Treatment of mice with IL2-complex enhances inflammasome-driven IFN-γ production and prevents lethal toxoplasmosis

Toxoplasmic encephalitis is an AIDS-defining condition in HIV+ individuals. 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 identify CD4-independent mechanisms to control acute T. gondii infection. Here we have investigated the targeted expansion and regulation of IFN-γ production by CD8+ T cells, DN T cells and NK cells in response to T. gondii infection using IL-2 complex (IL2C) pre-treatment in an acute in vivo mouse model. Our results show that expansion of CD8+ T cells, DN T cells and NK cell by S4B6 IL2C treatment increases survival rates of mice infected with T. gondii and this increased survival is dependent on both IL-12- and IL-18-driven IFN-γ production. Processing and secretion of IFN-γ-inducing, bioactive IL-18 is dependent on the sensing of active parasite invasion by multiple redundant inflammasome sensors in multiple hematopoietic cell types but independent from T. gondii-derived dense granule (GRA) proteins. Our results provide evidence for a protective role of IL2C-mediated expansion of CD8+ T cells, DN T cells and NK cells in murine toxoplasmosis and may represent a promising adjunct therapy for acute toxoplasmosis. Author Summary A third of the world’s population is chronically infected with the parasite Toxoplasma gondii. In most cases the infection is asymptomatic, but in individuals suffering from AIDS, reactivation of brain and muscle cysts containing T. gondii is a significant cause of death. The gradual decline of CD4 T cells, the hallmark of AIDS, is believed to be a major contributing factor in reactivation of T. gondii infection and the development of acute disease. In this study, we show that targeted expansion of non-CD4 immune cell subsets can prevent severe disease and premature death via increased availability of interferon gamma-producing immune cells. We also demonstrate that the upstream signaling molecule interleukin-18 is required for the protective immune response by non-CD4 cells and show that the sensing of active parasite invasion by danger recognition molecules is crucial. Our findings reveal that targeted cell expansion may be a promising therapy in toxoplasmosis and suggests that the development of novel intervention strategies targeting danger recognition pathways may be useful against toxoplasmosis, particularly in the context of AIDS.


24
Toxoplasmic encephalitis is an AIDS-defining condition in HIV + individuals. The decline of 25 IFN-γ-producing CD4 + T cells in AIDS is a major contributing factor in reactivation of 26 quiescent Toxoplasma gondii to an actively replicating stage of infection. Hence, it is important 27 to identify CD4-independent mechanisms to control acute T. gondii infection. Here we have 28 investigated the targeted expansion and regulation of IFN-γ production by CD8 + T cells, DN T 29 cells and NK cells in response to T. gondii infection using IL-2 complex (IL2C) pre-treatment 30 in an acute in vivo mouse model. Our results show that expansion of CD8 + T cells, DN T cells 31 and NK cell by S4B6 IL2C treatment increases survival rates of mice infected with T. gondii 32 and this increased survival is dependent on both IL-12-and IL-18-driven IFN-γ production. 33 Processing and secretion of IFN-γ-inducing, bioactive IL-18 is dependent on the sensing of [1]. It is estimated that one-third of the world's population is infected with T. gondii. In most 58 individuals, infection is asymptomatic and leads to chronic, life-long persistence of T. gondii-59 containing cysts, primarily in brain and muscle tissue [2]. Active disease, also known as 60 toxoplasmosis, usually occurs after reactivation of encysted parasites, and is often associated 61 with immunosuppression. If untreated, toxoplasmosis may be fatal. Additionally, serious eye 62 disease has been reported as a result of infection with T. gondii [3] and, if a primary infection 63 occurs during pregnancy, abortion, stillbirth and fetal abnormalities can occur [2,4]. Whereas 64 an acute infection is generally mediated by the fast-replicating tachyzoite stage of the parasite, cells and DN T Cells and up to 50% of all NK cells stained IFN-γ + , particularly following 134 inoculation with 40 and 100 cysts. Because these mice had never been exposed to apicomplexan 135 parasites before, these results ruled out antigen-specific responses.

