Treatment of mice with S4B6 IL-2 complex prevents lethal toxoplasmosis via IL-12- and IL-18-dependent interferon-gamma production by non-CD4 immune cells

Toxoplasmic encephalitis is an AIDS-defining condition. The decline of IFN-γ-producing CD4+ T cells in AIDS is a major contributing factor in reactivation of quiescent Toxoplasma gondii to an actively replicating stage of infection. Hence, it is important to characterize CD4-independent mechanisms that constrain acute T. gondii infection. We investigated the in vivo regulation of IFN-γ production by CD8+ T cells, DN T cells and NK cells in response to acute T. gondii infection. Our data show that processing of IFN-γ by these non-CD4 cells is dependent on both IL-12 and IL-18 and the secretion of bioactive IL-18 in response to T. gondii requires the sensing of viable parasites by multiple redundant inflammasome sensors in multiple hematopoietic cell types. Importantly, our results show that expansion of CD8+ T cells, DN T cells and NK cell by S4B6 IL-2 complex pre-treatment increases survival rates of mice infected with T. gondii and this is dependent on IL-12, IL-18 and IFN-γ. Increased survival is accompanied by reduced pathology but is independent of expansion of TReg cells or parasite burden. This provides evidence for a protective role of IL2C-mediated expansion of non-CD4 cells and may represent a promising lead to adjunct therapy for acute toxoplasmosis.


Toxoplasma-driven IFN-γ secretion by non-CD4 immune cells following oral infection with brain cysts or intravenous (i.v.) infection with tachyzoites.
Given that control of acute toxoplasmosis critically depends on IFN-γ 7 and non-CD4 immune cell types, such as CD8 + T cells, DN T cells and NK cells, are prime IFN-γ producers, we wanted to delineate the mechanistic requirements of IFN-γ production by these cell types in response to T. gondii. We furthermore wanted to explore whether responses were similar after oral infection (a common natural route of infection), i.v. infection with tachyzoites (modelling blood transfusion, a rare but significant-for the individual-route of infection 32 and the often used purely experimental i.p. route of infection with tachyzoites. We first inoculated naïve B6 mice with 10, 40 or 100 T. gondii ME49 cysts and assessed IFN-γ production by viable splenic CD3 + CD4 + , CD3 + CD8 + , CD3 + CD4 -CD8 -(DN) T cells and CD3 -NKp46 + cells 1 day and 5 days after inoculation. Whereas no IFN-γ production was observed 1 day after inoculation, a significant increase in IFN-γ-secreting cells was detected at 5 days after inoculation in spleen (Fig. 1a,b), mesenteric lymph nodes (MLN) (Fig. 1e,f) and Peyer's Patches PP (Fig. S1a,b). Up to 10% of CD8 + T cells and DN T Cells and up to 50% of all NK cells stained IFN-γ + , particularly following inoculation with 40 and 100 cysts.
We also investigated if rapid IFN-γ production could be induced by inoculation with tachyzoites via the i.v. and i.p. routes using a short-term in vivo exposure model in which naïve B6 mice were exposed to T. gondii tachyzoites for a maximum of 72 h. When mice were injected i.v. or i.p. with 10 5 tachyzoites, no significant IFN-γ production could be seen in either spleen, MLN or PP within 72 h (Fig. S1e). However, i.v. or i.p. inoculation with 10 7 tachyzoites led to secretion of IFN-γ by CD3 + CD8 + , CD3 + CD4 -CD8 -(DN) T cells and CD3 -NKp46 + cells in spleen, MLN and PP as early as 2-24 h after inoculation (Fig. 1c,d, g, h; Fig. S1c,d), mirroring the results seen 5 days after a cyst inoculation (Fig. 1b,f). Importantly, at 24 h after tachyzoite inoculation, levels of other acute inflammatory mediators, such as IL-6, TNFα and IL-10, were almost indistinguishable from naïve mice ( Fig. 1i-k), indicating that these cytokines were not impacting on protective IFN-γ responses 24 h after i.v. infection.
These results show that i.v., i.p. tachyzoite infections and oral brain cyst infections induce almost identical acute immune responses. Given that it is difficult to quantify the number of bradyzoites within brain cysts used for oral infection and, moreover, dissemination patterns following oral infection are erratic in individual mice 33, we subsequently focused on IFN-γ secretion 24 h after i.v. injection of tachyzoites as our primary readout for further dissection of the underlying mechanistic requirements.
