|Home | About | Journals | Submit | Contact Us | Français|
European and North American strains of the parasite Toxoplasma gondii belong to three distinct clonal lineages, type I, II and III, which differ in virulence. Understanding the basis of Toxoplasma strain differences and how secreted effectors work to achieve chronic infection is a major goal of current research. Here we show that type I and III infected macrophages, a cell type required for host immunity to Toxoplasma, are alternatively activated, while type II infected macrophages are classically activated. The Toxoplasma rhoptry kinase ROP16, which activates STAT6, is responsible for alternative activation. The Toxoplasma dense granule protein GRA15, which activates NF-κB, promotes classical activation by type II parasites. These effectors antagonistically regulate many of the same genes, and mice infected with type II parasites expressing type I ROP16 are protected against Toxoplasma-induced ileitis. Thus, polymorphisms in determinants that modulate macrophage activation influence the ability of Toxoplasma to establish a chronic infection.
Toxoplasma gondii is an obligate intracellular parasite capable of infecting a wide range of mammalian hosts including humans. Most Toxoplasma strains isolated in Europe and North America belong to just three distinct clonal lineages, the type I, II and III strains. These strains differ in virulence in mice and likely cause different sequelae in humans (Boothroyd and Grigg, 2002). Toxoplasma secretes effector molecules from two secretory organelles, rhoptries and dense granules, into the host cytosol which modulate host signaling pathways and dictate virulence (Reese et al., 2011; Rosowski et al., 2011; Saeij et al., 2006; Saeij et al., 2007; Taylor et al., 2006). A major aim in the field is to understand the genetic basis underlying Toxoplasma strain differences in virulence and how secreted effector molecules work to achieve chronic infection for the parasite.
A possible strategy to establish a chronic infection is through the modulation of the host’s Th1 response, which culminates in the production of IFN-γ the main mediator of resistance against Toxoplasma. However, the Th1 response must also be tightly regulated otherwise a lethal inflammatory response develops. For example, mice lacking the regulatory cytokines IL-10 or IL-27 die of severe immune pathology when infected with type II strains (Gazzinelli et al., 1996; Stumhofer et al., 2006). Th1 type cytokines (e.g. IFN-γ) synergize with pattern recognition receptors on macrophages to signal for the classical activation (or M1) of macrophages, which exert antimicrobial functions against intracellular pathogens and requires the activity of NF-κB, IRF and C/EBPδ transcription factors (Medzhitov and Horng, 2009). Macrophages are necessary for host protection to Toxoplasma (Dunay et al., 2008) and the IL-12 response by infected macrophages is greatly influenced by the strain type (Robben et al., 2004). The molecular mechanism underlying parasite strain-differences in pro-inflammatory cytokine induction by macrophages has not been resolved.
In contrast, alternatively activated macrophages (or M2) develop in a Th2 cytokine environment (IL-4, IL-13) and are inhibited by Th1 type cytokines (Martinez et al., 2009). M2 macrophages secrete anti-inflammatory molecules that can down-regulate Th1 immune processes and are important in the immune response against worm infections. M2 activation is promoted by STAT6 and PPARγ transcription factors (Odegaard et al., 2007) and mice deficient in IL-4, a potent mediator of M2 polarization, are more susceptible to acute Toxoplasma infection (Roberts et al., 1996). IL-4 treatment, as well as TLR stimulation (El Kasmi et al., 2008), alters L-arginine metabolism in macrophages through the induction of the arginase enzyme. L-arginine and the arginase catabolite ornithine, can be scavenged by parasites to generate ATP (Cook et al., 2007) and assist their replication (Iniesta et al., 2001). Thus, the ability of Toxoplasma, or other protozoan parasites, to induce specific macrophage activation states could have immediate consequences on virulence, local parasite burden and inflammatory related pathologies.
Given that Toxoplasma elicits a strong Th1 response, it is not known whether the parasite actively counteracts collateral damage following inflammation, or if there are strain differences in such a response. In this report, we show that mouse macrophages infected with type I and III Toxoplasma strains are polarized towards an M2 activation state, while type II infected macrophages resemble aspects of M1 activation. This difference is due to polymorphisms in the Toxoplasma rhoptry kinase ROP16 and dense granule protein GRA15, which activate STAT6 and NF-κB signaling pathways, respectively. Finally, type II strains that express ROP16 from the type I strain fail to cause intestinal inflammation in susceptible C57BL/6 (B6) mice. This report provides direct evidence that Toxoplasma infected macrophages can be alternatively activated, and describes parasite-specific factors that work independently of pattern recognition receptors to achieve M1 or M2 activation.
