Caspase-8 expression in CD8⁺ T cells promotes pathogen restriction in the brain during Toxoplasma gondii infection

Abstract

Cell death is an integral restriction mechanism against intracellular pathogens. We have previously reported extensive cell death in the brain during infection with the intracellular parasite, Toxoplasma gondii. Here, we focus on the role of caspase-8, a regulator of extrinsic apoptosis, during T. gondii infection. We find that Casp8^−/−Ripk3^−/− mice have increased brain parasite burden in comparison to controls and succumb to infection despite the generation of robust immune responses. We observed that neurons, astrocytes, and CD8⁺ T cells had high rates of parasite interactions in Casp8^−/−Ripk3^−/− mice compared to wild-type mice. While Casp8 deficiency in neurons and astrocytes did not affect control of infection, deletion of Casp8 in CD8⁺ T cells led to impaired survival, increased parasite burden, and direct infection of CD8⁺ T cells in the brain. We conclude that in addition to well-characterized effector functions, CD8⁺ T cells use caspase-8 to control T. gondii in the brain.

CD8⁺ T cells lacking caspase-8 harbor T. gondii and drive increased brain parasite burden despite robust effector function.

INTRODUCTION

Various forms of cell death contribute to immunity to intracellular pathogens (1–4). Understanding the role that cell death plays in pathogen defense in the brain is important due to the long-lived nature and minimal regenerative capacity of neurons. One important regulator of cell death is caspase-8. Caspase-8 has been classically defined as an initiator of noninflammatory extrinsic apoptosis, but more recent work has demonstrated an additional role for caspase-8 in promoting cytokine production and inflammation (5–9). While it is extremely rare, caspase-8 deficiencies have been reported in patients and linked to an increased susceptibility to bacterial and viral infections (10). Studying the role of caspase-8 in murine models of infection was initially limited because of the embryonic lethality of Casp8 deletion in mice. However, the codeletion of Casp8 and Ripk3 allows for viable mouse lines (11, 12). Since this advancement, several studies have tested the role of caspase-8 during infection. During Yersinia pestis infection, caspase-8 limits bacterial load primarily through the regulation of inflammation (13). In other infections, caspase-8 plays a more limited role (12, 14–19). Extensive cell death has been reported in the brain in response to Toxoplasma gondii infection (20), but whether caspase-8 regulates pathogen control in a cell-type–specific context in the brain has not been explored.

T. gondii is an obligate intracellular parasite that disseminates throughout the entire body of the host and takes up chronic residence in the brain and other immune-privileged tissues (21, 22). During the chronic stage of infection, the dormant form of the parasite, bradyzoites, are typically found in neurons, although all nucleated cell types can be infected in vitro (23–26). Prior studies have demonstrated that neurons are the primary cell type with sustained infection, perhaps because they are long-lived and lack the ability to clear intracellular parasites (24). Although the exact reason why T. gondii predominantly persists in neurons has yet to be completely elucidated, one hypothesis is that other cell types in the brain are infected but undergo cell death (24).

Throughout the course of infection, continual immune activity is required to control the infection and prevent lethality (27). Bradyzoites in cysts can revert back into replicating tachyzoites, a process called reactivation, where infective parasites are able to invade new host cells (27). During chronic brain infection, infiltrating T helper cells (T_H1 cells) produce interferon-γ (IFN-γ), a critical cytokine for the control of T. gondii that contributes to antiparasitic activation of glia and infiltrating immune cells (28–35). IFN-γ induces the expression of a variety of genes involved in immune cell activation and intracellular parasite killing through immunity-related guanosine triphosphatases and guanylate-binding proteins (36, 37). One important IFN-γ–mediated mechanism of parasite restriction involves the up-regulation of inducible nitric oxide synthase (iNOS) by brain-infiltrating inflammatory monocytes and macrophages (38–45). Beyond this, recruited macrophages and resident microglia, astrocytes, and endothelial cells secrete cytokines and/or chemokines to trigger immune cell infiltration into the brain, which assists in killing of the parasite (46, 47). For example, recruited CD8⁺ cytotoxic T lymphocytes mediate parasite killing through perforin-stimulated cell lysis (48, 49).

The role of cell death during T. gondii infection is a current topic of exploration. Batista et al. (20) identified numerous regions of dying cells in the brain at 4 weeks postinfection (wpi) and demonstrated that the cell death pathway known as pyroptosis is essential for parasite restriction during chronic brain infection. Follow-up studies have shown that Cx3cr1-expressing cells use caspase-1 to reinforce CD4⁺ T cell IFN-γ production during acute infection for optimal parasite control (50). Other studies have explored receptor-interacting protein kinase 3 (RIPK3) during the acute phase of infection but did not find evidence for necroptosis in parasite control (51). Work by DeLaney et al. (14) found that Casp8^−/−Ripk3^−/− mice infected with the Prugniard (Pru) strain of T. gondii succumb early to infection, while wild-type (WT) and Ripk3^−/− mice persist into the chronic stage. In addition, early work on T. gondii showed that one specific activator of caspase-8, Fas, is important in restricting ocular T. gondii infection, although its role in mediating parasite restriction in the brain has not been explored (52). Together, these studies have begun to demonstrate that cell death can be protective during chronic T. gondii infection.

