Inducible Nitric Oxide Synthase Is a Key Host Factor for Toxoplasma GRA15-Dependent Disruption of the Gamma Interferon-Induced Antiparasitic Human Response

Toxoplasma, an important intracellular parasite of humans and animals, causes life-threatening toxoplasmosis in immunocompromised individuals. Gamma interferon (IFN-γ) is produced in the host to inhibit the proliferation of this parasite and eventually cause its death. Unlike mouse disease models, which involve well-characterized virulence strategies that are used by Toxoplasma to suppress IFN-γ-dependent immunity, the strategies used by Toxoplasma in humans remain unclear. Here, we show that GRA15, a Toxoplasma effector protein, suppresses the IFN-γ-induced indole-2,3-dioxygenase 1-dependent antiparasite immune response in human cells. Because NLRP3-dependent production of IL-1β and nitric oxide (NO) in Toxoplasma-infected human cells is involved in the GRA15-dependent virulence mechanism, blocking NO or IL-1β production in the host could represent a novel therapeutic approach for treating human toxoplasmosis.

T oxoplasma gondii is an obligatory protozoan parasite that can infect nearly all warm-blooded animals, including humans (1, 2). It is estimated that one-third of the world's human population is infected with T. gondii; notably, most infections are asymptomatic. T. gondii, however, also causes toxoplasmosis in immunocompromised individuals; the clinical signs of toxoplasmosis comprise encephalitis, hepatitis, and myocarditis. Individuals with increased susceptibility to toxoplasmosis include those with AIDS, those undergoing chemotherapy, fetuses with congenital diseases, and newborn babies of women who initially contracted the infection during pregnancy (3)(4)(5). T. gondii is ranked among the top five human pathogens that cause economic loss and life impairment via food-borne illness in the United States (6). Thus, T. gondii is an important pathogen of both humans and animals.
T. gondii secretes various effector molecules into host cells upon infection to promote efficient parasite growth and dissemination in vivo (7,8). The effector mechanisms used by the parasite to subvert host immune responses have been extensively analyzed in mouse models. The proteins ROP5, ROP16, ROP17, ROP18, GRA7, and TgIST are secreted from rhoptries or dense granules to suppress anti-T. gondii cellautonomous immune responses; this results in increased parasite virulence in mice (9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19). GRA6, a dense granule protein, activates the host transcription factor NFAT4 to induce chemokines and recruit neutrophils to sites of infection, thereby promoting parasite dissemination and maximizing parasite virulence (20). GRA15, another dense granule protein, is secreted into host cells to activate another host transcription factor, NF-B, in both mouse and human cells (21)(22)(23). Similarly to GRA6, GRA15 activates host immune responses and mediates interleukin-1 (IL-1) production via activation of the NLRP3 inflammasome (23,24). Lack of GRA15 in T. gondii parasites promotes in vivo parasite proliferation in mice (22). Given that GRA15-deficient T. gondii is more virulent than wild-type (WT) T. gondii in mice, GRA15 might assist host survival by limiting parasite replication; hence, it may play an antiparasitic role in mice (19,22). However, the significance of GRA15 as a virulence factor in humans is not well understood.
The mechanisms underlying host resistance to T. gondii rely on innate and adaptive immunity and involve various immune/nonimmune cells and cytokines. Among these contributing factors, interferon-␥ (IFN-␥), which is the most important host cytokine that targets T. gondii, is largely produced by CD4 ϩ T cells and natural killer cells; it stimulates cell-autonomous responses in immune cells, including macrophages and dendritic cells, or nonimmune cells (e.g., fibroblasts) (25). IFN-␥ activates the STAT1 transcription factor and induces the expression of hundreds of genes (26). IFN-␥inducible GTPases and nitric oxide (NO) mediate parasite clearance and growth inhibition in mice, respectively (27); however, they may not play major roles in these processes in humans (28)(29)(30). IFN-␥-dependent nutrient deprivation and cell death have been established as anti-T. gondii responses in human cells (31,32). IFN-␥ stimulates the expression of indoleamine 2,3-dioxygenase (IDO) to degrade tryptophan, an essential nutritional amino acid for the intracellular growth of T. gondii in human cells (33,34). Thus, although IFN-␥ is important for responses against T. gondii in both humans and mice, the IFN-␥-inducible effector mechanisms differ greatly between these two hosts.