Rapid IFN-γ secretion in response to T. gondii requires IL-12 and IL-18. Whereas the role of IL-12 in IFN-γ secretion is well established for T. gondii 2, rapid production of IFN-γ in response to other intracellular pathogens, such as S. enterica, L. monocytogenes and M. tuberculosis has also been linked to the upstream effects of IL-18 28,29 . To interrogate whether or not, and how early, IFN-γ secretion in response to T. gondii also requires IL-18, we exposed naïve B6 mice to T. gondii ME49 tachyzoites and treated the animals with neutralizing monoclonal antibodies (mAb) to IL-12, IL-18 or IL-12 and IL-18 immediately after inoculation. We focused Scientific RepoRtS | (2020) 10:13115 | https://doi.org/10.1038/s41598-020-70102-1 www.nature.com/scientificreports/ on NK cells for these and subsequent experiments, since this was the cell type for which the highest proportion of cells stained positive for IFN-γ following inoculation with T. gondii (see Fig. 1). At 24 h after exposure, we assessed IFN-γ secretion by NK cells in the spleen ex vivo. Neutralization of IL-12 and IL-18 significantly reduced IFN-γ production, with IL-12 contributing approximately 50% and IL-18 approximately 30-40% of  www.nature.com/scientificreports/ the response (Fig. 2a). Consistent with this, where IL-12 levels in the serum of infected mice peaked at approximately 2 h after inoculation, the levels of IL-18 mirrored those of IFN-γ for up to 72 h (Fig. 2b). The significant reduction of rapid IFN-γ production in Il18 -/mice, and the almost complete absence of rapid IFN-γ production in anti-IL-12-treated Il18 -/mice, further confirmed a direct correlation between IL-12, IL-18 and IFN-γ secretion (Fig. 2c,d). Furthermore, treatment with anti-IL-12 and/or anti-IL-18 also reduced concentrations of IFN-γ, IL-12 and IL-18 in the serum of infected mice in an additive manner ( Fig. 2d-f). These results suggest a hierarchical relationship in which a primary IL-12-driven IFN-γ response is followed by an IL-18-dominant IFN-γ response. We concluded that innate IFN-γ secretion in response to T. gondii is driven by the secretion of IL-12 and IL-18.
As expected, Caspase1/11 -/mice did not secrete significant levels of IL-18 following T. gondii inoculation (Fig. 3b), indicating that the remaining IFN-γ response in Caspase1/11 -/mice is driven by IL-12. Surprisingly, when we tested mice deficient in the upstream NLR family pyrin domain-containing proteins 1 and 3 (NLRP1 and NLRP3), NLR molecules that had been implicated previously in recognition of T. gondii 24, both knockout strains secreted indistinguishable amounts of IL-18 compared with B6 mice ( Fig. 3b; Table S2). This data suggested a redundant role for NLRP1 and NLRP3. However, even double knockout and heterozygous Nlrp1 +/− Nlrp3 +/− , Nlrp1 −/− Nlrp3 +/− and Nlrp1 +/− Nlrp3 −/− mice secreted high levels of IL-18 after exposure to T. gondii ME49 tachyzoites ( Fig. 3b; Table S2), suggesting that additional PRR molecules must be involved in sensing of T. gondii invasion in vivo. Taken together these results indicate that rapid IFN-γ secretion in vivo in response to T. gondii depends on the inflammasome → caspase-1→ IL-18 axis, and that T. gondii likely activates at least three different inflammasomes in vivo.

Toxoplasma gondii activates inflammasomes in multiple cell types.
To further investigate the role of cytosolic PRRs in sensing T. gondii invasion, and to potentially target inflammasome activation for preventive or therapeutic intervention strategies, we next tried to identify the T. gondii-sensing cell type in vivo. To do this, we made use of a red fluorescent protein (RFP) tagged T. gondii ME49 (T. gondii ME49-RFP) strain to track parasite uptake by different immune cell subsets in the spleen. Twenty-four hours after tachyzoite injection, T. gondii ME49-RFP also induced rapid IFN-γ secretion by splenic CD3 + CD4 + , CD3 + CD8 + , CD3 + CD4 -CD8 -(DN) T cells and CD3 -NKp46 + cells (Fig. 4a) and high levels of serum IL-18 ( Fig. 4b), similar to wild-type T. gondii ME49 (see Figs. 1 and 2). Approximately 0.5% of all splenocytes contained T. gondii ME49-RFP in vivo 24 h after inoculation (Fig. 4c). Sorted RFP + cells secreted significantly more IL-18 ex vivo compared to RFP − cells (Fig. 4d), and further surface phenotyping revealed that T. gondii ME49-RFP was primarily contained in monocytes, neutrophils and CD8α + dendritic cells (Fig. 4e,f). Splenic MHC-II + CD11c + DCs, CD11b + Ly6G + neutrophils and CD11b + Ly6C + monocytes each comprised approximately 20-30% of all RFP-containing cells after i.v. tachyzoite injection. Only very few T cells, B cells and macrophages appeared to harbor parasites (Fig. 4e,f). To investigate if cell types that contained T. gondii ME49-RFP parasites also activated inflammasomes, we performed intracellular staining for the inflammasome adaptor molecule apoptosis-associated speck-like protein containing a carboxy-terminal CARD (ASC), and measured the activation of caspase-1 with a fluorescent inhibitor that only binds to activated caspase-1 (FLICA FAM-YVAD-FMK) 29 . Consistent with the uptake of T. gondii ME49-RFP by different cell types, T. gondii ME49-RFP parasite-harboring neutrophils, monocytes and DCs also expressed higher levels of ASC and FAM-YVAD compared with RFP − cells and FMO controls (Fig. 4g). Collectively, these results indicate that T. gondii infection activates multiple redundant inflammasomes in multiple different hematopoietic cell-types in vivo.