In this study, we explored the role of caspase-8 during chronic brain infection with T. gondii. We found that Casp8^−/−Ripk3^−/− mice infected with the Me49 strain of T. gondii survive to the chronic stage of infection but succumb by 6 wpi, while Casp8^+/−Ripk3^−/− and C57BL/6 control mice survive for several months. The Casp8^−/−Ripk3^−/− mice had no observed immune deficits yet displayed an eightfold increase in brain parasite burden at 4 wpi compared to control groups. Using a cre-reporter system that labels host cells that have interacted with the parasite, we observed a large population of parasite-interacted cells in the brain in the absence of caspase-8. We identified neurons, astrocytes, and brain-infiltrating CD8⁺ T cells as the main cell types to have an increase in the number of parasite interactions in Casp8^−/−Ripk3^−/−Ai6 mice compared to WT^Ai6 control mice. Using cell-type–specific knockouts of Casp8, we found that Casp8 deficiency in CD8⁺ T cells, but not neurons or astrocytes, led to a significant increase in parasite burden at 6 wpi compared to controls despite a comparable T_H1 cell immune response. We also examined the role of Fas, an activator of caspase-8. We found an increase in parasite burden at 6 wpi and a strong T_H1 cell immune response in Fas^lpr mice. Overall, this work demonstrates that in addition to the well-characterized effector functions, CD8⁺ T cells use caspase-8 to mediate pathogen resistance in the brain during T. gondii infection.

RESULTS

Caspase-8 deficiency limits control of T. gondii infection in the brain despite a robust T_H1 cell response

To explore the role of caspase-8 during a chronic brain infection, we infected Casp8^−/−Ripk3^−/−, Casp8^+/−Ripk3^−/−, and C57BL/6 WT mice intraperitoneally with 10 cysts of the Me49 type II strain of T. gondii. Casp8 deficiency is embryonically lethal due to unrestricted necroptosis; however, codeletion of Ripk3 rescues lethality (12). For this reason, all mice with Casp8 deletion in this study were also deficient in Ripk3, and the statistical analyses compare Casp8^−/−Ripk3^−/− mice to both WT and Casp8^+/−Ripk3^−/− control mice. At 4 wpi, WT and Casp8^+/−Ripk3^−/− mice had no significant difference in brain cyst burden. By contrast, the Casp8^−/−Ripk3^−/− mice had an eightfold increase in the number of tissue cysts compared to WT mice and a sevenfold increase compared to Casp8^+/−Ripk3^−/− mice, demonstrating that Casp8, but not Ripk3, affects brain parasite burden (Fig. 1A). Survival was also affected, as the Casp8^−/−Ripk3^−/− mice succumbed to the infection by 6 wpi (Fig. 1B). Hematoxylin and eosin (H&E) staining of infected WT mouse brain parenchyma demonstrated minimal parasite presence [single cysts per field of view (FOV)] and immune cells present in the blood vessels (Fig. 1C). Notably, the brains of Casp8^−/−Ripk3^−/− mice had clusters of cysts and extensive perivascular cuffing, indicating considerable immune cell recruitment to the brain (Fig. 1D). The classically defined mechanisms of parasite control in the brain include T cell production of IFN-γ and tumor necrosis factor–α (TNFα) and infiltrating inflammatory monocyte production of nitric oxide via iNOS (27). We examined the immune responses in the brain at 4 wpi using flow cytometry (fig. S1A). In the Casp8^−/−Ripk3^−/− mice, we observed an increase in the number of CD8⁺ T cells compared to WT mice, and an increase in CD8⁺IFN-γ⁺ T cells, and CD8⁺TNFα⁺ T cells compared to both control groups (Fig. 1, E to G). In addition, we looked at markers of cytotoxicity in CD8⁺ T cells and found an increase in the number of CD8⁺FasL⁺ T cells and CD8⁺GranzymeB⁺ T cells compared to controls (Fig. 1, H and I). Furthermore, we found an increase in CD4⁺ T cells, CD4⁺IFN-γ⁺ T cells, CD4⁺TNFα⁺ T cells, and Ly6C^hiiNOS⁺ inflammatory monocytes in the Casp8^−/−Ripk3^−/− mice compared to WT mice (Fig. 1, J to M). Measuring expression of critical mediators of host protection in whole-brain homogenate by reverse transcription quantitative polymerase chain reaction (RT-qPCR), we observed increased expression of Ifng, Tnf, and Nos2 in brains of Casp8^−/−Ripk3^−/− mice compared to control mice (Fig. 1N). We observed small increases in T cells and iNOS⁺ monocyte number in Casp8^+/−Ripk3^−/− mice, which did not translate to overall increases in Ifng, Tnf, or Nos2 mRNA in these mice, further demonstrating that Ripk3 has minimal impact on the immune response to T. gondii in the brain in comparison to Casp8.