The anti-T. gondii role for inducible NO synthase (iNOS), an IFN-␥-inducible protein, has been established in mice (35). Deletion or pharmacological inhibition of the iNOS gene in mouse macrophages results in profoundly reduced NO production in response to IFN-␥, along with concomitant parasite growth (28,36). However, blocking iNOS activity does not affect the IFN-␥-induced antiparasite response in human macrophages or monocytes (28). Thus, although iNOS may play different roles in mice and humans, its precise role in the IFN-␥-mediated interplay between humans and T. gondii remains unknown.
We describe here a novel virulence strategy for T. gondii whereby the pathogen utilizes the GRA15 effector protein and the iNOS host cofactor to suppress the IFN-␥induced IDO-dependent cell-autonomous immune response in human cells.
proliferated in a manner similar to that seen with wild-type parasites in THP-1 and Huh7 cells (Fig. S1C). During parasite dissemination in vivo, T. gondii infects CD11b ϩ cells and translocates itself from infected sites to multiple organs (20,37) where T. gondiiinfected monocytes can interact with various tissue cells. To mimic this in vivo monocyte-hepatocyte interaction using an in vitro design, various coculture models have been developed with human immune cell lines and tissue cell lines; one of these models utilizes the THP-1 and Huh7 cell lines (38). Thus, we tested cocultures of THP-1 cells and Huh7 cells, in which THP-1 cells were infected with T. gondii for 24 h; subsequently, infected THP-1 cells and culture supernatants were seeded on Huh7 cells, with or without IFN-␥. These cocultures were incubated for a further 48 h; numbers of parasites were then compared (Fig. 1B). Surprisingly, the GRA15-KO parasite-infected THP-1/Huh7 coculture condition resulted in a significant reduction in the numbers of parasites compared with the wild-type parasite-infected condition (Fig. 1B). To test whether the reduction of the number of GRA15-KO parasites in the THP-1/Huh7 coculture model was due to the lack of GRA15 in the parasites, we complemented GRA15 in the GRA15-KO parasites (Fig. S1E) and then analyzed parasite growth under the coculture conditions (Fig. 1C). The numbers of GRA15-KO parasites complemented with GRA15 significantly increased in the IFN-␥-stimulated THP-1/Huh7 coculture model compared with the numbers of parasites containing empty vector (Fig. 1C). Next, we quantified the parasite levels by measuring the amount of genomic DNA; we compared the growth levels of wild-type and GRA15-KO parasites in THP-1 and Huh7 cells or in the coculture ( Fig. 1D and E; see also Fig. S1F). The numbers of GRA15-KO parasites in IFN-␥-prestimulated THP-1 and Huh7 cells were similar to those seen with the wild-type parasites (Fig. 1D). In contrast, the numbers of GRA15-KO parasites were significantly lower than the numbers of wild-type parasites in cocultures of infected THP-1 and Huh7 cells in the presence of IFN-␥ (Fig. 1E); this is consistent with data from luciferase assays. Collectively, these data suggest that GRA15 has an advantageous effect on T. gondii growth under these human cell line coculture conditions.
Next, we tested whether the proparasitic effect of GRA15 was present in a coculture of primary mouse macrophages and hepatocytes (Fig. 1F). In cocultures of primary mouse cells, the advantageous effect of GRA15 in IFN-␥-stimulated cells was not observed (Fig. 1F), indicating that the proparasitic effect of GRA15 might be absent in mice.