Toxoplasma strain differences in the modulation of host cells at the site of infection might lead to variation in their ability to disseminate, replicate, or survive. In agreement with an earlier report studying the type I and II strains (Mordue and Sibley, 2003), we observed that all three strains equally infected macrophages during the first 3 days of intraperitoneal (i.p.) infection and this cell type accounted for approximately 60% of the total infected peritoneal exudate cells (PECs) (Figure S1). Neutrophils were the next most abundant cell type infected by Toxoplasma (15% of the total infected PECs). However, it was the macrophage that provided a niche permissive for parasite replication because when compared to neutrophils they exhibited a higher GFP intensity (parasites express GFP) (Figure S1) and had more parasites per vacuole when FACS sorted and analyzed under a microscope (≥ 4 compared to 1 parasites per parasitophorous vacuole, macrophages vs. neutrophils, not shown). No differences were observed in cellular recruitment, or infectivity of these or other cell types over the first 3 days of infection (Figure S1, not shown).
These observations suggest that parasite differences in virulence might be due in part to how these strains modulate or interact with macrophages. To explore this further, two mouse macrophage cell lines, RAW264.7 and J774, were infected with either the type II or III strain and gene expression profiles were analyzed by oligonucleotide arrays 18 h after infection. Dendritic cells are an important link between the innate and adaptive T cell immune responses and can be infected by Toxoplasma in vivo (Bierly et al., 2008; John et al., 2009) (data not shown). Thus, the transcriptional response of the dendritic cell line DC2.4 (Shen et al., 1997) to Toxoplasma infection was also analyzed. A total of 1173 annotated genes were differentially regulated (two-fold expression level difference) between type II and III infections: 670 in RAW264.7 cells, 238 in J447 cells and 930 in DC2.4 cells.
Bioinformatic analysis of these gene sets suggested that signaling pathways which intersect with NF-κB (pattern recognition receptor signaling, TNF, IL-8 and CD40 signaling), IRF (cytosolic receptor and interferon signaling), JAK/STAT transcription factors (interferon and IL-10 signaling) and nuclear receptors (LXR/RXR and glucocorticoid receptors) mediate gene expression differences between type II and III infections (Figure S2). Furthermore, type II-induced genes were significantly enriched in TFBSs for NF-κB, IRF, Ets, ATF/CREB and CEBP/AP1 transcription factors, as well as several nuclear receptors (Table S1). Generally these transcription factors are also activated in LPS-stimulated or M1 macrophages (Medzhitov and Horng, 2009). Conversely, type III-induced genes had significant enrichment for TFBSs of transcription factors involved in hematopoietic cell proliferation, survival, and differentiation (GATA1, E2F and HOXA9), as well as for NFIL3, a transcription factor which is regulated by STAT6 (Schroder et al., 2002). These results suggest that macrophages and dendritic cells undergo distinct programing when infected by different Toxoplasma strain types. Interestingly, many of the genes that differed in expression between cells infected with the type II and III strains are also highly expressed in either M1 or M2 macrophages, respectively (Martinez et al., 2006) (Table 1). Thus, type II- and III-infected cells appear to have been activated differently, specifically along the M1/M2 axis.
To confirm these observations, a variety of in vitro assays were performed to determine the alternative activation state of macrophages: arginase-I activity, expression of the mannose receptor type C (Mrc1/CD206) and the macrophage galactose/N-acetylgalactosamine specific lectin (mMgl or Mgl1/2). Macrophages infected with type I or III parasites, but not with type II parasites, had high arginase activity 24 h after infection (Figure 1A). In general, type III infected macrophages had higher arginase activity than type I infected macrophages. Furthermore, approximately 50% of macrophages infected with either the type I or III strain, but not the type II strain, expressed high levels of CD206 and mMgl (Figure 1B). Experiments with the DC2.4 dendritic cell line produced essentially the same results (Figure 1C). Compared to RAW264.7 cells, the DC2.4 cell line was more sensitive to arginase induction and under these conditions, the type II strain induced a small but significant increase of arginase activity compared to non-infected cells.