IL-18-driven IFN-γ secretion to T. gondii depends on viable parasites but is independent of secreted GRA proteins. Next, we assessed if rapid IFN-γ secretion in response to T. gondii required viable parasites or could be induced by soluble factors. To this end, naïve B6 mice were injected with either live, heatkilled or sonicated T. gondii ME49 tachyzoites. Inoculation with live parasites induced significantly increased IFN-γ secretion by NK cells and increased serum IL-18 levels compared to heat-killed or sonicated parasites (Fig. 5a,b). To exclude the possibility that heat inactivation and sonication destroyed soluble factors that could potentially drive this response, we also injected naïve B6 mice with HFF cell debris, which had been re-suspended in the T. gondii ME49 culture supernatant. This treatment also failed to induce IFN-γ and IL-18 secretion (Fig. 5a,b). These results indicated that viable parasites are required to initiate an IFN-γ response, suggesting that  IFNγ [ng/ml] IL12 [pg/ml] P B S a n t i I L 1 8 a n t i I L 1 2 a n t i I L 1 2 / 1 8 I l 1 8 -/ -I l 1 8 -/ -+ a n t i I L 1 2 P B S a n t i I L 1 8 a n t i I L 1 2 a n t i I L www.nature.com/scientificreports/ effector molecules with distinct host-modulatory properties 35 . In particular, dense granule (GRA) proteins have been shown to play important roles in the maintenance of the parasitophorous vacuole (PCV), for the intracellular lifestyle and to exert host-modulatory functions 36 . We further probed the parasite-derived factors that might drive early, IL-18-dependent IFN-γ secretion by exposing naïve B6 mice to a panel of T. gondii strains to test if GRA proteins are required for IL-18-driven IFN-γ secretion. Hence, we infected mice with a mutant strain of T. gondii ME49 that lacks ASP5, a critical requirement for secretion of GRA proteins 37, as well as strains lacking GRA20 or GRA23, two proteins that contain the PEXEL motif required for PCV exit. No significant difference in the levels of serum IL-18 and NK cell-produced IFN-γ was observed 24 and 48 h after inoculation with T. gondii ME49 ASP5-deficient tachyzoites compared with inoculation of a wildtype T. gondii ME49 (Fig. 5c,d), suggesting that ASP5-driven GRA export is dispensable for inflammasome activation. Similarly, inoculation with GRA20-deficient or GRA23-deficient parasites did not significantly reduce IFN-γ secretion in the absence of IL-12 ( Fig. S2). We also tested another Type II T. gondii isolate, DEG (T. gondii DEG), which had been implicated in reduced IL-1β secretion following in vitro infection of macrophages 24 but, similar to inoculation with T. gondii ME49 ASP5-deficient parasites, inoculation with T. gondii DEG did not lead to reduced levels of serum IL-18 and NK cell-produced IFN-γ in this model (Fig. 5c,d). At 48 h after DEG tachyzoite inoculation, the levels of serum IL-18 were, if anything, slightly higher compared with inoculation of T. gondii ME49, though this was not statistically significant (Fig. 5d). These data indicate that ASP5-dependent secretion of GRA proteins does not affect IL-18-driven IFN-γ secretion in vivo and highlights the diverging mechanisms that underlie in vitro IL-1β and in vivo IL-18 secretion in response to T. gondii.

IL2C treatment expands IL-18-responsive IFN-γ-secreting cell subsets. Collectively, the results
presented so far raise the prospect that, if the ability of non-CD4 cells to invoke inflammasome-dependent, IL18driven production of IFN-γ can be enhanced, it may be possible to control acute toxoplasmosis in AIDS. Hence, we investigated if targeted expansion of non-CD4 cells with IL2C treatment can achieve this. First, naïve mice were treated i.p. with IL2C complex on four consecutive days ( Fig. 6a) and, 24 h after the last IL2C injection, immune cell expansion was assessed by flow cytometry relative to untreated animals. As reported previously 38, IL2C treatment led to a significant expansion of memory CD8 + T cells, NK cells and DN T cells in spleen and MLN (Fig. 6b,c) though increases observed in the PP were not statistically significant (Fig. 6d). Next we assessed if IL2C-expanded and non-expanded CD8 + T cells, DN T cells and NK cells responded similarly to T. gondii infection. IL2C-treated and untreated mice were infected with 10 7 ME49 tachyzoites for 24 h (Fig. 6a). The percentage of CD8 + T cells, DN T cells and NK cells producing IFN-γ was almost indistinguishable between IL2C-treated and untreated mice (Fig. 6e). The number of IFN-γ + NK cells, IFN-γ + CD8 + T cells and IFN-γ + DN T cells increased 3-30 fold following IL2C treatment (Fig. 6f). Similarly, IL2C pretreatment significantly increased systemic IFN-γ levels in the serum after i.v. infection ( Fig. 6g) but, as expected, did not lead to a significant change in the levels of serum IL-18 (Fig. 6h). We also assessed the expression of IL18R and IL12R on the surface of IFN-γ + and IFN-γ − cells. IFN-γ + NK cells (data for CD8 + T cells and DN T cells not shown) expressed significantly higher levels of IL18R and IL12R compared to IFN-γ − NK cells (Fig. 6i,j). Taken together, these results show that IL2C-expanded cells respond similarly to non-expanded cells and that the effect of IL2C ---  Statistical analyses: Two-way ANOVA followed by Tukey's post hoc test. Significant differences within each mouse genotype are indicated by asterisks: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; n.s. not significant.