IL-1 signaling in Huh7 cells downregulates the IFN-␥-induced anti-T. gondii response. To investigate which host factors are responsible for the GRA15-dependent proparasitic effect, we first compared levels of cell viability between wild-type and GRA15-KO T. gondii under the THP-1/Huh7 coculture conditions (Fig. S2A). Rates of cell death in coculture with wild-type or GRA15-KO parasites were comparable among the various time points tested (Fig. S2A), indicating that cell viability is not a factor for the GRA15-mediated proparasitic effect. Next, the supernatants from wild-type or GRA15-KO T. gondii-infected THP-1 cultures were collected, filtered to remove parasites or THP-1 cells, and subsequently added to Huh7 cells newly infected with wild-type or GRA15-KO parasites. The numbers of parasites in Huh7 cells were then assessed ( Fig. 2A). We found that the presence of GRA15 in T. gondii parasites that infected the THP-1 cells, but not in those that infected the Huh7 cells, affected the GRA15dependent increase in parasite numbers in the Huh7 cells ( Fig. 2A). THP-1 cell infections with T. gondii show GRA15-dependent and strain-specific IL-1␤ production (23). We confirmed that IL-1␤ production occurred in THP-1 cells infected with wild-type or GRA15-KO parasites complemented with GRA15 but not in THP-1 cells infected with GRA15-KO parasites (Fig. 2B). On the other hand, IL-1␤ production did not occur in Huh7 cells infected with wild-type or GRA15-KO parasites (Fig. S2B). Because the results seen with the supernatant from the THP-1-infected cells explain the observed effect of GRA15 on the cells ( Fig. 2A), we tested whether IL-1␤ is involved in parasite growth in Huh7 cells stimulated with IFN-␥ (Fig. 2C). Stimulation with IFN-␥ strongly suppressed the growth of the wild-type parasite in Huh7 cells (Fig. 2C). In contrast, the addition of IL-1␤ significantly impaired the IFN-␥-induced reduction in parasite numbers (Fig. 2C), whereas the effect of IL-1␤ was not observed when IFN-␥-pretreated Huh7 cells were subsequently infected with T. gondii and then stimulated with IL-1␤ (Fig. S2C). To directly test the effect of IL-1␤ signaling in Huh7 cells on downregulation of the IFN-␥-induced response, we utilized CRISPR/Cas9 genome editing to generate Huh7 cells lacking IL-1R1 or MyD88 ( Fig. S2D and E), both of which are essential for the IL-1 receptor signaling pathway (39); we then examined the IL-1␤-induced impairment of the IFN-␥-induced reduction in parasite numbers in the IL-1R1-KO or MyD88-KO Huh7 cells (Fig. 2C). Notably, IL-1␤-induced impairment of the IFN-␥-mediated response against T. gondii was not observed in either IL-1R1-KO or MyD88-KO Huh7 cells (Fig. 2C). Furthermore, IL-1R1-KO or MyD88-KO Huh7 cells cocultured with the wild-type T. gondii-infected THP-1 cells contained significantly fewer parasites than the wild-type T. gondii-infected Huh7 cells (Fig. 2D). In addition, IL-1␣, as well as IL-1␤, had similar suppressive effects on the IFN-␥-induced reduction of T. gondii numbers in Huh7 cells (Fig. S2F). Collectively, these results suggest that IL-1 negatively regulates the IFN-␥induced anti-T. gondii response in the Huh7 hepatocyte line.