In contrast, the type II strain induced the expression of genes that are associated with the M1 phenotype, including many pro-inflammatory cytokines (Table 1). As has been previously noted (Robben et al., 2004), compared to other strain types, the type II strain highly induces the expression of transcripts for the pro-inflammatory cytokines IL-1β, IL-6 and IL-12p40/35 in infected bone marrow derived macrophages (BMDMs). Our transcriptional profiling largely confirmed these results, and also suggested that Il23 and Il18 transcripts were consistently induced by the type II strain in all three cell lines tested (Table 1). In agreement with these observations, the type II strain, but not the type I or III strains, elicited a strong IL-23 (p40/p19) and IL-12 (p40/p35) response by infected BMDMs as determined by ELISA (Figure 1D). Not all aspects of M1 activation were observed in type II infected cells. For example, M1 cells are known to produce high amounts of reactive oxygen and nitrogen intermediates. Although we found iNOS message (Nos2) to be enriched in type II infected cells (Table 1), compared to LPS/IFN-γ-stimulated BMDMs, infection with the type I, II or III strains generated little to no amounts of nitric oxide (NO) (not shown). The inability of Toxoplasma to directly induce iNOS protein has similarly been observed (Mordue and Sibley, 2003).
Previously, using human foreskin fibroblast (HFFs), we demonstrated that the type I and III strains, but not the type II strain, constitutively activate STAT6, a difference due to polymorphisms in the secreted Toxoplasma rhoptry kinase ROP16 (Saeij et al., 2007). Furthermore, ROP16 can directly phosphorylate the critical Y641 residue required for STAT6 activation (Ong et al., 2010), thus making ROP16 a likely candidate driving strain differences in M2 activation. To investigate this further, we analyzed if a type I strain that has the ROP16 gene deleted (type I Δrop16) (Ong et al., 2010) can induce M2 cells. Unlike wild-type type I (and III) strains, the type I Δrop16 strain failed to induce the expression of any of the M2 associated markers (arginase, CD206 and mMgl) (Figure 2), which correlated with its inability to phosphorylate STAT6 at any time point investigated (0.5, 1, 2, 4, 8, 18, not shown; 24 h shown in Figure 2C, D).
Gene expression profiling of BMDMs infected with the type I or type I Δrop16 strains indicated that ROP16 controlled the expression (≥ 2-fold) of 538 unique annotated genes. Indeed, the most differentially regulated genes were those associated with the M2 phenotype, with Arg1 at the top of the list of ROP16-induced genes (132-fold) (Table S2A). Other M2 markers including Cd206, Mgl2 and Cdh1 (E-cadherin) were also dependent on ROP16. Surface receptors that were induced by ROP16 were confirmed by FACS, which included the B7 family members B7-DC (Pdcd1lg2) and B7-H1 (Cd274), as well as the C-type lectin Dectin-1 (Clec7a) (Figure 2E). The M2 signature could clearly be seen in a TFBS analysis of ROP16-regulated genes which implicated STAT6 and other M2 transcriptional regulators (Table S2B). In summary, the M2 activation program induced by Toxoplasma is controlled by ROP16.
To confirm that ROP16-mediated differences in macrophage activation occur in vivo, we infected BALB/c mice i.p. with type I, type II, or type I Δrop16 parasites and stained for M2 markers 21 h post-infection. In type I-challenged animals, infected CD11b+ PECs (which includes monocytes/macrophages and granulocytes) stained positive for phospho-STAT6 and expressed the mannose receptor CD206 (Figure 2F, G) and this was dependent on type I ROP16. In contrast, type II infected cells did not express CD206 and were less likely to stain positive for phospho-STAT6 (Figure 2F, G). These data show that strain-specific induction of M2 cells is apparent in vivo and that the rhoptry kinase ROP16 plays a fundamental role in this process.