Further significance values are shown in Tables S1 and S2. IL2C pre-treatment protects mice from acute lethal toxoplasmosis independently of T Reg expansion and parasite burden. To definitively assess if IL2C-mediated expansion of IL-18-responsive IFN-γ-secreting non-CD4 cell subsets can prevent lethal toxoplasmosis in mice, we used the well-established oral inoculation model with T. gondii ME49 bradyzoite-containing brain cysts. As above, naïve B6 mice were treated i.p. with IL2C for four consecutive days (Fig. 7a). IL2C treatment was accompanied by a weight loss from which mice recovered within a few days (data not shown). Forty-eight hours after the last IL2C treatment, mice were inoculated orally with 10 or 40 T. gondii ME49 cysts and were assessed for weight loss and consequent survival, whereby mice were euthanized when weight loss exceeded 20% of body weight, in accordance with Animal Ethics Committee of James Cook University Approvals A2138 and A2324. All mice that had been inoculated with 40 cysts and 87% of mice that had been inoculated with 10 cysts, but had not received IL2C injections, succumbed within 14 days after inoculation (Fig. 7b,c). In contrast, IL2C pre-treatment extended survival in mice that had been inoculated with 40 cysts up to 36 days, and approximately 40% of mice that had been inoculated with 10 cysts survived until day 60 (Fig. 7b,c).  Results are presented as individual data points (d,f), pooled data means ± SEM (a,b) and representative FACS plots (c,e) and histograms (g) of 6-9 mice from two or three pooled independent experiments. Statistical analyses: One-way ANOVA followed by Dunnett's multiple comparison test (a) or Student's t-test (b,d,g); significant differences are indicated by asterisks: *p < 0.05; **p < 0.01; ***p < 0.001.  39 or IFN-γ from mice that had been treated with IL2C for four days and had been inoculated with 10 T. gondii ME49 cysts with neutralizing antibodies reversed IL2C-mediated increase in survival (Fig. 7d,e), indicating that IL2C-mediated cell expansion directly correlated with increased survival. Similarly, neutralization of IL-18, IL-12 or IFN-γ, reversed the protective phenotype (Fig. 7d,f). All mice that were not treated with IL2C succumbed to the infection by day 16, with a median survival of 11 days (Fig. 7f). Whilst 67% of IL2C-treated mice that received control rat IgG survived until day 60, the median survival for mice treated with anti-IFN-γ was 10.5 days, 13 days for mice treated with anti-IL-12 and 14 days for mice treated with anti-IL-18 (Fig. 7f). All mice that survived until day 60 were assessed for T. gondii brain cysts and all contained cysts in their brain, indicating that all mice were infected and that survival was not due to a failure of the infection to establish.

Scientific RepoRtS
To assess if IL2C pre-treatment also impacts on measurable disease parameters other than survival, we also assessed pathology, parasite burden, serum cytokine levels and T Reg numbers in MLN and lamina propria (LP) at 2, 4 and 9 days following oral cyst infection. At 9 days after infection, IL2C pre-treated mice displayed significantly reduced gross pathology of gut and liver (Fig. 7g,h) in the absence of any effect on parasite burden (Fig. 7i); thus, parasite burden was below detection limits in IL2C-treated and untreated animals at 2 days after infection and progressively increased at the same rate in both groups through days 4 and 9 post-infection (Fig. 7i). Due to the low infectious dose of 10 cysts, inflammatory cytokines such as IFN-γ, IL-6, IL-12 and TNF were not detectable in any mice at 2 and 4 days after infection and, although detectable at day 9 post-infection, there was no significant difference in the levels of these cytokines in IL2C-treated versus untreated mice (data not shown).
Similarly, systemic IL-10 levels (Fig. 7j) and T Reg numbers in MLN and lamina propria were not increased by IL2C injections (Fig. 7k-m) suggesting a role for IL2C pre-treatment independent of the previously reported T Reg expansion with JES6-1A12-containing IL2C 40,41 . Collectively, these results demonstrate a protective role of IL2C pre-treatment in acute lethal murine toxoplasmosis that is dependent on IL-12, IL-18 and IFN-γ but is independent from effects on parasite burden.