GRA15 indirectly downregulates the IDO1-induced anti-T. gondii cell-autonomous immune response. IFN-␥-induced tryptophan degradation by IDO plays a central role in the IFN-␥-induced anti-T. gondii response in humans (34). Interestingly, we found that IL-1␤ costimulation severely reduced IDO1 protein levels in a MyD88-dependent fashion (Fig. 4A). Because the reported results of previous studies in this area were based on the use of the 1-MT IDO inhibitor, there is no genetic evidence for the involvement of IDO in the anti-T. gondii response. Furthermore, ATG16L1, an autophagy protein, has recently been reported to participate in the IFN-␥-induced antiparasite response in a HeLa cervical carcinoma cell line (41). Therefore, we generated IDO1-deficient (IDO1-KO) and ATG16L1-deficient (ATG16L1-KO) Huh7 cells by using CRISPR/Cas9 genome editing to determine whether IDO or ATG16L1 or both are important for the IFN-␥induced anti-T. gondii response in Huh7 cells ( Fig. S3A and B). Because kynurenine is a tryptophan metabolite of IDO (42), we measured the kynurenine concentration in IDO1-KO Huh7 cells (Fig. S3C). While the kynurenine concentration increased upon IFN-␥ treatment of wild-type Huh7 cells, such an increase was not observed in IDO1-KO cells (Fig. S3C). We then assessed whether an IFN-␥-induced reduction in T. gondii growth and parasite numbers in vacuoles had occurred in the wild-type, ATG16L1-KO, or IDO1-KO cells ( Fig. 4B and C). The wild-type and ATG16L1-KO cells showed similar reductions in parasite numbers in response to IFN-␥. In sharp contrast, impaired IFN-␥-induced reductions in parasite numbers, as well as increased parasite growth in   Fig. 4B and C), suggesting that IDO1, but not ATG16L1, plays an important role in the IFN-␥-induced anti-T. gondii cell-autonomous response. Comparing the levels of IL-1␤-induced impairment of the IFN-␥-dependent response between wild-type and IDO1-KO Huh7 cells, the effect of IL-1␤ was found to have been completely abolished in IDO1-KO cells (Fig. 4D). Consequently, we tested whether the parasite number reduction caused by GRA15 deficiency involves IDO1 (Fig. 4E). We compared IDO1 protein levels in uninfected and wild-type or GRA15-KO parasite-infected THP-1/Huh7 cocultures (Fig. 4E). Although the wild-type parasite-infected condition resulted in low IDO1 protein levels, the levels of IDO1 protein in the GRA15-KO parasite-infected condition were higher than those seen under the wild-type parasite-infected conditions (Fig. 4E). Moreover, the IFN-␥-induced reduction in the numbers of GRA15-KO parasites was abolished when Huh7 cells lacked IDO1 (Fig. 4F), suggesting that the reduction in parasite numbers related to GRA15 deficiency was caused by increased IDO1 protein levels. In addition, we consistently found that IDO1 protein levels in cocultures of wild-type parasite-infected wild-type THP-1 and MyD88-KO Huh7 cells were higher than those seen with wild-type Huh7 cells (Fig. 4G). Furthermore, IDO1 protein levels in cocultures of CASP1-KO or NLRP3-KO, but not NLRP1-KO, THP-1 cells and wild-type Huh7 cells were higher than those seen with the wild-type THP-1 cells (Fig. 4H). Taken together, these data showed that infection of THP-1 cells with GRA15-intact T. gondii produced IL-1␤ in a manner dependent on NLRP3 and caspase-1; this led to an indirect reduction in IDO1 proteins, thereby supporting parasite growth in human cells.
The proparasitic effect of GRA15 in the presence of IFN-␥ was not observed under the mouse macrophage/hepatocyte coculture conditions (Fig. 1F). Next, we tested whether the proparasitic effect of IL-1␤ was present in IFN-␥-activated mouse hepatocytes, in terms of IDO1 ( Fig. 4I and J). Mouse hepatocytes also showed strong IFN-␥mediated suppression of T. gondii growth (Fig. 4I). However, in sharp contrast to the observations in human cells, IL-1␤ stimulation in mouse hepatocytes could not reverse the IFN-␥-induced reduction of T. gondii levels (Fig. 4I). Furthermore, mouse hepatocytes lacking IFN-inducible GTPases, such as Irgm1 and Irgm3 (well-known anti-T. gondii factors in mice) (43), were completely defective with respect to IFN-␥-induced parasite number reduction (Fig. 4I). Comparing the IDO1 protein levels in human and mouse hepatocytes, IDO1 proteins were detected in human cells but not in mouse cells (Fig. 4J). In contrast, Irgm1 proteins were detected only in mouse cells, and not in human cells, in response to IFN-␥ (Fig. 4J); notably, humans do not express Irgm1 but possess IRGM, a single human homolog of IRG whose significance remains unclear with respect to the anti-T. gondii human response (44). To directly examine the role of IRGM in Huh7 cells, we generated IRGM-KO Huh7 cells by CRISPR/Cas9 genome editing and tested for the ability to mount an IFN-␥-induced anti-T. gondii response ( Fig. 4K and L; see also Fig. S3D). IFN-␥-prestimulated IRGM-KO cells reduced T. gondii numbers in a manner similar to that seen with the wild-type cells (Fig. 4K). Furthermore, the IL-1␤mediated increases of parasite numbers and concomitant reductions of IDO1 protein levels in IFN-␥-stimulated IRGM-KO Huh7 cells were similar to those observed in wild-type cells (Fig. 4K and L). Taken together, these results indicate that the IL-1mediated proparasitic effect is not present in mouse hepatocytes, where IFN-inducible GTPases such as Irgm1 and Irgm3 play major roles in the anti-T. gondii system. In addition, human IRGM is not involved in the IFN-␥-induced anti-T. gondii response and the IL-1-mediated proparasitic effect in human cells.