Recently, we identified a dense granule protein GRA15, encoded by the type II strain (GRA15II), that is secreted into the host cytosol and causes the activation and nuclear translocation of host p50/REL-A NF-κB heterodimers (Rosowski et al., 2011). Furthermore, GRA15II accounted for most of the IL-12p40 response by infected macrophages in vitro, which prompted us to explore in greater detail the role of GRA15II in M1 activation. Therefore, gene expression profiles of type II and type II Δgra15 infected BMDMs were analyzed. Genes that were dependent (≥ 2-fold) on the expression of Toxoplasma GRA15II were those typically induced in M1 cells (Table S3A) and included many pro-inflammatory cytokines (Il23a, Il6, Il12a, Il12b, Il1a and Tnf), ligands for T cell costimulatory receptors (Cd40, Cd80, Cd70, Tnfsf9 and Tnfsf4), NF-κB modulators (Nfkb1e, Nfkb1b, Nfkb1z), and some of the M1 associated chemokines (Cxcl1 and Cxcl11). Transcripts for Il10 and the IL-27/35 subunit Ebi3, which are necessary to prevent inflammatory damage following Toxoplasma infection (Gazzinelli et al., 1996; Stumhofer et al., 2006), were also highly dependent on the expression of GRA15II (Table S3A). Interestingly, Il23a was the most differentially regulated gene transcript between these stains (58-fold) and secretion of IL-23 (p40/19), IL-12 (p40/35) and IL-10 by infected BMDMs was entirely dependent on the expression of type II GRA15 (Figure 3C). TFBS analysis of the 710 genes that were regulated by GRA15II also revealed a significant enrichment of TFBSs for transcription factors known to be activated in M1 macrophages (Medzhitov and Horng, 2009), which include NF-κB, IRF, C/EBP, PU1 and ATF (Table S2B). GRA15II also repressed many genes which are associated with M2 activation (Mrc1, Mgl1, Ccl24, Ear11, Ptgs1) (Table S3A). In particular, PPARγ, which encodes a transcription factor that promotes M2 activation (Odegaard et al., 2007), was the gene most repressed by GRA15II. In summary, many aspects of M1 activation by the type II strain were largely controlled by GRA15.
To test whether a simple combination of GRA15 and ROP16 will determine the M1/M2 phenotype of infected macrophages, a series of type II parasites was generated that expressed or lacked the STAT3/6 and/or NF-κB activating versions of these genes. Gene expression profiles, pro-inflammatory cytokine secretion and arginase activity of BMDMs infected with type II, type II +ROP16I, type II Δgra15 and type II Δgra15 +ROP16I parasites were analyzed. Type II +ROP16I infected macrophages possessed significantly higher arginase activity than type II infected RAW264.7 cells (not shown) and BMDMs (Figure 3B), but lower than the response elicited by the type III strain (Figure 3B). Furthermore, by removing GRA15 from the type II +ROP16I strain (i.e. type II Δgra15 +ROP16I) arginase activity of infected BMDMs was enhanced (Figure 3B, P=0.06), suggesting that GRA15 can inhibit arginase-1 induction by ROP16. Although the expression of the other arginase isozyme, Arg2, was dependent on Toxoplasma GRA15II (5.7-fold) (Table S3A), only a small decrease of urea production was observed between macrophages infected with type II Δgra15 and type II parasites (Figure 3B), implicating Arg1 as the major producer of urea following Toxoplasma infection. Arginase activity was also determined in Stat6 knockout BMDMs. In keeping with an earlier report that Toxoplasma can induce the expression of Arg1 independently of STAT6 (El Kasmi et al., 2008), Toxoplasma elicited urea production in Stat6−/− BMDMs, albeit at levels much lower than in wild type BMDMs (Figure 3B). Importantly, strain-differences in urea production were absent in Stat6−/− BMDMs, suggesting that polymorphisms in ROP16 affect signaling via STAT6 to achieve strain-specific differences in the macrophage arginase response.
To obtain a complete picture of the macrophage response to these strains, the gene expression profiles of BMDMs infected with type II, type II Δgra15, type II +ROP16I and type II Δgra15 +ROP16I parasites were compared. Indeed, there was significant overlap between ROP16 and GRA15 regulated genes, as many genes were either synergistically or antagonistically regulated by both GRA15 and ROP16 (Table S3B). For example, M1 cytokines that were highly induced by GRA15II were downregulated following type II+ROP16I infection (Figure 3C); likely because ROP16 can inhibit the activation of NF-κB by type II strains (Rosowski et al., 2011). Conversely, many M2 markers were more highly expressed in BMDMs infected with type II Δgra15 +ROP16I than with type II +ROP16I parasites (Figure 3A, B), indicating that activation of NF-κB can inhibit the expression of M2 markers, possibly through its effect on PPARγ. Finally, GRA15 and ROP16 retained the ability to induce M1 or M2 markers in polarizing environments in vitro and in vivo (see discussion in Figure S3). Thus, a combination of GRA15 and ROP16 can control macrophage activation along the M1/M2 axis.