Discussion
Non-CD4 cells, such as CD8 + T cells, DN T cells and NK cells, have been implicated in early control of severe infections with intracellular pathogens, including T. gondii, M. tuberculosis and Salmonella 2,29 . Our study provides a mechanistic framework for how T. gondii activates IFN-γ secretion by protective CD8 + T cells, DN T cells and NK cells. In particular, we demonstrate that in vivo IL-18-driven IFN-γ secretion in response to T. gondii likely requires the activation of at least three different inflammasomes. The involvement of both NLRP1 and NLRP3 has been shown in other, distinct models of toxoplasmosis 24 but, in the model presented here, only Caspase1/11 −/− mice but not Nlrp1 −/− , Nlrp3 −/− and Nlrp1 ±/− Nlrp3 ±/− mice were devoid of circulating IL-18 after T. gondii infection. These results suggest that a third sensor for in vivo T. gondii detection must exist in addition to NLRP1 and NLRP3 24,42 . This conclusion is underscored by the fact that Caspase1/11 −/− mice but not Nlrp1 −/− , Nlrp3 −/− and Nlrp1 ±/− Nlrp3 ±/− mice all maintained significant levels of IFN-γ + NK cells after 24 h infection    www.nature.com/scientificreports/ inflammasome signaling pathway. This is in contrast to the often very specific recognition of viral and bacterial infections by one particular inflammasome in a distinct cell type 28,29,[43][44][45][46] . It is likely that this divergence highlights the evolutionary complexity of parasites and suggests that more highly evolved organisms have developed a more complex inflammasome-dependent interplay with their hosts. In line with this hypothesis, it was shown recently in vitro that T. gondii also activates the NLRC4 and AIM2 inflammasomes in human fetal small epithelial cells 47, as well as the expression of NLRP6, NLRP8 and NLRP13 in THP-1 macrophages 48 . Given that myeloid cell subsets often express distinct arsenals of PRRs on the cell surface and intracellularly, their ability to recognize and interact with T. gondii differs. Subsequently, the identification and characterization of distinct myeloid cell types producing IL-18 in response to T. gondii may foster innovative strategies for targeted interventions. Toxoplasma gondii appears to activate both NLRP1 and NLRP3 24, yet the specificity of this activation remains elusive. While NLRP3 activation in response to T. gondii is influenced by P2X 7 receptor-dependent potassium efflux and the induction of reactive oxygen species 47,[49][50][51] , the exact mechanisms of how T. gondii activates multiple inflammasomes remain enigmatic. It is also interesting to note that in vitro infection of mouse macrophages and human monocytes with T. gondii only leads to the secretion of IL-1β, but not IL-18 24,52 . In contrast, in vivo infection in mice leads to significant secretion of IL-18 but not IL-1β 24 . It has even been suggested that in vitro infection of human neutrophils leads to evasion of NLRP3 activation and IL-1β secretion 53 . Furthermore, in vitro activation of inflammasomes differs between T. gondii strains, and is predominantly induced by Type II parasites 24 . These findings suggest that T. gondii has evolved sophisticated diverging effector mechanisms to manipulate inflammasome biology in different host cell subsets, and suggest that secreted effector molecules and/or distinct structural proteins may underlie inflammasome activation. It is, therefore, interesting that Nlrp1 ±/− Nlrp3 ±/− mice did not show reduced IL-18 secretion after infection with T. gondii. It is important to note that in mice the Nlrp1 locus is on the same chromosome as the Nlrp3 gene, meaning that the generation of rare double knockout offspring relies on recombination rather than inheritance. It will therefore be important to further investigate the role of Nlrp1 and 3 with alternative methods, such as CRISPR/Cas9 and/or chemical inhibition.
Our study has ruled out ASP5-dependent GRA proteins 37, the most abundant family of T. gondii-derived effector molecules 35, as the primary activator of inflammasomes. GRA molecules influence several host cell pathways 54 and are required for the transport of small molecules across the parasitophorous vacuole 55 . These results do not exclude GRA proteins that don't depend on ASP5 for export, and further studies will have to investigate the role of ASP5-independent GRA proteins as well as rhoptry proteins and other surface structures in driving this process. In particular, the recently described MYR1 protein export system 56-58 may be valuable in answering if secreted effector molecules are at all required to initiate inflammasome activation.