iNOS is essential for the GRA15-dependent reduction in IDO1 protein levels. We attempted to gain further insight into the mechanism by which IL-1␤-mediated suppression of IDO1 protein levels is dependent on MyD88. NO is known to inhibit IDO activity in macrophages (45). Because iNOS is important for IFN-␥-mediated NO production (46), we examined iNOS mRNA expression in Huh7 cells (Fig. 5A). Stimulation with IL-1␣ or IL-1␤ in addition to IFN-␥ strongly induced iNOS mRNA expression and NO production in a MyD88-dependent manner ( Fig. 5A; see also Fig. S4A). Furthermore, THP-1 cells did not express iNOS protein and produce NO in response to IFN-␥ or/and IL-1␤ stimulation ( Fig. S4B and C). Therefore, to assess the role of iNOS in Huh7 cells, we generated iNOS-deficient (iNOS-KO) Huh7 cells by CRISPR/Cas9 genome editing (Fig. S4B). The iNOS-KO Huh7 cells lacked NO production in response to IL-1␤/IFN-␥ ( Fig. S4D and E). Furthermore, IL-1␤-induced reduction of IDO1 protein levels and impaired IFN-␥-dependent parasite reduction were not observed in iNOS-KO cells ( Fig. 5B and C), suggesting that IL-1␤ might stimulate iNOS expression and thereby cause IDO1 inhibition, serving as the GRA15-dependent virulence mechanism. We tested this possibility by using the parasite-infected THP-1/Huh7 coculture model ( Fig. 5D and E). iNOS-KO Huh7 cells cocultured with wild-type parasite-infected wildtype THP-1 cells exhibited a complete loss of NO production but showed no reduction in IDO1 protein levels ( Fig. 5D and E). Additionally, the IFN-␥-mediated reduction in numbers of parasites was enhanced in cocultures of iNOS-KO Huh7 cells, in comparison with wild-type cells (Fig. 5F). When CASP1-KO THP-1 cells, IL-1R1-KO Huh7 cells, MyD88-KO Huh7 cells, or GRA15-KO T. gondii were used in the coculture model, iNOS protein expression and NO production were defective (Fig. 5G to J; see also Fig. S4F), suggesting that IL-1␤-induced NO is not important for the anti-T. gondii response in Huh7 cells. Moreover, complementation of GRA15 in GRA15-KO parasites restored iNOS expression and reduction of IDO1 protein levels in the THP-1/Huh7 coculture model (Fig. 5K). Taken together, these data indicate that iNOS is critical for the GRA15dependent virulence mechanism in the THP-1/Huh7 coculture model.
