Naïve CD8 T cell IFNγ responses to a vacuolar antigen are regulated by an inflammasome-independent NLRP3 pathway and Toxoplasma gondii ROP5

Host resistance to Toxoplasma gondii relies on CD8 T cell IFNγ responses, which if modulated by the host or parasite could influence chronic infection and parasite transmission between hosts. Since host-parasite interactions that govern this response are not fully elucidated, we investigated requirements for eliciting naïve CD8 T cell IFNγ responses to a vacuolar resident antigen of T. gondii, TGD057. Naïve TGD057 antigen-specific CD8 T cells (T57) were isolated from transnuclear mice and responded to parasite-infected bone marrow-derived macrophages (BMDMs) in an antigen-dependent manner, first by producing IL-2 and then IFNγ. T57 IFNγ responses to TGD057 were independent of the parasite’s protein export machinery ASP5 and MYR1. Instead, host immunity pathways downstream of the regulatory Immunity-Related GTPases (IRG), including partial dependence on Guanylate-Binding Proteins, are required. Multiple T. gondii ROP5 isoforms and allele types, including ‘avirulent’ ROP5A from clade A and D parasite strains, were able to suppress CD8 T cell IFNγ responses to parasite-infected BMDMs. Phenotypic variance between clades B, C, D, F, and A strains suggest T57 IFNγ differentiation occurs independently of parasite virulence or any known IRG-ROP5 interaction. Consistent with this, removal of ROP5 is not enough to elicit maximal CD8 T cell IFNγ production to parasite-infected cells. Instead, macrophage expression of the pathogen sensors, NLRP3 and to a large extent NLRP1, were absolute requirements. Other members of the conventional inflammasome cascade are only partially required, as revealed by decreased but not abrogated T57 IFNγ responses to parasite-infected ASC, caspase-1/11, and gasdermin D deficient cells. Moreover, IFNγ production was only partially reduced in the absence of IL-12, IL-18 or IL-1R signaling. In summary, T. gondii effectors and host machinery that modulate parasitophorous vacuolar membranes, as well as NLR-dependent but inflammasome-independent pathways, determine the full commitment of CD8 T cells IFNγ responses to a vacuolar antigen.


65
Toxoplasma gondii is a globally spread intracellular parasite that can infect nearly all warm-blooded 66 vertebrates, including humans. Transmission between hosts occurs following ingestion of oocysts shed 67 from the definitive feline host or predation of chronically infected animals harboring infectious 'tissue 68 cysts'. Immune modulation by the parasite during the first weeks of infection is therefore critical for T.

69
gondii to establish latency and lifecycle progression. The parasite accomplishes this by hiding and 70 manipulating the immune system from within a specialized parasitophorous vacuole (PV) that is created 71 during invasion. T. gondii releases 'effector' proteins from secretory organelles, including rhoptry proteins 72 (ROP) that are injected into the host cytosol upon invasion, as well as dense granules (GRA) that are 73 secreted into the lumen of the PV and aid its internal structure and formation. Many of these secreted 74 'effectors' manipulate host cell signaling pathways and shield the PV from host immune attack (1). In mice,

256
Due to the conserved nature of the TGD057 peptide epitope (Fig S1), a unique opportunity arose to explore approximately 25% or less Irgb6-PV coating is observed (75), the CD8 T cell responses to these strains 276 differ dramatically (Fig 5A, S3B). Furthermore, one notable outlier among clade A strains is the relatively 277 high T57 response to BOF (Fig 5B, S3B). BOF encodes a single copy of ROP5B that is marginally 278 expressed (75). When BOF is complemented with the LC37 cosmid, which encodes the entire ROP5 locus 279 from the clade A type I genetic background, the CD8 T cell IFNγ response is largely reduced (Fig 5B). We 280 infer from these assays the clade A genetic background inhibits T57 IFNγ responses, but the identity of the 281 ROP5-host interacting partner and why this genetic background leads to repressed responses is currently 282 unknown.