Fig. 1. Caspase-8 is essential for controlling T. gondii infection in the brain, but Casp8 deficiency did not affect the T_H1 cell immune response.

Casp8^−/−Ripk3^−/−, Casp8^+/−Ripk3^−/−, and WT C57BL/6 mice were infected with 10 cysts of the Me49 strain of T. gondii intraperitoneally and analyzed at 4 wpi. (A) Cyst burden was quantified by light microscopy. (B) Survival curve of infected WT, Casp8^+/−Ripk3^−/−, and Casp8^−/−Ripk3^−/− mice. Representative H&E image of (C) WT and (D) Casp8^−/−Ripk3^−/− brain at 4 wpi. Arrows indicate inflamed blood vessels, and arrowheads indicate T. gondii cysts. Scale bars, 50 μm. Brain immune cell populations were quantified by spectral flow cytometry at 4 wpi: (E) CD8⁺ T cells, (F) CD8⁺IFN-γ⁺ T cells, (G) CD8⁺TNFα⁺ T cells, (H) CD8⁺FasL⁺ T cells, (I) CD8⁺GranzymeB⁺ T cells, (J) CD4⁺ T cells, (K) CD4⁺IFN-γ⁺ T cells, (L) CD4⁺TNFα⁺ T cells, and (M) infiltrating iNOS⁺ monocytes (CD45^hiCD11b⁺Ly6G⁻Ly6C^hiNOS⁺). (N) Gene expression was measured by RT-qPCR for Ifng, Tnf, and Nos2. Statistical significance determined by randomized block analysis of variance (ANOVA) and least-squares means: (A) WT C57BL/6 (n = 16), Casp8^+/−Ripk3^−/− (n = 14), and Casp8^−/−Ripk3^−/− (n = 22) (four experiments); (E, J, and M) WT C57BL/6 (n = 15), Casp8^+/−Ripk3^−/− (n = 9), and Casp8^−/−Ripk3^−/− (n = 14) (three experiments); (F and K) WT C57BL/6 (n = 15), Casp8^+/−Ripk3^−/− (n = 8), and Casp8^−/−Ripk3^−/− (n = 18) (three experiments); (N) WT C57BL/6 (n = 10), Casp8^+/−Ripk3^−/− (n = 10), and Casp8^−/−Ripk3^−/− (n = 14) (two experiments). Statistical significance determined by log-rank (Mantel-Cox) test: (B) WT C57BL/6 (n = 10), Casp8^+/−Ripk3^−/− (n = 8), and Casp8^−/−Ripk3^−/− (n = 9) (two experiments). Statistical significance determined by ordinary one-way ANOVA: (G, I, and L) WT C57BL/6 (n = 5), Casp8^+/−Ripk3^−/− (n = 5), and Casp8^−/−Ripk3^−/− (n = 9) (one experiment). (H) WT C57BL/6 (n = 5), Casp8^+/−Ripk3^−/− (n = 6), and Casp8^−/−Ripk3^−/− (n = 5) (one experiment). Data are presented as mean ± SEM; *P < 0.05, **P < 0.01, and ***P < 0.001.

Notably, Casp8^−/−Ripk3^−/− mice develop extensive lymphadenopathy and splenomegaly partially due to a population of T cells that are CD4⁻CD8⁻B220⁺ and emerge with age (fig. S1B) (12). T cells have been observed to up-regulate B220 expression before undergoing apoptosis (53). At 4 wpi, there was a large CD3⁺CD4⁻CD8⁻B220⁺ population in the spleens of the Casp8^−/−Ripk3^−/− mice that was minimal in the Casp8^+/−Ripk3^−/− and WT mice (fig. S1C). In the brain at 4 wpi, the population of CD3⁺CD4⁻CD8⁻B220⁺ cells was comparable between the three groups and more than 10-fold smaller than the populations of both CD4⁺ and CD8⁺ T cells, demonstrating that this population makes up only a minor population (<5%) of the T cells in the brain during infection (fig. S1, D and E). However, there was an increase in the number of CD3⁺CD4⁻CD8⁻B220⁺IFN-γ⁺ cells in the brains of Casp8^−/−Ripk3^−/− mice compared to both controls at 4 wpi (fig. S1F). Overall, we observed an amplified T_H1 cell immune response and CD8⁺ T cell cytotoxicity markers in the brains of Casp8^−/−Ripk3^−/− mice in comparison to controls. This finding was unexpected, as higher parasite burdens are often driven by a deficiency in activation and/or trafficking of peripheral immune cell populations into the infected brain (30, 42, 43, 54).