Recently, it was demonstrated that arginase-I expression in macrophages is a susceptibility factor during Toxoplasma infection and Arg1−/− macrophages produce more NO in response to inflammatory stimuli (El Kasmi et al., 2008). iNOS like arginase, uses L-arginine as its substrate, but converts it to L-citrulline and NO which is required for long term control of Toxoplasma infection (Scharton-Kersten et al., 1997). We therefore investigated whether the induction of arginase by Toxoplasma ROP16 has a similar effect on the production of NO by macrophages. By adding different concentrations of L-arginine to L-arginine-free medium, it was determined that below ~40 mg/L of L-arginine there is a rapid decline in NO production by mouse macrophages stimulated with LPS and IFN-γ (Figure 4A) (commercial DMEM contains 84 mg/L L-arginine, mouse serum contains ~35 mg/L L-arginine). RAW264.7 macrophages were infected for 24 h with type I or type I Δrop16 parasites and activated with LPS and IFN-γ for 24 h. Although high arginase activity was induced only in macrophages infected with type I parasites, both strains inhibited NO production equally well at L-arginine concentrations of 15, 40 and 80 mg/L (not shown). Since Toxoplasma infected cells are refractory to IFN-γ signaling (Zimmermann et al., 2006), infection and stimulation was performed concurrently to bypass this refractory state. Indeed, NO production was restored under these conditions. NO production by stimulated macrophages was dependent on the concentration of L-arginine and multiplicity of infection (MOI), however, no significant difference in NO production was observed between cells infected with type I or type I Δrop16 parasites (Figure 4B), even though type I parasites induced high levels of arginase under these conditions (Figure 4C).
Arginase converts L-arginine to urea and ornithine, the latter of which can be utilized in a variety of metabolic pathways including amino acid and polyamine synthesis. Toxoplasma lacks arginase activity and is a polyamine auxotroph (Cook et al., 2007). Induction of host arginase could therefore be a strategy to increase the availability of host polyamines to support Toxoplasma replication. Since DC2.4 cells are extremely responsive to induction of arginase-I (Figure 1C), we assayed parasite growth in these cells cultured in 35 mg/L L-arginine and either stimulated with IL-4 to induce arginase-I and/or treated with the arginase-I inhibitor nor-NOHA. In general, parasite replication for all strains was enhanced in IL-4 treated cells. The difference was apparent at 48 hours for the type II/III strains, and at 24 hours for the type I strains (Figure 4D, E). Proliferative enhancement was through IL-4-induced host arginase, since IL-4 stimulated cells treated with nor-NOHA elicited a proliferative response that was similar in magnitude to parasite growth in non-treated cells (Figure 4D, E). Strains that express ROP16I grew slightly faster (in particular type II +ROP16I compared to type II) and nor-NOHA inhibited parasite replication of all strains tested. These results suggest that Toxoplasma could use the arginase metabolic pathway to promote its own growth, possibly through the action of ROP16.
Following natural peroral infection with the type II strain, C57BL/6 mice rapidly die during the acute phase of infection, which correlates with severe intestinal pathology (Liesenfeld et al., 1996). Toxoplasma-induced ileitis can be cured by removing a variety of pro-inflammatory cytokines including IFNγ (Liesenfeld et al., 1996), IL-23 and IL-22 (Munoz et al., 2009). Given the anti-inflammatory capacity of the Jak-STAT6 and -STAT3 signaling pathways to regulate these mediators, we decided to explore the role of ROP16 in inflammatory modulation at the site of infection.
To this end, B6 mice were perorally challenged with type II +ROP16I or +ROP16III parasites, or with the type II or III strains. With high dose infection (800 cysts), the majority of mice survived infection with the type II ROP16I/III strains, and those that died, did so at a much later time point than mice infected with the type II strain (Figure 5A). Survival following type II+ROP16I infection correlated with reduced intestinal inflammation across the entire length of the intestine (Figure 5D). By day 8, cellular infiltrate into the lamina propria, villi blunting and necrosis was nearly absent in type II+ROP16I infected mice (Figure 5B, D). Protection also correlated with reduced submucosa thickening (type II = 72 uM; vs. type II +ROP16I 56 uM, average over 400 measurements p <10−7, not shown). Whereas the type II strain elicited a significant influx of granulocytes into the villi and Peyer’s patches, as detected by Gr-1 and myeloperoxidase (MPO) staining, granulocyte recruitment was significantly reduced in type II +ROP16I infected animals (Figure 5E). Villus expression of iNOS (NOS2) in type II +ROP16I infected mice was reduced but not significantly different from type II infection. Lymphocytes isolated from the Peyer’s patches of type II +ROP16I infected animals produced less IL-22 and IFN-γ when stimulated with plate bound anti-CD3 and anti-CD28 antibodies compared to lymphocytes from type II infected mice (Figure 5C); while lamina propria lymphocytes from type II +ROP16I infected animals produced less IL-22 but similar amounts of IFN-γ compared to lymphocytes from type II infected animals. Thus, type I/III ROP16 expression correlates with a general dampening of Toxoplasma-induced ileitis after oral infection of B6 mice.