Blockade of NO production inhibits T. gondii growth in THP-1/Huh7 coculture condition. The essential role of iNOS in the proparasitic effect of GRA15 prompted us to examine whether this step can be targeted by a specific inhibitor. To test whether NO produced by iNOS is critical for the GRA15-mediated reduction in IDO1 levels, we used aminoguanidine, an inhibitor of NO synthase, and calculated the numbers of parasites (47). Addition of aminoguanidine to the coculture reduced the NO concentration and, conversely, increased IDO1 protein levels; in contrast, the iNOS level remained unchanged between the control and aminoguanidine-treated cocultures ( Fig. 6A and B), indicating that aminoguanidine inhibits NO production downstream of iNOS and restores IDO1 protein levels. Consequently, the number of parasites in the coculture system was significantly reduced in the presence of aminoguanidine (Fig. 6C). Although aminoguanidine treatment of the IDO1-KO Huh7 cells also inhibited NO production, the IFN-␥-induced reduction in parasite numbers was unchanged between the aminoguanidine-treated and untreated cells (Fig. 6D and E), suggesting that the NO levels might be insufficient for the suppression of T. gondii growth in Huh7 cells or that NO might have an anti-T. gondii effect in Huh7 cells. Taken together, these results demonstrate that inhibition of NO production by aminoguanidine may ameliorate the proparasitic effect of GRA15 and the corresponding IL-1-mediated virulence mechanism.
GRA15-dependent IDO1 reduction by iNOS-mediated NO in coculture of primary human monocytes and hepatocytes. Finally, we tested whether the T. gondii GRA15-mediated virulence program is observed in primary human cells. Wild-type T.

gondii infection in human CD14
ϩ monocytes from peripheral blood caused IL-1␤ production, whereas infection by GRA15-KO parasites did not (Fig. 7A). Next, responses to IFN-␥ and IL-1␤, in terms of IDO1 and iNOS expression, were examined in primary hepatocytes (Fig. 7B). Levels of IDO1 proteins were severely reduced in primary hepatocytes (Fig. 7B), suggesting that primary monocytes and hepatocytes behaved in   a manner similar to that shown by THP-1 and Huh7 cells. Then, we examined parasite growth in cocultures of T. gondii-infected primary monocytes with primary hepatocytes (Fig. 7C). The growth of GRA15-deficient T. gondii in the cocultures was much lower than the growth of wild-type parasites (Fig. 7C). Furthermore, the levels of NO concentration and iNOS expression in cocultures containing GRA15-deficient parasites were lower than those in cocultures containing wild-type parasites (Fig. 7C and D). Conversely, the level of IDO1 protein was higher in cocultures containing GRA15deficient parasites than in cocultures containing wild-type parasites (Fig. 7D), indicating that the presence of GRA15 is advantageous for T. gondii in primary human cells. Next, we asked whether pharmacological blockade of NO production inhibits the growth of GRA15-intact wild-type T. gondii in cocultures of primary monocytes and hepatocytes ( Fig. 7E and F). Aminoguanidine treatment in the coculture resulted in increased IDO1 protein levels and reduced NO concentrations, although iNOS protein levels were unchanged (Fig. 7E). Furthermore, wild-type T. gondii growth was significantly reduced in the presence of aminoguanidine (Fig. 7F). Taken together, these results indicate that the IFN-␥-induced anti-T. gondii human cell-autonomous response requires IDO1, which is inhibited by the GRA15 parasite effector in a manner that utilizes host NLRP3 inflammasome and iNOS, in cocultures of T. gondii-infected human cells (Fig. 7G).

DISCUSSION
The present study has shown that iNOS, a well-known, critical anti-T. gondii host factor in mice (35), acts as a "pro-T. gondii" host factor in humans. iNOS participates in the GRA15-dependent virulence mechanism to suppress the IFN-␥-induced IDO1mediated antiparasite human response.