Having observed an increased brain parasite burden but an intact immune response in the Casp8^−/−Ripk3^−/− mice, we explored whether an immune defect during acute infection could be contributing to the increased parasite burden in the central nervous system. Previous studies have explored the role of caspase-8 in cytokine production during acute T. gondii infection and proposed a role for caspase-8 in promoting dendritic cell (DC) interleukin-12 (IL-12) production (14). During the acute phase of infection, DCs produce IL-12, which is critical for T cell production of IFN-γ, a main mediator of parasite killing and control (55–57). At 5 days postinfection (5 dpi), we found no differences in the number of Ly6C^hi monocytes or IL-12⁺CD11c⁺MHCII⁺ DCs by number or frequency in peritoneal exudate cells (PECs) (fig. S2, A to C and E to H). In addition, we found a significant increase in IL-12p40 in the serum of the Casp8^−/−Ripk3^−/− mice compared to WT mice (fig. S2D). At 8 dpi, we observed equivalent levels of serum IFN-γ (fig. S2I), suggesting that T cell responses were intact. In the PEC population, we found an increase in the number of CD8⁺ and CD4⁺ T cells and equivalent numbers of CD11b⁺Ly6C^hi inflammatory monocytes between the Casp8^−/−Ripk3^−/− and WT mice (fig. S2, J to L), demonstrating a robust immune response during the acute phase of infection. Despite the intact immune response, we did find that caspase-8 deficiency led to impaired parasite restriction at 8 dpi, demonstrated by a significant increase in parasite DNA in the brain, heart, lung, and liver of the Casp8^−/−Ripk3^−/− mice compared to the two control groups (fig. S3, A to D). At 4 wpi, there was a large increase in parasite DNA remaining in the brain of Casp8^−/−Ripk3^−/− mice, but parasite burden in the heart, lung, and liver was markedly reduced compared to the levels observed at 8 dpi and was equivalent in all three experimental groups (fig. S3, E to H), suggesting that parasites are cleared effectively in the periphery of Casp8^−/−Ripk3^−/− mice at later time points. Overall, these results demonstrate that Casp8^−/−Ripk3^−/− mice mount a robust T_H1 cell immune response throughout the course of infection, but caspase-8 is necessary for parasite restriction particularly in the brain.

Casp8 is highly expressed in immune cell populations during chronic T. gondii infection

Cell death of differentiated cells is a rare occurrence in the healthy brain. However, prior studies have demonstrated increased Casp8 expression in various cell types in the brain during multiple neuroinflammatory conditions (58–64). Caspase-8 can be activated by ligation of death receptors, which induces recruitment of proteins that scaffold on the death receptor. Within the scaffold, caspase-8 homodimerization allows for self-cleavage between catalytic subunits and release of the active caspase-8 enzyme (5, 6, 65). However, cells with active caspase-8 often undergo cell death before they can be identified in vivo. Thus, identifying cell populations that express Casp8 allows for an assessment of cell populations that could potentially use caspase-8. For this reason, we sought to identify which cell types express Casp8 in the brain using multiplexed error-robust fluorescence in situ hybridization (MERFISH) technology. This technology was used to measure spatially resolved transcripts in individually segmented nuclei in whole-brain slices. The benefit of using this technology is that it allowed us to perform transcriptomics on neuronal and glial cell populations that are otherwise difficult to isolate using standard brain single-cell purification processes. In addition, this assay avoids inducing an extensive stress response in cells. We designed a 342-gene panel that included housekeeping genes, genes to differentiate cell types, and genes related to inflammation and cell death (sheet S1). In naïve mouse brains, we identified 12 cell populations (n = 55,399 cells, 32,247 transcripts per FOV) (Fig. 2A and fig. S4, A to C). We mapped normalized z-score expression of Casp8 to the identified cell populations and quantified normalized z-score expression of cell death–related genes (Fig. 2, B and C). We found that in a naïve mouse brain, Casp8 is predominantly expressed in microglia, which is consistent with the literature (66). During chronic T. gondii infection, there was a robust increase in the population of macrophages and T cells due to immune cell infiltration into the brain (n = 58,479 cells, 29,758 transcripts per FOV) (Fig. 2D and fig. S4, D to F). Mapping normalized z-score expression of Casp8 to the cell populations demonstrated that Casp8 is predominantly expressed in immune cell populations (Fig. 2E). Furthermore, we identified Casp8 as most highly expressed in infiltrating CD8⁺ and CD4⁺ T cells, infiltrating macrophages, and microglia with a lower degree of expression in endothelial cells, the choroid plexus, and pericytes (Fig. 2F). Neuronal populations, oligodendrocytes, oligodendrocyte precursor cells, and astrocytes had comparably low levels of Casp8 expression (Fig. 2F). Thus, during T. gondii infection, Casp8 is most highly expressed in brain-infiltrating and brain-resident immune cells. To measure the active form of caspase-8 (cleaved caspase-8), we used immunohistochemistry (IHC) in chronically infected WT mouse brain sections (Fig. 2G). We used chronically infected Casp8^−/−Ripk3^−/− mouse brain sections and confirmed a lack of cleaved caspase-8 staining. In WT brain sections at 4 wpi, we were able to visualize cleaved caspase-8 in the brain (Fig. 2, H to J). Cells with cleaved caspase-8 staining were rare (average one cell per four sagittal sections) and often very small, likely due to the activation of downstream caspases and the induction of apoptosis. To increase the amount of cleaved caspase-8 we could measure by IHC, we inhibited caspase-3 using Ac-DEVD-CHO. Beginning at 4 wpi, we injected mice with 3 mg/kg of Ac-DEVD-CHO intraperitoneally a total of five times over the course of 2 weeks with the final treatment the day before analysis (Fig. 2K). At this time, we were able to identify more cleaved caspase-8⁺ cells and clusters of cleaved caspase-8⁺ cells; however, these cells were still relatively uncommon (average of three cleaved caspase-8⁺ cells per sagittal section) (Fig. 2, L to Q). Here, we demonstrate caspase-8 expression and activity in the brain during chronic T. gondii infection.