It is tempting to speculate that the overall purpose of activating multiple inflammasomes in multiple cell types is to drive an inflammatory host response that mediates the progression of T. gondii into the chronic cyst phase, while at the same time preventing activation of parasite-killing mechanisms. Toxoplasma can invade and replicate in virtually all nucleated cell types of warm-blooded animals. From an evolutionary perspective, it is not surprising that the arms race between the host and the parasite has led to the evolution of numerous strategies to activate the immune system (from the parasite's perspective) and to sense the invasion (from the host's perspective). The fundamental differences between the habitats and the composition of the immune system of susceptible warm-blooded host species may require T. gondii to activate as many different inflammasome sensors as possible. It is well established that T. gondii requires a pro-inflammatory, IFN-γ-dominated immune response to form cysts 7 . Because transmission is critical for the parasite's survival and completion of the life cycle, it is maladaptive for T. gondii to kill its host. This may explain why IFN-γ neutralization is fatal, because IFN-γ deficiency favors tachyzoite replication and prevents cyst formation. Furthermore, these findings may also explain why T. gondii cysts reactivate after HIV co-infection in humans; HIV destroys CD4 + T cells, a prime IFN-γ producer. Hence, we reasoned that a viable adjunct therapy in T. gondii/ HIV co-infection might be achieved by boosting IFN-γ-producing NK cells, CD8 + T cells and DN T cells to prevent acute toxoplasmosis. Figure 7. IL2C pre-treatment protects mice from acute, lethal toxoplasmosis independently of T Reg expansion and parasite burden. (a) Experimental plan showing that naïve B6 mice were treated i.p. with IL2C on four consecutive days or left untreated. Two days after the last IL2C treatment, mice were inoculated orally with 10 or 40 T. gondii ME49 brain cysts and survival was assessed over time. (b,c) Percent survival of B6 mice, untreated or treated with IL2C before being inoculated orally with 40 (b) or 10 (c) T. gondii ME49 brain cysts. (d) Experimental plan showing that naïve B6 mice were treated as described in (a) with IL2C-treated animals receiving weekly i.p. injections with mAb and survival being assessed over time. (e,f) Percent survival of B6 mice, inoculated orally with 10 T. gondii ME49 brain cysts, untreated or treated with IL2C and mAb against IFN-γ, CD8, NK1.1, Thy1.2 or control rIgG (e), or with mAb against IL-12, IL-18, IFN-γ or control rIgG (f). (g,h) Gross pathology of the intestines (g) or livers (h) and parasite burden in the spleen (i) and IL-10 levels in the serum (j) of B6 mice, untreated or treated with IL2C, 2, 4 and/or 9 days after being inoculated orally with 10 T. gondii ME49 brain cysts. (k) Representative FACS plots of CD3 + CD4 + CD25 + Foxp3 + regulatory T cells from MLN and lamina propria (LP) of B6 mice, untreated or treated with IL2C, 2 days after being inoculated orally with 10 T. gondii ME49 brain cysts. (l,m) Numbers of regulatory T cells from MLN (l) and LP (m) of B6 mice, untreated or treated with IL2C, 2, 4 or 9 days after being inoculated orally with 10 T. gondii ME49 brain cysts. Results are presented as individual data points (g,h,i), pooled data means (b,c,e,f,j,l,m) or representative FACS plots (k) from two to three pooled independent experiments with 5-15 mice per group. Statistical analyses: One-way ANOVA followed by Dunnett's multiple comparison test (g,h,l,m), Student's t-test (i) or Log-rank (Mantel-Cox) test (b,c,e,f); significant differences are indicated by asterisks: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. BDL, below detection limit. www.nature.com/scientificreports/ Interleukin-2 and Interleukin-15 are critical cytokines for the maturation and survival of IL-18 responsive DN T cells, NK cells and CD8 + T cells [59][60][61][62][63][64] . The role of IL-15 in immunity to T. gondii remains controversial 65,66 , but IL-2 knockout mice are highly susceptible to T. gondii infection 67, and injection of recombinant IL-2 enhances survival of Toxoplasma-infected mice 68,69 . Because IL-2 and IL-15 signaling depends on trans-presentation 70,71 , complexing IL-2 with anti-IL-2 (IL2C) or IL-15 with IL-15RαFc (IL15C) significantly enhances their biological activity in vivo 31,72 . Importantly, the binding site of the anti-IL2 clone used in the IL2C determines whether a preferential expansion of regulatory T cells (T Reg ; anti-IL-2 clone JES6-1A12) or CD8 + T cell, NK cells and DN T cells occurs (anti-IL-2 clone S4B6) 31,73 .
Using JES6-1A12-containing IL2C, Akbar et al. 41 showed that selective expansion of T Reg cells in Type I T. gondii RH-infected animals improved control of the parasite. It was also demonstrated that T Reg expansion with JES6-1A12-containing IL2C can overcome the competition for bioavailable IL-2 by regulatory and effector T cells, leading to reduced immunopathology and morbidity during acute Type II T. gondii ME49 infection 40 . These studies are in line with other reports showing a collapse of T Reg cells during acute T. gondii infection due to IL-2 starvation and an overall protective role of T Regs in acute T. gondii-mediated immunopathogenesis [74][75][76][77] . In contrast to JES6-1A12-containing IL2C, S4B6-containing IL2C has been shown to boost NK cell and memory CD8 + T cell numbers in mice and to enhance their cytolytic capacity against viral infections, malaria 78 and cancer cells 70,79-81 . Short-term exposure of naïve mice to IL2C containing S4B6 has also been shown to enhance resistance and immunity against Listeria monocytogenes infection 82 . Our study is the first to show a protective effect of S4B6-containing IL2C pre-treatment in toxoplasmosis and our results suggest that IL2C pre-treatment can protect mice from lethal toxoplasmosis via distinct mechanisms, depending on the IL-2 mAb clone used to prepare the cytokine complex. Thus, JES6-1A12-containing IL2C seems to compensate for the limited bioavailability of IL-2 for Treg survival during acute T. gondii infection, leading to reduced immunopathology, whereas S4B6containing IL2C, whilst also reducing pathology without affecting parasite burden, does so in a Treg-independent manner. Our findings do not definitively rule out a role for Treg function or local (i.e., gut) IL-10 in SB46 IL2C treated mice (e.g., via use of Treg ot IL-10 depleted or knockout mice) but our results do show that S4B6 IL2C treatment has no effect on systemic IL-10 levels, further pointing towards a Treg-independent function. Based on our results, we conclude that S4B6-containing IL2C seems to favor survival and expansion of IL-18-driven IFN-γ secretion, possibly driving parasites towards stage conversion and cyst formation. It is, hence, tempting to speculate that both types of IL2C could have a synergistic effect if applied together.