In this report, we have demonstrated that Toxoplasma type I and type III strains can induce the M2 phenotype, while the type II strain induces M1 like macrophages. The alternative activation of macrophages is dependent in large part on the Toxoplasma polymorphic protein kinase ROP16 while the classical activation of macrophages by the type II strain is due to the unique ability of its GRA15 protein to activate the NF-κB pathway and elicit pro-inflammatory cytokines. In total, polymorphisms in these two factors account for approximately 25–50% of the gene expression differences between type II or III infected cells (not shown).
Interestingly, ROP16 and GRA15, seem to affect signaling pathways that are themselves differentially ‘wired’ between different hosts. For example, there is considerable mouse strain variation in the ability to generate M1 or M2 cells (Mills et al., 2000). One hypothesis is that parasite effectors from different Toxoplasma strains evolved to work optimally in hosts predisposed to certain types of immune responses, such as those along the Th1/Th2/Th17 or M1/M2 axes. Conversely, ending up in the wrong host might lead to severe disease and failure to establish chronic infection. Evidence surrounding this idea include: 1) C57BL/6, but not BALB/c, mice challenged orally with type II strains, but not type III strains (not shown), can develop severe immune pathology caused by high levels of pro-inflammatory cytokines in the small intestine (ileitis) (Liesenfeld et al., 1996); 2) Chronic infections with type II strains, but not other strains, can cause severe pathology in the brain (encephalitis) of susceptible mice (Suzuki and Joh, 1994); 3) Type I or type I/III recombinant strains are more often sampled in patients that present severe ocular toxoplasmosis (Grigg et al., 2001); and 4) Although the type I strain is lethal in laboratory mice, rats are extremely resistant to Toxoplasma infection in general (Sibley and Ajioka, 2008). Whether, the type II or other strains are uniquely suited to spread in nature is an unresolved question, however the type II strain is the most abundant strain identified in livestock and in humans in North America and Europe.
The fact that type I and III parasites initially induce M2 macrophages seems to contradict the well-established fact that Toxoplasma infections with the type I strain induces a strong Th1 response (Mordue et al., 2001). We favor the hypothesis that early alternative activation of macrophages, decreased macrophage IL-12 secretion (Robben et al., 2004), and reduced dendritic cell expansion following type I infection (Tait et al., 2010) contribute to a lower IFN-γ and cytotoxic host response against the parasite. Later during infection, the higher type I parasite burden could lead to increased levels of pro-inflammatory cytokines (Mordue et al., 2001). This could be due to ‘danger signals’ released from lysing parasites such as profilin that can directly stimulate dendritic cells to produce high levels of IL-12 (Yarovinsky et al., 2005).
If both the type I and III strains induce alternative macrophage activation, why then are type III strains avirulent? The major Toxoplasma locus responsible for the difference in virulence between type I and III strains encodes another rhoptry protein kinase, ROP18 (Saeij et al., 2006; Taylor et al., 2006), which can phosphorylate and thereby inactivate mouse p47 GTPases, which are crucial for IFN-γ mediated killing of Toxoplasma (Fentress et al., 2010; Steinfeldt et al., 2010). Type III strains express extremely low levels of ROP18 due to an insertion in its promoter (Saeij et al., 2006) and the addition of the type I ROP18 locus to a type III strain causes this strain to be as virulent as type I strains (Taylor et al., 2006). In contrast, we recently demonstrated that the virulence differences in mice between type II and III strains is multifactorial with five Toxoplasma loci being involved, including ROP18, ROP16 (Saeij et al., 2006), ROP5 (Reese et al., 2011) and most likely GRA15 (Rosowski et al., 2011). With the exception of ROP5, evidence now exists that these virulence factors can directly manipulate macrophage function and the likelihood that other Toxoplasma effectors, polymorphic or not, would do likewise seems probable, especially considering the findings presented here.