Fig. 2. Casp8 is highly expressed in infiltrating immune cells during chronic T. gondii infection.

MERFISH was used to measure spatially resolved transcriptomics at a single nuclei level in whole-brain slices of WT C57BL/6 mice. A 342-gene panel was used to call cell types and measure the expression of numerous inflammation and cell death–related genes. (A) Nuclei from uninfected brains were annotated through Leiden clustering and differential expression testing of marker genes by Wilcoxon rank sum test (n = 55,399 cells, 32,246.9 transcripts per FOV), and cell populations were identified. OPC, oligodendrocyte precursor cell. (B) Casp8 expression was mapped to the cell populations. (C) Expression of genes associated with cell death was analyzed by cell population. (D to F) Cell clustering, Casp8 expression mapping, and gene expression were similarly examined in the infected brain at 4 wpi (n = 58,479 cells, 29,757.7 transcripts per FOV). (G) Outline of experiments in WT mice. (H) Confocal microscopy was used to image cells with cleaved caspase-8 (magenta) with 4′,6-diamidino-2-phenylindole (DAPI) (blue). (I) Magnified image of cleaved caspase-8 (magenta) with DAPI (blue) and (J) cleaved caspase-8 (magenta) signal alone in WT mice at 4 wpi. (K) Outline of experiments using caspase-3 inhibition in WT mice. (L to Q) Confocal microscopy of cleaved caspase-8 (magenta) with DAPI (blue) and magnified images of WT mice treated with caspase-3 inhibitor. Scale bars, 50 μm.

Neurons, astrocytes, and CD8⁺ T cells have increased interactions with T. gondii in Casp8-deficient mice

We next sought to understand whether caspase-8 plays an essential role in restricting T. gondii infection in specific cell populations. To identify the cell types that encounter parasite in Casp8-deficient mice, we used a Pru strain of T. gondii in which the cre recombinase gene is fused to the toxofilin gene (TF-Pru). Toxofilin is a rhoptry protein that is secreted into host cells upon parasite invasion. During infection with TF-Pru parasites, cre recombinase is injected into host cells (23). We crossed Casp8^−/−Ripk3^−/−, Casp8^+/−Ripk3^−/−, and WT mice onto a ROSA26^Ai6/Ai6 background, a Cre reporter allele designed to drive green fluorescent protein variant (ZsGreen1) (67). As the parasite interacts with a cell, parasite-driven cre activity will induce ZsGreen fluorescent labeling. At 4 wpi, WT^Ai6 and Casp8^+/−Ripk3^−/−Ai6 mice had sparse ZsGreen⁺ cells, and most were neurons, consistent with previous reports (Fig. 3A and fig. S5A) (24, 68, 69). At a site of reactivation where immune cells cluster and parasites invade new cells, we found no ZsGreen⁺ cells in the WT^Ai6 and Casp8^+/−Ripk3^−/−Ai6 mice (Fig. 3B and fig. S5B), suggesting that immune cells infected during a reactivation event may not be long lived. On the other hand, the Casp8^−/−Ripk3^−/−Ai6 mice demonstrated a marked increase in the number of ZsGreen⁺ cells, with the predominant populations being neurons [neuron-specific nuclear protein⁺ (NeuN⁺); Fig. 3C, inset, and fig. S5, C to E], astrocytes [glial fibrillary acidic protein⁺ (GFAP⁺); Fig. 3D, inset, and fig. S5, F to H], and a population of cells morphologically consistent with immune cells [Fig. 3, E (inset) and C to F]. Populations of ZsGreen⁺ cells visualized via IHC in the Casp8^−/−Ripk3^−/−Ai6 mice were highly variable; in some locations, we identified up to 75 ZsGreen⁺ neurons, while in other locations, there were numerous ZsGreen⁺ glia and immune cells. In FOV with ZsGreen⁺ cells, on average, there were 16 neurons and two astrocytes that were ZsGreen⁺; however, there were many FOV with no ZsGreen⁺ neurons and large clusters of ZsGreen⁺ astrocytes. The ZsGreen⁺ immune cell populations were also highly variable throughout the brain. In some FOV, there were no ZsGreen⁺ immune cells, while in other locations, there were more than 50 identifiable immune cells. However, because of clustering of ZsGreen⁺ cells, there were likely more immune cells that we were unable to be identified purely by IHC. To specify and quantify the immune populations that had interacted with the parasite, we performed flow cytometric analysis. As expected, we observed very few ZsGreen⁺ cells in the two control groups, as neurons do not survive the isolation protocol for flow cytometry. Of the few ZsGreen⁺ cells purified in the control groups, we identified CD8⁺ cells as the most highly abundant population (fig. S5, I and J). Notably, the Casp8^−/−Ripk3^−/−Ai6 mice had a fivefold increase in the frequency of ZsGreen⁺ cells compared to WT mice (Fig. 3, G to I). In the Casp8^−/−Ripk3^−/−Ai6 mice, we found that about 60% of the ZsGreen⁺ cells were CD8⁺, about 30% were CD4⁺, and about 10% were CD11b⁺ (Fig. 3, J to M). These data suggested that neurons, astrocytes, and brain-infiltrating immune cells depend on caspase-8 for parasite control during chronic brain infection.