137
We also investigated if rapid IFN-γ production could be induced by inoculation with tachyzoites 138 via the i.v. and i.p. routes using a short-term in vivo exposure model in which naïve B6 mice 139 were exposed to T. gondii tachyzoites for a maximum of 72 hours. When mice were injected 140 i.v. or i.p. with 10 5 tachyzoites, no significant IFN-γ production could be seen in either spleen, 141 MLN or PP within 72 hours (S1E Fig). However, i.v. or i.p. inoculation with 10 7 tachyzoites 142 led to secretion of IFN-γ by CD3 + CD8 + , CD3 + CD4 -CD8 -(DN) T cells and CD3 -NKp46 + cells 143 in spleen, MLN and PP as early as 2-24 hours after inoculation (Fig 1C, D and S1C, D Fig), 144 mirroring the results seen 5 days after a cyst inoculation (Fig 1B). Importantly, at 24 hours after 145 tachyzoite inoculation, levels of other acute inflammatory mediators, such as IL-6, TNFα and 146 IL-10, were almost indistinguishable from naïve mice (Fig 1E-G). These results indicate that Furthermore, these results show that i.v., i.p. tachyzoite infections and oral brain cyst infections 152 induce almost identical acute immune responses. Toxoplasma gondii cyst production in mice is 153 a slow and laborious process. In addition, it is difficult to quantify the number of bradyzoites 154 within brain cysts used for oral infection and, moreover, dissemination patterns following oral 155 infection are erratic in individual mice [33]. Therefore, we subsequently focused on IFN-γ 156 secretion by splenic NK cells 24 hours after i.v. injection of tachyzoites as our primary readout 160 Whereas the role of IL-12 in IFN-γ secretion is well established for T. gondii [2], rapid 161 production of IFN-γ in response to other intracellular pathogens, such as S. enterica, L. 162 monocytogenes and M. tuberculosis has also been linked to the upstream effects of 163 29]. To interrogate whether or not, and how early, IFN-γ secretion in response to T. gondii also 164 requires IL-18, we exposed naïve B6 mice to T. gondii ME49 tachyzoites and treated the 165 animals with neutralizing monoclonal antibodies (mAb) to IL-12, IL-18 or IL-12 and IL-18 166 immediately after inoculation. At 24 hours after exposure, IFN-γ secretion by NK cells in the 167 spleen was assessed directly ex vivo. Neutralization of IL-12 and IL-18 significantly reduced 168 IFN-γ production, with IL-12 contributing approximately 50% and IL-18 approximately 30-169 40% of the response (Fig 2A). The significant reduction of rapid IFN-γ production in Il18 -/-170 mice, and the almost complete absence of rapid IFN-γ production in anti-IL-12-treated Il18 -/-171 mice, further confirmed a direct correlation between IL-12, IL-18 and IFN-γ secretion (Fig 2C, 172 D). Consistently, where IL-12 levels in the serum of infected mice peaked at approximately 2 173 hours after inoculation, the levels of IL-18 mirrored those of IFN-γ for up to 72 hours (Fig 2B). 174 Furthermore, treatment with anti-IL-12 and/or anti-IL-18 also reduced concentrations of IFN-175 γ, IL-12 and IL-18 in the serum of infected mice in an additive manner (Fig 2D-F). These results 176 suggest a hierarchical relationship in which a primary IL-12-driven IFN-γ response is followed 177 by an IL-18-dominant IFN-γ response. We concluded that innate IFN-γ secretion by CD8 + T  Given that the molecular mechanisms that lead to T. gondii-mediated IL-12 secretion are well 183 characterized, we focused our attention on the host signaling pathways required for IL-18-184 driven IFN-γ production, using a panel of genetically modified mouse strains. Secretion of 185 bioactive IL-18 depends on the enzymatic cleavage of pro-IL-18 by caspase-1 [23]. Activation 186 of caspase-1 involves the sensing of danger molecules or stress signals via upstream cytosolic 187 PRRs, so called inflammasomes, a process that can be enhanced and controlled via TRIF-188 dependent caspase-11 activation. Caspase1/11 -/double KO mice produced significantly less 189 IFN-γ following injection with T. gondii ME49 tachyzoites compared with B6 mice, and this 190 response could be almost completely prevented by additional anti-IL-12 treatment (Fig 3A). As 191 expected, Caspase1/11 -/mice did not secrete significant levels of IL-18 following T. gondii 192 inoculation (Fig 3B), indicating that the remaining IFN-γ response in Caspase1/11 -/mice is 193 driven by . Surprisingly, when we tested mice deficient in the upstream NLR family pyrin 194 domain-containing proteins 1 and 3 (NLRP1 and NLRP3), NLR molecules that had been 195 implicated previously in recognition of T. gondii [24], both knockout strains secreted 196 indistinguishable amounts of IL-18 compared with B6 mice (Fig 3B). This data suggested a 197 redundant role for NLRP1 and NLRP3. However, even double knockout and heterozygous 198 Nlrp1 ±/-Nlrp3 ±/mice secreted high levels of IL-18 and IFN-γ after exposure to T. gondii ME49 199 tachyzoites (Fig 3A, B), suggesting that additional PRR molecules must be involved in sensing 200 of T. gondii invasion in vivo. Taken together these results indicate that rapid IFN-γ secretion 201 in vivo in response to T. gondii depends on the inflammasome  caspase-1  IL-18 axis, and 202 that T. gondii activates at least three different inflammasomes in vivo.