Activation and IFNγ differentiation of CD8 T cells are only partially inhibited by T. gondii ROP5
284 Following antigen-driven TCR stimulation (or 'signal 1'), early activated T cells receive secondary cues 285 from the environment including co-stimulation ('signal 2') and cytokines ('signal 3') to commit to the 286 production of cytokines like IFNγ. Whether clade A strains, through ROP5 or other effectors, intersect one 287 or several these activation steps to lower T57 IFNγ responses is unclear. To explore this issue further, we  infected Nlrp3-/-BMDMs (Fig 8A-8B). In contrast, the IFNγ response was only partially decreased to Asc-348 /-, Casp1/11-/-, and Gsdmd-/-BMDMs infected with ME49 ( Fig 8A), and no consistent difference was 349 observed between knockout and wildtype BMDMs infected with MAS ( Fig 8B) credence to this hypothesis, in that the parasite's export machinery appears dispensable for inducing 382 TGD057-specific CD8 T cell responses (Fig 2). One curious observation in this regard, is that the TGD057 383 peptide epitope, SVLAFRRL, encodes the lone ASP5 recognition TEXEL motif (underlined) (Fig S1). In 384 fact, ASP5 cleavage would preferentially produce parasite peptides with a terminal leucine, which is the the opposite occurred (Fig 2), ruling against parasite-assisted antigen processing of TGD057. In addition to 389 its role in protein export, ASP5 is required for targeting dense granules to the PVM (59,60,113 GBPs may also assist in the MHC 1 antigen presentation, for which we found some evidence (Fig 3).

420
Although the role of p62 using mouse knockout cells was not directly tested, a comparison of p62-PV 421 localization patterns between stimulatory and non-stimulatory parasites strains ( Fig S4) argues against a 422 dominant role for this pathway in our system. Other differences include lessons learned from the GRA6 423 antigen. When the C-terminal GRA6 epitope is facing the host cytosol it is highly stimulatory to CD8 T 424 cells (42). The protruding nature of the GRA6 epitope into the host cytosol may bypass need for host 425 recruitment of IFNγ-induced IRG/GBP machinery, thus facilitating its immuno-dominance. However,

426
TGD057 is not an integral membrane protein nor is it associated with the membranous fractions of the PV 427 (41). It is therefore unclear how the initial antigen is first detected to start T57 IFNγ responses, which 428 paradoxically require IFNγ signaling to begin with (Fig 3). A clue may come from the OVA system. Host 429 derived ER-GICs fuse with the PVM in a Sec22b SNARE-dependent process to initiate MHC 1 presentation 430 of T. gondii expressed OVA (43). Whether this pathway seeds the initial antigen-specific response to 431 TGD057 is unknown. Yet even in response to clade A strains, which are poor inducers of TGD057-specific 432 CD8 T cell IFNγ and IL-2 responses (Fig 5, S3B), the early activation marker CD69 was readily detected 433 on T57 CD8 T cells (Fig 6). The immune system is therefore robust in its ability to perceive T. gondii 434 antigens, which employs multiple non-redundant pathways to acquire antigens from vacuolated pathogens 435 (116).

436
Third, our studies demonstrate an absolute requirement for the pathogen sensor NLRP3, and to a 437 similar extent NLRP1, for promoting naïve TGD057-specific CD8 T cell IFNγ responses to parasite-438 infected cells. Moreover, it appears NLR-mediated regulation can occur in the absence of other components 439 of the inflammasome cascade. This is inferred because T57 IFNγ production was still detected in response 440 to parasite-infected ASC, caspase-1/11, and gasdermin D deficient cells, or when IL-1/18 cytokine 441 signaling was inhibited. In contrast, when NLRP3 is removed, there was no IFNγ response (Fig 8), even 442 though the CD8 T cells were activated (Fig 9). NLRs have several inflammasome independent functions,

443
including the ability to form a bridge between ER and mitochondria to initiate inflammasome signaling 444 (117). NLRs also bind to and directly activate transcription factors, such as IRF4 (118)

693
The ∆asp5 and ∆myr1 T. gondii strains listed were assayed for host CD8 T cell response as previously 694 described in Fig 1A. The IFNγ response at 48h, as analyzed by ELISA, is normalized to that of the clade D

759
The T57 CD8 T cell IFNγ response to TGD057 was analyzed by ELISA and normalized to that of WT