Fig. 3. CD8⁺ T cells, astrocytes, and neurons had increased parasite interactions during chronic infection in Casp8-deficient mice.

Mice were intraperitoneally infected with TF-Pru to identify cell populations with parasite-driven cre activity at 4 wpi. Confocal microscopy was used to image cre activity (green) and T. gondii parasites (magenta) in (A and B) WT^Ai6 and (C to F) Casp8^−/−Ripk3^−/−Ai6 mice. (C) inset represents neurons, (D) inset represents glia, and (E) inset represents immune cells. Images are representative of three independent experiments. Scale bars, 100 μm. ZsGreen⁺ cell populations in the brain of (G) WT^Ai6 and (H) Casp8^−/−Ripk3^−/−Ai6 mice were (I) quantified by spectral flow cytometry. SS-A, side scatter-area. ZsGreen⁺ cell populations in the brains of Casp8^−/−Ripk3^−/−Ai6 mice were analyzed for (J) CD8, (K) CD4, and (L) CD11b expression and (M) quantified. Statistical significance determined by randomized block ANOVA and least-squares means: (I) WT^Ai6 (n = 11), Casp8^+/−Ripk3^−/−Ai6 (n = 16), and Casp8^−/−Ripk3^−/−Ai6 (n = 19) (five experiments). Data are presented as mean ± SEM; *P < 0.05.

Casp8 deletion in neurons, astrocytes, microglia, and macrophages did not affect chronic T. gondii infection control

On the basis of the observation that neurons have increased interactions with T. gondii parasites in the Casp8^−/−Ripk3^−/−Ai6 mice (Fig. 3C), we generated Syn^CreCasp8^fl/flRipk3^−/−Ai6 mice to explore the role of Casp8 in neuronal populations. We confirmed excision of Casp8 in a purified neuronal population from naïve mouse brain using RT-qPCR (Fig. 4A). Furthermore, we observed robust cre activity–induced ZsGreen expression solely and broadly in neurons from naïve adult mouse brains (Fig. 4, B to D). At 6 wpi, we found no differences in brain parasite burden in the Syn^CreCasp8^fl/flRipk3^−/−Ai6 mice compared to the control Casp8^fl/flRipk3^−/−Ai6 mice (Fig. 4E). There was a slight increase in the number of CD8⁺ T cells in the neuronal Casp8 knockouts in comparison to controls, but there was no difference in the number of CD4⁺ T cells and Ly6C^hiiNOS⁺ inflammatory monocytes (Fig. 4, F to H). In addition, using the caspase-3 inhibition paradigm to measure cleaved caspase-8⁺ cells in the brain during chronic infection (Fig. 2K), we did not identify caspase-8 activity in NeuN⁺ cells (fig. S6, A to D). These results suggest that neuronal expression of Casp8 is likely not contributing to the increase in parasite burden observed in the whole-body Casp8^−/−Ripk3^−/− mice.

Fig. 4. Casp8 deletion in neurons or astrocytes did not affect parasite burden or immune responses during chronic T. gondii infection.

