite extract (Extract), or recombinant TgPRF, while secreting significant amounts of p40 IL-12 and TNF-α in response to live parasites (Fig. 4A). In parallel experiments, monocyte-derived and peripheral blood DC were also shown to be unresponsive to STAg (data not shown). As a positive control, we demonstrated that murine splenocytes produce p40 IL-12 after stimulation with STAg and Extract (Supplemental Fig. 3A) (6). Additionally, we confirmed that this difference in cytokine production by human monocytes is unrelated to host cell viability (Supplemental Fig. 3B) or kinetic differences in cytokine production (Supplemental Fig. 3D). To test the possibility that the unresponsiveness of human monocytes to STAg could be the result of active inhibition of cytokine secretion, we stimulated monocytes with LPS or tachyzoites and increasing amounts of STAg. We found that STAg failed to inhibit the cytokine response elicited by these two agonists (Supplemental Fig. 3C). In addition to failing to stimulate TNF-α and IL-12 production, STAg was also unable to elicit the production of 16 other cytokines and chemokines measured by multiplex (data not shown). Because lack of direct recognition of TgPRF does not preclude the possibility that the activity of this protein is relevant in the context of live parasites, we tested the responsiveness of human monocytes to TgPRF KO tachyzoites. As shown in Fig. 4B, the cytokine responses induced by live TgPRF KO tachyzoites were indistinguishable from those induced by WT parasites.

To determine whether physical contact with live parasites is necessary for cytokine production, we cultured monocytes with freshly isolated tachyzoites either together or separated by a semipermeable membrane in a Transwell assay. Monocytes in contact with T. gondii produced p40 IL-12 and TNF-α, as previously shown, but when separated from the parasite failed to respond (Fig. 4C). To establish whether the cytokine response is dependent on live or replicating parasites, we cultured monocytes with live, irradiated, or dead tachyzoites killed by heating at 65°C or PFA fixation.

Both live and irradiated parasites (which are alive but cannot replicate) induced strong cytokine responses. However, heat-killed and PFA-fixed parasites, despite being taken up by human monocytes (data not shown), failed to stimulate cytokine production (Fig. 4D). Together, these data indicate that monocyte cytokine production is dependent on physical contact with live parasites and is not induced or inhibited by STAg.

T. gondii is an obligate intracellular parasite that can actively invade any nucleated cell. Because physical contact with live parasites is required for cytokine production, we investigated the role of parasite invasion in cytokine production by primary human monocytes. To test this, we pretreated parasites with either of two different drugs that block parasite invasion through distinct mechanisms: MYB, an irreversible inhibitor of actin cytoskeleton polymerization (24) and 2,4′-dibromoacetophenone, an irreversible inhibitor of phospholipase A2s that prevents rhoptry protein secretion (23). Interestingly, monocytes secreted cytokines in response to drug-treated parasites, indicating that invasion is dispensable for monocyte cytokine production (Fig. 5A). Moreover, TNF-α levels tended to be higher in the cultures with drug-treated tachyzoites than with the control parasites treated with DMSO alone. To ensure that the drugs blocked invasion, cytospins were prepared from each culture condition 24 h postinfection and examined by light microscopy, and PV were observed only in the control samples containing DMSO-treated tachyzoites (data not shown). Additionally, we failed to observe extracellular parasites in cultures where invasion was inhibited, suggesting that the monocytes might be phagocytizing and degrading the parasites.