204
Toxoplasma gondii activates inflammasomes in multiple cell types. 205 To further investigate the role of cytosolic PRRs in sensing T. gondii invasion, and to potentially 206 target inflammasome activation for preventive or therapeutic intervention strategies, we next 207 tried to identify the T. gondii-sensing cell type in vivo. To do this, we made use of a red 208 fluorescent protein (RFP) tagged T. gondii ME49 (T. gondii ME49-RFP) strain to track parasite 209 uptake by different immune cell subsets in the spleen. Twenty-four hours after tachyzoite 210 injection, T. gondii ME49-RFP also induced rapid IFN-γ secretion by splenic CD3 + CD4 + , 211 CD3 + CD8 + , CD3 + CD4 -CD8 -(DN) T cells and CD3 -NKp46 + cells (Fig 4A) and high levels of 212 serum IL-18 (Fig 4B), similar to wild-type T. gondii ME49 (see Figs. 1 and 2). Approximately 213 0.5% of all splenocytes contained T. gondii ME49-RFP in vivo 24 hours after inoculation (Fig   214   4C). Sorted RFP + cells secreted significantly more IL-18 ex vivo compared to RFPcells (Fig   215   4D), and further surface phenotyping revealed that T. gondii ME49-RFP was primarily 216 contained in monocytes, neutrophils and CD8α + dendritic cells (Fig 4E, F)  parasites induced IFN-γ secretion and increased serum IL-18 levels (Fig 5A, B). To exclude 236 the possibility that heat inactivation and sonication destroyed soluble factors that could 237 potentially drive this response, we also injected naïve B6 mice with HFF cell debris, which had 238 been re-suspended in the T. gondii ME49 culture supernatant. This treatment also failed to 239 induce IFN-γ and IL-18 secretion (Fig 5A, B). These results indicated that active parasite 240 invasion is required to initiate an IFN-γ response, suggesting that T. gondii virulence factors  (Fig 5C, D), suggesting that ASP5-driven GRA export is dispensable for inflammasome 256 activation. Similarly, inoculation with GRA20-deficient or GRA23-deficient parasites did not  gondii ME49 ASP5-deficient parasites, inoculation with T. gondii DEG did not lead to reduced 261 levels of serum IL-18 and NK cell-produced IFN-γ in this model (Fig 5C, D). At 48 hours after 262 tachyzoite inoculation, the levels of serum IL-18 were even significantly higher compared with 263 inoculation of T. gondii ME49 (Fig 5D). These data indicate that ASP5-dependent secretion of 264 GRA proteins does not affect IL-18-driven IFN-γ secretion and highlights the diverging 265 mechanisms that underlie in vitro IL-1β and in vivo IL-18 secretion in response to T. gondii. 266 267 IL2C treatment expands IL-18-responsive IFN-γ-secreting cell subsets 268 Collectively, the results presented so-far raise the prospect that, if the ability of non-CD4 cells 269 to invoke inflammasome-dependent, IL18-driven production of IFN-γ can be enhanced, it may 270 be possible to control acute toxoplasmosis in AIDS. Hence, we investigated if targeted 271 expansion of non-CD4 cells with IL2C treatment can achieve this. First, naïve mice were treated 272 i.p. with IL2C complex on four consecutive days (Fig 6A) and, 24 hours after the last IL2C 273 injection, immune cell expansion was assessed by flow cytometry relative to untreated animals.

274
As reported previously [38], IL2C treatment led to a significant expansion of memory CD8 + T 275 (Fig 6B, C) and to a minor increase in the 276 Peyer's Patches (PP) (Fig 6D).To further assess if IL2C-expanded and non-expanded CD8 + T  We also assessed the expression of IL18R and IL12R on the surface of IFN-γ + and IFN-γcells.

IFN-γ + NK cells (data for CD8 + T cells and DN T cells not shown) expressed significant higher
287 levels of IL18R and IL12R compared to IFN-γ -NK cells (Fig 6H, I). Taken together, these 288 results show that IL2C-expanded cells respond identically to non-expanded cells and that IL2C 289 treatment numerically expands IFN-γ producing cells that maintain a higher IL18R level 290 expression compared to IFN-γcells. To definitively assess if IL2C-mediated expansion of IL-18-responsive IFN-γ-secreting non-295 CD4 cell subsets can prevent lethal toxoplasmosis in mice, we used the well-established oral 296 inoculation model with T. gondii ME49 bradyzoite-containing brain cysts. As above, naïve B6 297 mice were treated i.p. with IL2C for four consecutive days (Fig 7A). IL2C treatment was 298 accompanied by a weight loss from which mice recovered within a few days (data not shown).