(A) Levels of Casp8 were measured by RT-qPCR from neurons purified from naïve Syn^CreCasp8^fl/flRipk3^−/−Ai6 mice (Syn^Cre) and Casp8^fl/flRipk3^−/−Ai6 mice (control). (B to D) Cre-reporter activity (ZsGreen) in neurons (NeuN; magenta) was assessed by IHC in naïve Syn^CreCasp8^fl/flRipk3^−/−Ai6 mice. Scale bars, 50 μm. (E) Syn^CreCasp8^fl/flRipk3^−/−Ai6 mice and control Casp8^fl/flRipk3^−/−Ai6 mice were infected for 6 weeks, and cyst burden was measured. Immune responses in the brain were measured by flow cytometry: (F) CD8⁺ T cells, (G) CD4⁺ T cells, and (H) monocytes producing iNOS (CD45^hiCD11b⁺Ly6G⁻Ly6C^hiNOS). (I) Levels of Casp8 were measured by RT-qPCR from ACSA2⁺ astrocytes purified from naïve Aldh1l1^CreERT2Casp8^fl/flRipk3^−/−Ai6 (Casp8^fl/fl) and Aldh1l1^CreERT2Casp8^fl/wtRipk3^−/−Ai6 (Casp8^fl/wt) control mice. (J to L) Cre-reporter activity (ZsGreen) in astrocytes (GFAP/ALDH1L1/S100β; magenta) was assessed in naïve Aldh1l1^CreERT2Casp8^fl/flRipk3^−/−Ai6 mice. Scale bars, 50 μm. (M) Aldh1l1^CreERT2Casp8^fl/flRipk3^−/−Ai6 and control Aldh1l1^CreERT2Casp8^fl/wtRipk3^−/−Ai6 mice were infected for 6 weeks, and cyst burden was measured. Immune responses in the brain were measured by flow cytometry: (N) CD8⁺ T cells, (O) CD4⁺ T cells, and (P) brain-infiltrating inflammatory monocytes producing iNOS (CD45^hiCD11b⁺Ly6G⁻Ly6C^hiNOS⁺). Statistical significance determined by unpaired t test: (A) Casp8^fl/flRipk3^−/−Ai6 (n = 4) and Syn^CreCasp8^fl/flRipk3^−/−Ai6 (n = 4) (one experiment). Statistical significance was determined by randomized block ANOVA and least-squares means: (E to H) Casp8^fl/flRipk3^−/−Ai6 (n = 14) and Syn^CreCasp8^fl/flRipk3^−/−Ai6 (n = 11) (three experiments); (I) Aldh1l1^CreERT2Casp8^fl/wtRipk3^−/−Ai6 (n = 6) and Aldh1l1^CreERT2Casp8^fl/flRipk3^−/−Ai6 (n = 5) (two experiments); (M) Aldh1l1^CreERT2Casp8^fl/wtRipk3^−/−Ai6 (n = 9) Aldh1l1^CreERT2Casp8^fl/flRipk3^−/−Ai6 (n = 11) (three experiments); (N to P) Aldh1l1^CreERT2Casp8^fl/wtRipk3^−/−Ai6 (n = 6) and Aldh1l1^CreERT2Casp8^fl/flRipk3^−/−Ai6 (n = 6) (two experiments). Data are presented as mean ± SEM; *P < 0.05.

Given the increased interactions between parasites and astrocytes seen in the TF-Pru experiments (Fig. 3, D to F), we also explored the role of caspase-8 in astrocytes using Aldh1l1^CreERT2Casp8^fl/flRipk3^−/−Ai6 mice. We confirmed excision of Casp8 in astrocyte cell surface antigen-2 (ACSA2⁺) cells purified from the brain by RT-qPCR (Fig. 4I) and also demonstrated cre-induced recombination in astrocytes by IHC (Fig. 4, J to L). At 6 wpi, we found no difference in brain parasite burden or immune cell populations between the Aldh1l1^CreERT2Casp8^fl/flRipk3^−/−Ai6 and Aldh1l1^CreERT2Casp8^fl/wtRipk3^−/−Ai6 control mice (Fig. 4, M to P). In addition, we did not identify caspase-8 activity in GFAP⁺ cells using the caspase-3 inhibition paradigm (fig. S6, E to F). While Aldh1l1^CreERT2 is used to target astrocytes with no detectible targeting of neurons, it has been well documented that Aldh1l1 is highly expressed in hepatocytes and proximal tubular cells and mucus glandular cells (70–72). To assess whether Aldh1l1^CreERT2 expression outside of astrocytes affects acute/peripheral infection, we measured parasite burden at 8 dpi. We found no differences in parasite burden in the brain, heart, lung, or liver between the experimental and control mice (fig. S6, G to J). Together, this demonstrates that capsase-8 in Aldh1l1-expressing cells does not play a role in parasite restriction.

Brain-resident microglia and infiltrating myeloid cells play a critical role in restricting T. gondii infection (32, 42). Because both cell populations highly express Casp8, we explored the role of caspase-8 expression in microglia using the Cx3cr1^CreERT2 mice and in peripheral myeloid cells using Lysozyme M (LysM)^Cre mice. After confirming excision of Casp8 from microglia in naïve brains of the Cx3cr1^CreERT2Casp8^fl/flRipk3^−/−Ai6 mice by RT-qPCR (fig. S7A), we examined the role of microglial Casp8 during chronic T. gondii infection. At 6 wpi, there were no differences in brain parasite burden or in the number of T cells, infiltrating inflammatory monocytes producing iNOS, or microglia (CD45^intCD11b⁺) when comparing the Cx3cr1^CreERT2Casp8^fl/flRipk3^−/−Ai6 mice and the control Cx3cr1^CreERT2Casp8^fl/wtRipk3^−/−Ai6 mice (fig. S7, B to E). This suggests that caspase-8 expression in microglia is not essential for parasite restriction.