To determine whether phagocytosis was indeed occurring, we infected monocyte cultures with GFP-expressing RH88 tachyzoites treated with MYB or DMSO and then examined the cells by confocal microscopy after 3 h (Fig. 5B). Using LAMP-1 as a marker for the phagolysosome and GRA7 as a specific marker for the PV membrane, we were able to distinguish among the GFP-positive monocytes between cells containing PV formed by parasite invasion on the one hand and phagocytized parasites on the other. In monocytes that had successfully been invaded, GRA7 was clearly observed around the GFP signal, and LAMP-1, although present, did not associate with the PV. In contrast, in monocytes that had phagocytized parasites, the GFP signal was present albeit weaker, GRA7 was absent, and LAMP-1 was associated with the GFP-expressing parasites. When analyzed by FACS, the percentage of monocytes containing intracellular GFP⁺ parasites was on average 27 and 12% in cultures exposed to RH88-DMSO and RH88-MYB, respectively (Fig. 5C). Further analysis by confocal microscopy showed that among GFP-positive cells in RH88-DMSO cultures, 16% contained phagocytized parasites, whereas 84% displayed evidence of invasion events (Fig. 5D). In contrast, 98% of GFP-positive monocytes in RH88-MYB cultures acquired the parasites through phagocytosis and only 2% through invasion. Thus, after 3 h in culture, ∼4.5% of cells are phagocytized in the DMSO condition, whereas 11% are in the MYB condition. This 3-fold increase in phagocytic events in monocyte cultures exposed to MYB-treated parasites is proportional to the 2-fold increase in cytokine production observed. Similarly, when intracellular staining was performed, the frequency of IL-12⁺ monocytes was 2-fold higher in cultures exposed to invasion-incompetent than competent parasites (Fig. 5E). Importantly, in both cases, the majority of monocytes producing p40 IL-12 contained intracellular parasites. Together, these data indicate that, although phagocytosis and invasion of tachyzoites both occur in monocyte cultures, parasite invasion is not required for monocyte cytokine production.

To determine whether the phagocytosis of tachyzoites is necessary for eliciting a cytokine response from monocytes, we inhibited phagocytosis by treating the monocytes with MYB and culturing them with invasion-competent or incompetent parasites, as depicted in Fig. 6A, whereas the representative cytospins of corresponding cultures are shown in Fig. 6B. As predicted, when monocytes were capable of phagocytizing parasites, a strong TNF-α and p40 IL-12 response was observed, and even higher levels of the cytokines were detected in cultures with invasion-incompetent parasites. However, when monocyte phagocytosis was inhibited, the cells failed to produce cytokines in response to T. gondii regardless of the parasite’s ability to invade. Nevertheless, the same MYB-treated cells responded normally to LPS stimulation (Fig. 6C).

We next asked whether endosomal acidification, a key event in phagosome maturation, is required for cytokine induction. To do so, we treated monocytes with NH₄Cl (Fig. 6D) or Bafilomycin A1 (Fig. 6E), two agents that inhibit the acidification of endosomes via distinct mechanisms. Monocytes infected in the presence of either drug failed to secrete cytokines, but remained responsive to LPS stimulation. Light-microscopic examination of the cultured cells demonstrated that, at the concentrations used, the drugs failed to inhibit parasite invasion or replication, arguing against an indirect effect on the parasites themselves. Moreover, FACS-purified CD16⁺ NC and IM monocytes and CD1c⁺ DC displayed the same requirements for cytokine production in response to T. gondii (Fig. 7A, 7B). Together, these findings lend further support to the hypothesis that phagocytosis of live parasites is required for the induction of cytokines in human monocytes and DC.

In this study, we have identified CD16⁺ monocytes and CD1c⁺ DC as the major myeloid cell populations in human peripheral blood that produce IL-12 and TNF-α following exposure to T. gondii. Moreover, we demonstrate that, in both subsets, this cytokine response requires the phagocytosis of live parasites and endosomal acidification, which is in stark contrast to the response of murine myeloid cells that is driven primarily by soluble parasite products without the need for host–parasite interaction.

T. gondii infection in humans is fundamentally different from that of mice in that, whereas murine infection is thought to play a major role in the transmission cycle of the parasite, humans are largely an end-stage host for this protozoan (28). Because of this distinction, whereas the mouse is a convenient and biologically relevant host for examining mechanisms of host resistance to T. gondii, it need not directly reflect the mechanisms involved in the human response to this parasite. Indeed, the absence of functional TLR11 and TLR12 in humans predicts the existence of alternative parasite recognition mechanisms that can function in the absence of these major receptors used by mice and other species. Similarly, although DC and inflammatory monocytes appear to be the dominant myeloid cells in the innate response to T. gondii in mice (7, 8, 29), other cell populations could play major roles in the human response to the parasite.

Our findings establish that, in humans, monocytes, in addition to DC, are important sources of IL-12 and TNF-α in the response to T. gondii. Human peripheral blood monocytes have previously been shown to produce IL-1β upon infection with T. gondii (30), consistent with their known constitutive expression of active caspase-1 (31), and this observation was confirmed in the current study. However, we also found that T. gondii–stimulated monocytes and DC simultaneously produce high levels of IL-1RA, which might limit the effects of IL-1.