299
Forty-eight hours after the last IL2C treatment, mice were inoculated orally with 10 or 40 T. 300 gondii ME49 cysts and were assessed for weight loss and survival over 60 days. All mice that 301 had been inoculated with 40 cysts and 87% of mice that had been inoculated with 10 cysts, but 302 had not received IL2C injections, succumbed within 14 days after inoculation (Fig 7B, C). In the protective phenotype (Fig 7D, F). All mice that were not treated with IL2C succumbed to 313 the infection by day 16, with a median survival of 11 days (Fig 7F). Whilst 67% of IL2C-treated 314 mice that received control rat IgG survived until day 60, the median survival for mice treated 315 with anti-IFN-γ was 10.5 days, 13 days for mice treated with anti-IL-12 and 14 days for mice 316 treated with anti-IL-18 (Fig 7F). All mice that survived until day 60 were assessed for T. gondii 317 brain cysts. Mice contained 100 -200 cysts per brain (data not shown), indicating that all mice 318 were infected and that survival was not due to a failure of the infection to establish. Taken 319 together, these results further substantiate the proposal that IL2C pre-treatment protects mice 320 from lethal toxoplasmosis via IL-12-and IL-18-driven IFN-γ secretion.

322
To assess if IL2C pre-treatment also impacts on measurable disease parameters other than 323 survival, we also analyzed parasite burden and immunopathology at 2, 4 and 9 days following 324 oral cyst infection. Due to the low infectious dose of 10 cysts, only minimal changes in 325 immunopathology were observed at 2 and 4 days after infection in all groups (data not shown).

326
At 9 days after infection, IL2C pre-treated mice displayed significantly reduced gross pathology 327 of gut and liver (Fig 7G, H) in the absence of any effect on parasite burden (Fig 7I). TReg Nlrp3 -/and Nlrp1 ±/-Nlrp3 ±/mice were devoid of circulating IL-18 after T. gondii infection, a 341 third sensor must exist in addition to NLRP1 and NLRP3 [24,42]. Furthermore, we show that 342 inflammasome activation occurred in CD8α + DCs, inflammatory monocytes and neutrophils, 343 cell types that have also been implicated in IL-12 secretion in response to T. gondii [2]. These 344 results imply a high level of redundancy in the cell type that senses T. gondii infection as well 345 as in the host inflammasome signaling pathway. This is in contrast to the often very specific 346 recognition of viral and bacterial infections by one particular inflammasome in a distinct cell 347 type [28,29,[43][44][45][46]. It is likely that this divergence highlights the evolutionary complexity of 348 parasites and suggests that more highly evolved organisms have developed a more complex  Furthermore, in vitro activation of inflammasomes differs between T. gondii strains, and is 368 predominantly induced by Type II parasites [24]. These findings suggest that T. gondii has 369 evolved sophisticated diverging effector mechanisms to manipulate inflammasome biology in 370 different host cell subsets, and suggest that secreted effector molecules and/or distinct structural 371 proteins may underlie inflammasome activation. It is, therefore, interesting that Nlrp1 ±/-Nlrp3 ±/-372 mice did not show reduced IL-18 secretion after infection with T. gondii. It is important to note 373 that in mice the Nlrp1 locus is on the same chromosome as the Nlrp3 gene, meaning that the 374 generation of rare double knockout offspring relies on recombination rather than inheritance. It 375 will therefore be important to further investigate the role of Nlrp1 and 3 with alternative 376 methods, such as CRISPR/Cas9 and/or chemical inhibition.

378
Our study has ruled out ASP5-dependent GRA proteins [54], the most abundant family of T. It is tempting to speculate that the overall purpose of activating multiple inflammasomes in 389 multiple cell types is to drive an inflammatory host response that mediates the progression of

1003
Statistical analyses: One-way ANOVA followed by Dunnett's multiple comparison test; enumerated 24 hours later, and serum IL-18 (B) levels were measured. Statistical analyses: 1108 One-way ANOVA followed by Dunnett's multiple comparison test; significant differences are 1109 indicated by asterisks: * p<0.05; ** p<0.01; *** p<0.001; n.s. not significant. [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 1 2 / 1 8