We recently reported that following i.p. infection with type II Δgra15 parasites, a lower host production of IL-12p70 and IFN-γ at 1–2 days of infection preceded a higher parasite burden at day 5 compared to the type II strain (Rosowski et al., 2011). It is possible that M1 activation by GRA15II helps drive IFN-γ production to keep parasite numbers low, thus facilitating chronic infection. The induction of IL-12 by GRA15II may be especially important in hosts that do not express TLR11, as in humans (Roach et al., 2005). With respect to the in vivo effects of ROP16, we previously observed that the type II +ROP16I or +ROP16III strains were less virulent following i.p. infection compared to the parental type II parasites (Saeij et al., 2006). We have extended this analysis to the oral model, and found that ROP16I prevents Toxoplasma-induced ileitis. The mechanism by which ROP16 promotes host survival is currently under investigation, but it is unlikely that ROP16 mediates its effect by inhibiting macrophage NO production. L-arginine depletion by arginase-1 also inhibits T cell responses (Pesce et al., 2009). Thus, ROP16-induced arginase activity, suppression of IL-12 family members, and its induction of the T cell co-inhibitory molecules B7-DC and B7-H1, might all feed into a general dampening of the CD4 T cell response, which is implicated in driving Toxoplasma-induced intestinal inflammation.
In conclusion, Toxoplasma appears unique in that strains of the same species can elicit polar-opposite responses in infected macrophages. The rationale for alternative activation by Toxoplasma might be two fold; 1) to limit inflammatory damage in the infected host by quelling the Th1 response aimed at parasite elimination, and 2) to scavenge polyamines or other metabolites for energy consumption. On the other hand, the promotion of M1 cells could assist the host to develop a better Th1 response required for parasite clearance, and paradoxically allow the establishment of chronic infection. It appears that the survival of both the host and parasite hangs in the delicate balance between opposing pro- and anti-inflammatory responses. Understanding how Toxoplasma regulates these immune decisions may provide important clues to their widespread success in establishing chronic infection.
Six to ten week old female BALB/cJ, C57BL/6J or B6.129S2(C)-Stat6tm1Gru/J (Jackson Laboratories) mice were used in all experiments. All mice were maintained in specific pathogen-free conditions in accordance with institutional and federal regulations. The Toxoplasma strains used in this study are described in the supplementary methods. RAW264.7, J774 and DC2.4 cell lines were maintained in the same medium as HFFs with an additional 10 mM HEPES (GIBCO Invitrogen). BMDMs were generated by culturing bone marrow cells in 20% L929 cell-conditioned medium, as previously described (Rosowski et al., 2011). This method yielded a highly pure population of CD11b+ F480+ macrophages (>99%) by FACS. All parasite strains and cell lines were routinely checked for mycoplasma contamination and it was never detected.
BALB/c mice were infected i.p. with 3×106 parasites and 21 h later, PECs were harvested and all manipulations were done in the presence of Phosphatase Inhibitor Cocktail 2 (Sigma) and 1x Roche protease inhibitor (Roche). PECs were blocked with 4% FBS, FcBlock (BD Pharmingen), normal rat and mouse serums (Caltag Laboratories), and stained with α-CD11b M1/70 PE or Pacific Blue (eBioscience), washed, fixed in 4% paraformaldehyde and permeabilized with 0.5% saponin in blocking solution, stained with rabbit α-SAG-1, α-CD206 MR5D3 biotin or PE (Serotec), and α-pSTAT6 pY641 J71-773.58.11 Alexa Flour 647 (BD Pharmingen), washed and detected with goat α-rabbit Alexa Fluor 488 (Invitrogen) and streptavidin PerCP (BD Pharmingen) all in permeabilization solution. For the staining of infected BMDMs (MOI=0.5), cells were blocked with 10% normal mouse and hamster serums (Jackson Immuno Research), FcBlock and stained with PE labeled antibodies to CD86 (eBioscience), B7-H1, B7-DC (BD Pharmingen), Dectin-1 or a RatIgG2a isotype control (R&D Systems), followed by washing. Cells were fixed with 2% formaldehyde and washed before FACS analysis.
Parasites (MOIs of 20, 10, 5 or 1) were added to cells (105 cells/well) grown in 96-well plates and arginase activity was measured 20–24 h after infection as described (Corraliza et al., 1994). Wells with similar numbers of viable parasites were compared between parasite strains as inferred from a plaque assay. To measure the NO response, RAW264.7 cells were grown in DMEM without L-arginine complemented with 10% dialyzed serum and defined concentrations of L-arginine (MP Biomedicals). Nitrite was measured by using the Griess reaction.