In the murine host, inflammatory monocytes have been shown to produce IL-12 and TNF-α in the lamina propria (32) and to produce IL-12 in the peritoneum (8) following parasite inoculation. Interestingly, in our experiments examining the monocyte subsets responding to T. gondii, CD14⁺CD16^neg monocytes, which are the human equivalent of murine inflammatory monocytes, failed to secrete significant levels of IL-12 and TNF-α following stimulation with the parasite. At present, it appears that this distinction reflects a genuine human–mouse species difference and not a divergence between circulating versus inflammation-recruited tissue monocytes, because, when examined in murine peripheral blood, inflammatory monocytes were again a major source of IL-12⁺ following STAg stimulation (N. Wu and D. Jankovic, unpublished data). CD14^dimCD16⁺ monocytes are known to play a major role in clearing cellular debris from damaged endothelia cells, recruiting neutrophils to the site of infection (33) and responding to viral PAMPs (34). In addition, the two minor CD16⁺CD14^dim NC and CD14⁺CD16⁺ IM monocyte subsets that produced high levels of IL-12 and TNF-α following T. gondii exposure have previously been identified as major cytokine-producing cells in the response to Aspergillus fumigatus (35) and, in recently published work, to Plasmodium vivax (36). Our findings extend those of latter study (36) in demonstrating that human CD16⁺ monocytes can play a selective role in the cytokine response to another major protozoan pathogen.

In parallel with our observations on the response of monocyte subsets, we showed that IL-12 and TNF-α production following T. gondii stimulation is restricted to only one subset of human myeloid DC (i.e., the CD1c⁺). Interestingly, the nonresponsive CD141⁺ DC population is the human equivalent of the murine CD8α⁺ DC subset, which dominates the IL-12 response of mice to T. gondii (7). Moreover, the CD141⁺ subset appeared to be totally unresponsive to parasite stimulation based on multiplex assay of 16 different cytokines and chemokines (data not shown). Previous studies on the human CD1c⁺ DC have indicated that this subset can produce IL-12 in response to a variety of stimuli (37, 38), suggesting that the discrepancy highlighted above may reflect a genuine human versus mouse difference in the behavior of this cell population.

A major finding of the current study is that the IL-12 response of human monocytes and DC is distinct from that of murine myeloid cells, which are highly responsive to soluble parasite products and in particular the TLR11/12 agonist, TgPRF (12). In addition, TgPRF failed to trigger cytokine response in human monocytes even after priming with IFN-γ (data not shown). Moreover, our data show that TgPRF does not contribute to the induction of cytokine response in human monocytes exposed to live parasites (Fig. 4B). These findings are inconsistent with a recent report by Salazar Gonzalez et al. (39), who showed that TgPRF stimulates human monocytes to secrete IL-6 and IL-12 in a TLR5-dependent manner. The basis of this discrepancy is unclear, but we speculate that it may stem from monocyte purification techniques employed by these authors, which differ from those used in this work.

A major conclusion of the current study is that IL-12 production by human monocytes, although not stimulated by killed parasites, requires the phagocytosis of live tachyzoites. Because both heat-killed and PFA-fixed parasites are actively phagocytized (data not shown), these results argue against phagocytosis as a nonspecific initiator of cytokine secretion in our system. Moreover, killed parasites opsonized with human anti–T. gondii Abs also failed to elicit a cytokine response from primary human monocytes (data not shown). In contrast to the need for phagocytosis of live parasites, active tachyzoite invasion of these cells was found not to be essential for cytokine production. Previous studies with human monocyte-derived DC and THP-1 cells also documented a requirement for live tachyzoites in the activation of these cells by toxoplasma, but did not examine the differential role of invasion versus phagocytosis in this process (40, 41). A requirement for phagocytosis in the induction of IFN-β mRNA has recently been demonstrated (42) in tachyzoite-exposed murine bone marrow inflammatory monocytes; however, the relevance of this finding to the human IL-12 response is complicated by the likely presence of the TLR11/12 pathway in these cells. At a more general level, phagocytosis and endosomal acidification-dependent cytokine production have been shown to be essential features of the response to Gram-positive bacteria, and this requirement was proposed to reflect the need to liberate cryptic pattern recognition receptor (PRR) ligands from the pathogen (43). Such a mechanism could explain why active infection, in which tachyzoites persist in an unacidified PV (44), is insufficient on its own in stimulating a significant cytokine response in human monocytes. Indeed, it is well documented that active infection of myeloid cells can inhibit cytokine responses induced by other stimuli such as TLR agonists (45).