There are three major pathways by which iNOS downregulates IDO1 expression levels. NO produced by iNOS negatively regulates IFN-␥-stimulated transcription of IDO mRNA (48). In addition, NO directly binds to IDO and inhibits its enzymatic activity (49). Finally, NO induces degradation of IDO in a proteasome-dependent fashion (50). Notably, the molecular mechanisms by which iNOS-generated NO negatively regulates IDO1 expression levels have been well studied. In the current study, we demonstrated that T. gondii exploits the iNOS (NO)-dependent downregulation of IDO1 by the secreting effector GRA15 to proliferate efficiently in a variety of human cell types. We found that GRA15-intact T. gondii infection indirectly reduces IDO1 protein levels via iNOS expressed in certain secondary cells, which is responsive to IL-1␤ derived from infected THP-1 cells or primary monocytes. In mice, genetic ablation of IDO1 has been shown to inhibit T. gondii proliferation in vivo (51), suggesting that IDO1 promotes T. gondii growth in mice. Conversely, pharmacological blockade of IDO by 1-MT in mice leads to defective parasite clearance in vivo (52). Thus, controversy exists regarding the anti-T. gondii role of IDO1 in mouse models. In contrast, iNOS is well established as a critical anti-T. gondii factor in mice (35). IFN-␥-activated mouse macrophages show neither IDO activity nor tryptophan degradation but can inhibit T. gondii growth by using the NO produced via iNOS (35). However, this study and another (28) have both shown that NO may not play a significant role in human hepatocytes, monocytes, or macrophages; in contrast, a different study showed that NO can inhibit T. gondii growth in human astrocytes (53). Thus, the role of NO and iNOS in the human response against T. gondii has been unclear. Although nitrite levels in murine cells can reach concentrations greater than 100 M (54), those in human cells remain less than 10 M, which might be too low for parasite growth inhibition. Given that IFN-␥ stimulates IDO1 expression in many human cell types (33,34), humans may have evolutionarily adopted IDO1 for primary anti-T. gondii cell-autonomous immunity in hepatocytes; mice may have preferentially selected iNOS and IFN-␥-inducible GTPases for this purpose. Indeed, we failed to observe IDO1 expression in IFN-␥-stimulated mouse hepatocytes. Furthermore, the absence of IFN-␥-inducible GTPases in mouse hepatocytes completely abolished the IFN-␥-induced reduction of T. gondii numbers. Thus, such species-specific differences regarding dependence on IDO1 and IFN-␥-inducible GTPases in antiparasitic responses in humans and mice, respectively, may make the role of iNOS as a hostderived covirulence factor for GRA15 difficult to clearly elucidate in the mouse model (21,22). The role played by GRA15 in suppressing IFN-␥-induced anti-T. gondii cellautonomous immunity is evident only when GRA15-intact T. gondii parasites are used to infect normal monocytes that are then cocultured with normal hepatocytes. Because T. gondii preferentially infects CD11b ϩ cells (e.g., monocytes) and spreads throughout various tissues from the primary sites of infection in vivo (37), GRA15 might play a role in parasite proliferation in human liver tissue, where newly arrived monocytes infected with GRA15-intact T. gondii might produce IL-1␤ and inhibit IDO1 expression in an iNOS-dependent fashion.
Our current study has shown that GRA15, a protein in type II T. gondii, promotes parasite survival in IFN-␥-stimulated human cells. Of note, type II T. gondii is the most prevalent cause of both human congenital and acquired toxoplasmosis in North America and Europe (55)(56)(57). It is possible that type II GRA15-dependent IL-1␤ production, which suppresses the IFN-␥-induced IDO1-dependent immune response via iNOS, might be involved in the pathogenesis of human toxoplasmosis. Of note, treatment with an inhibitor of NO production reduced T. gondii numbers to levels below those seen under both THP-1/Huh7 and primary monocyte/hepatocyte coculture conditions, suggesting that pharmacological blockade of NO production might ameliorate human toxoplasmosis. Furthermore, a previous study also reported that costimulation with IL-1␤ and IFN-␥ inhibited IDO activity mediated by NO in a Staphylococcus aureusinfected human cell line, thereby resulting in an impaired IFN-␥-mediated reduction in bacterial numbers (50). Thus, IL-1␤ might have a pathogenic function in the suppression of the IFN-␥-mediated, IDO1-dependent, cell-autonomous response to T. gondii and to various other human intracellular pathogens.