Mouse macrophage (RAW264.7 and J774) and dendritic (DC2.4) cell lines were grown in a T75 to 80% confluency and were infected with the Me49 or CEP strains (MOI=7). BMDMs were plated at 3 × 106 cells per well (6 well plate) infected (MOI=3) for eighteen hours after which RNA was isolated using TRIzol (Invitrogen). RNA was labeled and hybridized to a mouse Affymetrix array (Mouse 430 2.0 or Mouse 430A 2) and analyzed as described (Rosowski et al., 2011). Microarray data has been uploaded to the public data repository GEOarchive (NCBI).
Cells were fixed with 3% formaldehyde, treated with methanol, blocked and permeabilized with 3% bovine serum albumin, 5% goat serum, 0.2% Triton X-100 in PBS. Coverslips were incubated in permeabilization solution with antibodies specific for Mrc1/CD206 sc-58987, mMgl sc-56109 or polyclonal antibodies against p-STAT6 sc-11762 (Santa Cruz Biotechnology), and a mouse monoclonal antibody against the surface antigen SAG-1, DG52. Alexa Fluor 488 or 594 (Molecular Probes) coupled secondary antibodies and Hoechst dye were used for antigen and DNA visualization with a fluorescence microscope.
Lysates were prepared from infected (MOI=10 for 24 h) macrophages grown in a 6 well plate. Western blots were performed as described (Rosowski et al., 2011) using antibodies against GAPDH sc-32233 (Santa Cruz Biotechnology), p-STAT6 558241 (BD Pharmingen), mMgl AF-4297 (R&D systems) and SAG-1.
DC2.4 cells were cultured in DMEM containing 10% dialyzed serum and 35 mg/ml L-arginine and infected with an MOI of ~1 and simultaneously treated with 50 ng/ml mouse IL-4 (PeproTech) and/or 100 μM nor-NOHA (Nω-hydroxy-nor-Arginine) (Cayman Chemical) to induce or inhibit arginase, respectively. After 24–48 h, cells were fixed, blocked, permeabilized and stained as described above. The number of parasites per vacuole for 50 vacuoles was counted per each condition.
Brain homogenate of chronically infected mice were stained with dolichos biflorus-FITC (Vector Laboratories), and cysts were enumerated by microscopy. B6 mice were orally gavaged with 800 cysts and monitored for survival. Compared to the ME49 strain, the Pru strain is less virulent and higher cyst numbers are needed to cause intestinal inflammation. For tissue sectioning, 8 days after infection the entire length of the small intestine (duodenum to ileum) was cut longitudinally, fixed in 10% buffered formalin, rolled into a ‘jelly roll’, sliced in two, mounted into cassettes, and processed for sectioning and Haematoxylin Eosin staining. Antibodies to Gr-1 (Ly6C/G) RB6-8C5 (BD), MPO Ab-1 RB-373-A (Neo Marker) and iNOS sc-650 (Santa Cruz Biotechnology) were detected with HRP-linked secondary antibodies (Dako, BD). For ex vivo cytokine analysis, the Peyer’s patches were dissected and lymphocytes were obtained by crushing. Pieces of the small intestine were then incubated in 5 mM EDTA in HBSS, washed, digested with 1.25 mg/ml collagenase V and 50 U/ml DNAse and lamina propria lymphocytes were purified over a 40/80% Percoll (GE) gradient. Cells were washed, and 3×105 cells were plated in 96-well V bottom plates (Costar) coated with plate bound α-CD3ε and α-CD28 (5 ug/ml) in RPMI medium with 10% FBS, supplements and antibiotics. Twenty h later, supernatant was analyzed by ELISA for mouse IFN-γ or IL-22 (eBioscience).
KJ is supported by the Irvington Postdoctoral Fellowship Program of the Cancer Research Institute (CRI). Work in the Hunter laboratory was supported by the State of Pennsylvania and NIH AI42334, and EDTW was supported by T32 AI07532. YCO was supported in part by a Stanford Graduate Fellowship and a NSF Predoctoral Fellowship. AJ was supported by a T32 and a NSF GRFP Fellowship. Work in the Boothroyd lab was supported by NIH AI21423 and AI73756. Work in the Chien lab was supported by NIH AI33431 and U19 AI057229. JS was supported by NIH AI080621. We also thank Chakib Boussahmain for excellent assistance with tissue sectioning and staining.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.