Cytokine complex-mediated immunotherapy has not only attracted attention in models of infectious diseases but also in the cancer field 83 . IL2C treatment reduces viral load in a mouse model of gamma-herpesvirus infection 84 and impacts positively on mouse melanoma 85 and BCL1 leukemia 86 . More recently, IL2C treatment has also been tested successfully in cancer models in combination with immune checkpoint blockade 87 . IL-15/ IL-15Rα-Fc complexes (IL15C) have also been shown to expand CD8 + T cell, DN T cell and NK cell populations, and to protect mice against cerebral malaria via the induction of IL-10-producing NK cells 78 . Whether IL15C would also be protective in our model of lethal toxoplasmosis remains to be investigated. Taken together, these results suggest that cytokine complex treatment may be a more broadly applicable adjunct therapy in infectious diseases, but also highlight that the protective mechanisms may differ between different pathogens and cytokine complex types used. To our knowledge, no data are available yet on any clinical use of IL2C and IL15C in humans. It will be important to consider the hyper-inflammatory response that can be attributed to IL2C and IL15C treatment and, hence, careful consideration should be taken before using cytokine complexes clinically in the context of toxoplasmosis.
In summary, here we delineate a mechanistic framework for how IFN-γ is produced by non-CD4 cell types in vivo in response to T. gondii, including a crucial role for parasite viability and inflammasome-dependent IL-18 secretion. Our results demonstrate that in vivo inflammasome activation in response to T. gondii occurs in multiple myeloid cell types and indicate the existence of an unidentified T. gondii-sensing component. Additionally, our study excludes T. gondii-derived, ASP5-dependent, dense granule proteins as the main activators of inflammasomes in vivo. The observation that both IL-12 and IL-18 neutralization reverses the host protective role of CD8 + T cells, DN T cells and NK cell-produced IFN-γ during T. gondii infection highlights the redundancy and functional interchangeability of both cytokines during T. gondii infection. This combination of observations led us to the hypothesis that enhancement of inflammasome-dependent, IL18-driven production of IFN-γ by non-CD4 cells may be a route to control acute toxoplasmosis in AIDS. In accord with this, we provide compelling evidence for a protective role of IL2C pre-treatment in lethal toxoplasmosis. We demonstrate that IL2Cmediated expansion of CD8 + T cells, NK cells and DN T cells protects mice against acute disease and death in an IFN-γ-dependent manner. Hence, we conclude that inducing immune responses that lead to the expansion of IFN-γ-secreting CD8 + T cells, DN T cells and NK cells could be a crucial feature of improved toxoplasmosis intervention strategies, perhaps most particularly in the context of HIV co-infection and AIDS.

Methods
Mice. C57BL/6 J and Arc(S) mice were purchased from the Animal Resource Center (Perth, Australia). www.nature.com/scientificreports/ experiments, all mice were sex-and age-matched, and kept in our BSL 2 animal facility under specific pathogenfree (SPF) conditions. Type II T. gondii strains ME49, ME49-RFP, ME49 GRA20-deficient, ME49 GRA23-deficient, ME49 ASP5-deficient and DEG (ATCC, ATC50855) were maintained by continuous passage in human foreskin fibroblasts (HFF; ATCC, ATCSCRC1041) in DMEM supplemented with 10% FCS, penicillin, streptomycin and L-glutamine at 37 °C and 5% CO 2 . Parasites were harvested from recently lysed cell monolayers, passed through a 26G needle and a 3 µm TSTP Isopore™ membrane filter and concentrated by centrifugation at 500 g for 10 min. The pellet of tachyzoites was re-suspended in sterile PBS. Parasites were counted using a Neubauer hemocytometer and diluted to the required infectious dose in sterile PBS.

Parasites.