An important question raised by our findings concerns the nature of the parasite PAMPs involved in stimulation of IL-12 and TNF-α in human monocytes and DC. Clearly, in our study, these molecules are distinct from the tachyzoite TgPRF responsible for cytokine induction in murine myeloid cells (12, 14, 46). As discussed above, one hypothesis is that the relevant PAMPs require enzymatic digestion in the phagolysosome to be released in a form that can interact with the relevant PRR either within the endosome itself or perhaps in the cytosol following transfer to that compartment. Although TLR3/7/8/9 are all known to function within endosomes, TLR3 is not expressed by human monocytes and therefore is unlikely to be involved in the response to T. gondii. A role for the TLR nucleic acid–sensing receptors was suggested by the UNC93B1 dependency of host resistance to the parasite in mice and the greater susceptibility of TLR3/7/9/11-deficient versus TLR11-deficient animals to T. gondii infection (46). In addition, the same group demonstrated that human PBMC stimulated with T. gondii RNA and DNA produce IL-12 and TNF-α. However, in contrast to the findings reported in this work, these responses were only observed when the human PBMC were primed with IFN-γ.

As noted above, it is also possible that cytokine production in T. gondii–exposed cells is stimulated by cytosolic receptors that encounter parasite-derived PAMPs as a result of transfer from the phagosome. In the case of exogenous proteins, such a transfer from phagosome to cytosol is known to occur during Ag cross-presentation in DC (47, 48). Release of PAMPs into the cytosol is also thought to occur as a consequence of pathogen or damage-induced leakage of the phagosomal membrane (49–51). This mechanism has been proposed as an explanation for the phenomenon of vita-PAMPs, in which the phagocytosis of live bacteria is required for the induction of NLRP3-dependent cytokine production (52). In that case, evidence was presented that the relevant PAMPs are labile mRNA molecules, thereby explaining the dependence on phagocytic uptake of live organisms, a requirement also observed in our experiments on the induction of cytokines by T. gondii in human cells. In this hypothesis, leakage of toxoplasma-derived nucleic acids into the cytosol would trigger IL-12 and TNF-α via receptors such as RIG-I or MDA5. Experiments are in progress to examine the specific role of parasite-derived nucleic acids as the PAMPs involved in the response of human monocytes and DC and to assess the relative contribution of endosomal versus cytosolic PRR in this process.

The immune response to T. gondii plays a major role in determining the outcome of human infection and disease caused by this protozoan parasite. Because of its important role as an intermediate host in the toxoplasma life cycle, the mouse (and related rodents) may have evolved distinct immunological mechanisms for controlling this parasite (28, 53). Due to the emerging discrepancies between the mouse and human immune responses to T. gondii, studies directly identifying the specific requirements for recognition of toxoplasma by the human host are essential both for understanding pathogenesis and for the design of effective immunologic interventions.

We thank Elina Stregevsky, Bishop Hague, and Thomas Moyer for performing FACS sorting and Olena Kamenyeva and Juraj Kabat for assistance with confocal microscopy. We are also grateful to Drs. Isabelle Coppens (Johns Hopkins University), Michael Grigg (National Institute of Allergy and Infectious Diseases), and Christopher Hunter and David A. Christian (University of Pennsylvania) for providing anti-GRA7 Ab, GFP-RH88, and mCherry-RH88, respectively; Dr. Carl G. Feng for invaluable advice at the start of this study; and Dr. Bruno Andrade for insightful discussions and help with the statistical analysis. In addition, we thank Drs. Daniel S. Green and Roshanak Tolouei-Semnani for helpful discussions and both them and Dr. Katrin Mayer-Barber for critical reading of the manuscript.

The authors have no financial conflicts of interest.