In summary, we have shown that T. gondii can suppress the IFN-␥-induced antiparasitic response by indirectly targeting IDO1 in hepatocytes cocultured with monocytes. By investigating differences between human and mouse immune responses, unidentified virulence mechanisms associated with known or unknown T. gondii effectors may be discovered in the future. Additionally, pharmacological blockade of NO production could offer a novel therapeutic strategy for treating human toxoplasmosis.
Generation of gene-targeted human cell lines by CRISPR/Cas9 genome editing. Human cells were electroporated with the pEF6-hCas9-Puro vector containing target gRNA1/2 using NEPA21 (Nepa Gene). At 24 h postelectroporation, 0.5 to 5 g/ml puromycin was added for 5 to 10 days to select for cells with stably integrated genes. Cells were plated in limiting dilution in 96-well plates to isolate single-cell clones. To confirm complete target gene deficiency, the expression levels of IDO1, iNOS, IL-1R1, IRGM, CASP1, NLRP1, NLRP3, MyD88, and ATG16L1 protein were analyzed by Western blotting.
Generation of gene-targeted T. gondii strains by CRISPR/Cas9 genome editing. Prugniaud (Pru) T. gondii parasites expressing luciferase were filtered, washed, and resuspended in Cytomix (10 mM KPO 4 , 120 mM KCl, 0.15 mM CaCl 2 , 5 mM MgCl 2 , 25 mM HEPES, 2 mM EDTA). Parasites were mixed with 50 g of sgGRA15-1 and sgGRA15-2 CRISPR plasmid along with 40 g of the targeting vector linearized by KpnI and SacI and were supplemented with 2 mM ATP and 5 mM glutathione (GSH). Parasites were electroporated by the use of a Gene Pulser II instrument (Bio-Rad Laboratories). Selection by growth for 14 days in 25 g/ml mycophenolic acid (Sigma)-25 g/ml xanthine (Wako) was used to obtain a stably resistant clone. And then parasites were plated in limiting dilution in 96-well plates to isolate single clones. To confirm the disruption of the gene encoding GRA15, we analyzed mRNA of GRA15 from WT and GRA15-KO parasites by quantitative reverse transcription-PCR (RT-PCR). In addition, we observed comparable levels of in vitro growth and in vivo virulence with respect to each of the mutants and to the parental line.
Complementation of GRA15 in GRA15-KO T. gondii. To complement the GRA15-KO parasites, the GRA15 coding region and putative promoter region (1,940 bp upstream of the start codon) were amplified from Pru T. gondii genomic DNA using primers TgGRA15_F and TgGRA15_R (Table S1), subcloned into pCR-Blunt II TOPO (Thermo Fisher), and sequenced. The plasmid was cut by BamHI and collected at the indicated time points and measured by the use of an LDH cytotoxicity colorimetric assay kit according to the manufacturer's instructions. Culture supernatant of uninfected cells was used as a negative control. Culture supernatant of Triton X-100 (0.1%)-treated cells (the Triton X-100 was used to kill all the cells) was used as a positive control.
Measurement of parasite numbers by quantitative PCR. Total DNA was extracted from infected cells by using a DNeasy blood and tissue kit (Qiagen) according to the manufacturer's instructions and diluted to optimum concentrations for the quantitative PCR (qPCR) assay. The qPCR assays were performed with a CFX Connect real-time PCR system (Bio-Rad Laboratories) using a Go-Taq real-time PCR system (Promega). The parasite number was calculated by determination of the amounts of genomic DNA of the SAG1 gene using the standard curve (see Fig. S1F in the supplemental material). The standard curve was established by analysis of the numbers of parasites (ranging from 5 ϫ 10 2 to 5 ϫ 10 6 parasites) and of the qPCR cycle numbers of the SAG1 gene DNA.
Statistical analysis. All statistical analyses were performed using Prism 7 (GraphPad). All experimental points and n values represent averages of results from each of three biological replicates (three independent experiments). The statistical significance of differences in mean values was analyzed by using an unpaired two-tailed Student's t test. P values of less than 0.05 were considered to be statistically significant.