Generation of T. gondii ME49 Gra20 and Gra23 knockouts. We employed a CRISPR/Cas9 approach to insert frameshifts within the first 20 nt of the start of the coding sequence of gra20 and gra23 in T. gondii Me49 with consequential disruption of the final translated proteins. Inverse PCR was used to exchange the sgRNA of UPRT with the sgRNA for GRA20 with Ph-sgRNA_TgGRA20mutF (5′-ATG CAT AGC CGG AAC TGC GTG TTT TAG AGC TAG AAA TAGC-3′) and Ph-genCas9mutR (5′-AAC TTG ACA TCC CCA TTT AC-3′) to yield plasmid pCAS9sgGRA20. Similarly, inverse PCR was used to exchange the sgRNA of UPRT with the sgRNA for GRA23 with Ph-sgRNA_TgGRA23mutF (5′-GCA GCG CGT GCG GGA AGC AGG TTT TAG AGC TAG AAA TAGC-3′) and Ph-genCas9mutR (5′-AAC TTG ACA TCC CCA TTT AC-3′) to yield plasmid pCAS9sgGRA23. Transfection of T. gondii Me49 was carried out as described previously 89 . Twenty-four hours post-transfection, transiently transfected GFP + parasites were purified by flow cytometry as previously described 90 and individual GRA20 and GRA23 KO clones were further purified using two rounds of limiting dilution cloning. Sanger sequencing of PCR products was used to confirm disruption of the gra20 and gra23 ORFs.

Infections and monitoring.
To isolate T. gondii ME49 bradyzoite containing cysts, the brains of chronically infected Arc(S) mice (injected i.p. with 500 tachyzoites of T. gondii ME49 > 8 weeks prior) were removed, homogenized in sterile PBS, and subjected to centrifugation in a discontinuous Percoll gradient. Cysts were counted using a Neubauer hemocytometer and diluted in sterile PBS. For experiments, B6 mice were inoculated with 10, 40 or 100 cysts by oral inoculation. For mechanistic studies, B6 mice were injected i.v. in the lateral tail vein with 10 7 tachyzoites of T. gondii ME49, mutant strains on the T. gondii ME49 background or the Type II strain T. gondii DEG in a volume of 200 µl. For heat inactivation, T. gondii ME49 tachyzoites were grown as described above, enumerated, and washed twice with PBS before incubation at 62º C in a water bath for 1 h. Effective killing was verified by addition of heat-killed parasites to a HFF cell monolayer. All mice were monitored as stipulated by Animal Ethics Committee of James Cook University Approvals A2138, A2324. Chronically infected Arc(S) mice were monitored weekly for signs of morbidity and were euthanized using carbon dioxide asphyxiation for brain cyst harvesting. Mice infected orally with T. gondii-containing brain cysts were monitored daily for signs of disease and were euthanized using carbon dioxide asphyxiation at distinct time-points after infection for immunological readouts or when ethical endpoints were reached for survival experiments. Death was never used as an endpoint. Mice infected i.v. or i.p. with T. gondii tachyzoites were monitored daily for signs of disease and were euthanized using carbon dioxide asphyxiation 2 to 72 h after infection.
Isolation of leukocytes. Spleens, mesenteric lymph nodes and Peyer's Patches were extracted and mechanically disrupted by pushing cells through a 70 µm cell strainer. Subsequently, red-cell depleted, single-cell suspensions were prepared as described elsewhere 39 . Lamina propria cells were isolated from the ileum as published previously with minor modifications 91 .
Scoring of pathology. Gross pathology of ileum and liver was scored visually using a scoring system adapted from Melgar et al. 92 . For the ileum, the consistency of the intestinal contents, the degree of swelling and amount of angiogenesis were assessed. This system is based on an ascending scale of severity, for each parameter, as follows: 0 (no abnormality),1 (minimal),2 (moderate),or 3 (severe). For the liver, the colour and appearance of the organ were assessed on an ascending scale of severity from 0 (normal colour and appearance); 1 (blotchy appearance with some areas exhibiting change in colour); 2 (entire organ pale in colour); or 3 (entire organ pale in colour with visible signs of necrosis). Scores for each parameter were added together to give a total score for each animal.
Parasite burden. Parasite burden was measured in the whole spleen of individual mice using a microtitre dilution method adapted from Buffet et al. 93 It was necessary to determine parasite burden in the spleen rather than the intestine because it was impossible to harvest immune cells for analysis from the intestine and determine parasite burden in the same animal, however, we have demonstrated previously that the parasite burden in the spleen accurately mirrors that in the intestine 50 . Briefly, prior to the experiment, 96 well plates were seeded with HFF cells and allowed to become confluent. One row was allocated per mouse and each mouse was done in duplicate. Spleens were removed and single-cell suspensions were made by passing through a 70-µm cell strainer. Cells were pelleted at 1500g, and then resuspended in RPMI 1,640 containing 5% FCS at a concentration of 1 × 10 7 cells/ml. Two hundred microliters of spleen cell suspension was added to the first well of a 96-well plate and then serially diluted 1/2 across the plate. Plates were incubated at 37 °C in 5% CO2 for 7 days before wells Ethics statement. All experiments were approved and conducted according to Australian animal protection law and in accordance with requirements of the Animal Ethics Committee of James Cook University (Approvals A2138, A2324). Death was never used as an endpoint.