To assess the role of Casp8 in myeloid cells, we created LysM^CreCasp8^fl/flRipk3^−/−Ai6 mice. We confirmed excision of Casp8 in purified monocytes from the peritoneal cavity by RT-qPCR (fig. S7F). We then assessed parasite burden during the acute infection between the LysM^CreCasp8^fl/flRipk3^−/−Ai6 mice and the LysM^CreCasp8^fl/wtRipk3^−/−Ai6 control mice. We did not find any difference in parasite genomic DNA in the brain, heart, lung, or liver between the experimental and control group during the acute phase of infection (fig. S7, G to J). At 6 wpi, we did not find any difference in brain parasite burden or the number of infiltrating T cells, infiltrating monocytes, or infiltrating inflammatory monocytes producing iNOS (fig. S7, K to N). Overall, these experiments demonstrate that caspase-8 expression in brain-resident neurons, astrocytes, microglia, or infiltrating monocytes is not required for parasite control during T. gondii infection.

Casp8 deletion in CD8⁺ T cells led to decreased survival, increased brain parasite burden, and infected CD8⁺ T cells during chronic T. gondii infection

We next hypothesized that caspase-8 expression in CD8⁺ T cells would contribute to parasite restriction in the brain. To explore the role of caspase-8 in CD8⁺ T cells during T. gondii infection, we generated experimental Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice and control Casp8^fl/fl-Ripk3^−/−Ai6 mice. The Cd8α^Cre transgenic mice ensure CD8⁺, not CD4⁺, T cell specificity by targeting the Cd8α E8I enhancer region to drive cre recombinase. E8I regulates Cd8α expression and is active only in mature CD8 single-positive thymocytes and CD8⁺ T cells, as well as in innate-like CD8αα⁺ intraepithelial lymphocytes of the gut (73). We confirmed Casp8 excision in CD8⁺ T cells isolated from lymph nodes in naïve animals using RT-qPCR (fig. S8A). To further confirm the specificity of Casp8 deletion, we performed flow cytometry to identify cre-driven ZsGreen⁺ populations. At 6 wpi, more than 95% of the ZsGreen⁺ cells were CD3⁺CD8⁺ in the brains of Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice (fig. S8, B to C). In addition, the ZsGreen population was composed of less than 1% CD3⁻CD8⁺CD11c⁺MHCII⁺ DCs, CD3⁺B220⁺ cells, and CD3⁺CD4⁺ T cells (fig. S8C). In the spleens of Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice, at 6 wpi, the ZsGreen cells identified by flow cytometry were 60% CD3⁺CD8⁺, 30% CD3⁺CD8⁻B220⁺, 10% CD3⁺CD8⁻B220⁻CD4⁻, less than 2% CD3⁻CD8⁺CD11c⁺MHCII⁺, and less than 1% CD8⁻CD4⁺ (fig. S8, D and E). In the brains of the Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice, the frequency of B220⁺ cells within the CD3⁺ populations was around 8% and equivalent to the control group (fig. S8F) but markedly lower than in the spleen where ~35% of CD3⁺ cells were also B220⁺ (fig. S8G). We also assessed cre activity (ZsGreen expression) in CD8α⁺ DCs (CD11c^hiMHCII^hi) in the brain (fig. S8H) and found that ~20% of the CD8α⁺ DCs in the brain expressed ZsGreen, while ~5% of the CD8α⁺ DCs in the spleen were ZsGreen-positive (fig. S8I). Overall, the Cd8α^Cre was quite specific for CD8⁺ T cells within the immune cell compartment, as previously reported (74).

We next assessed the impact of CD8⁺ T cell–specific caspase-8 deletion on parasite control in the brain during chronic infection. We first asked whether caspase-8 is active in CD8⁺ T cells during chronic infection. In the brain of chronically infected Cd8α^CreAi6 mice, we identified cleaved caspase-8⁺ in ZsGreen⁺ CD8⁺ T cells (Fig. 5, A to C). This would suggest that caspase-8 could potentially play a role in parasite restriction in CD8⁺ T cells. Exploring parasite control, there was no difference in brain cyst burden at 4 wpi; however, there was a significant increase in the number of cysts per brain in the Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice at 6 and 8 wpi (Fig. 5, D to F). Around the 8 wpi time point, the Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice began to succumb to the infection, while the Casp8^fl/flRipk3^−/−Ai6 mice persisted (Fig. 5G). This demonstrated that caspase-8 in CD8⁺ T cells regulates parasite restriction and is critical for host survival. To assess whether Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice had differences in parasite control during acute infection that could potentially impact the density of parasite that infiltrates into the brain, we measured parasite burden at 8 dpi. We found no differences in parasite burden in the brain, heart, lung, or liver between the experimental and control groups, suggesting that any difference in brain parasite burden is not due to an inability to control parasite in peripheral tissues (fig. S9, A to D). Notably, we did identify sparse ZsGreen-labeled neurons in naïve mouse brains by microscopy (fig. S9, E to G) (71). However, on the basis of the lack of difference in brain parasite burden in the Syn^CreCasp8^fl/flRipk3^−/−Ai6 mouse line, any defects in parasite control in Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice are most likely due to CD8⁺ T cells.

Fig. 5. Casp8 deletion in CD8⁺ T cells increased parasite burden during chronic T. gondii infection.

Cd8α^CreAi6 mice were infected with 10 cysts of the Me49 strain of T. gondii intraperitoneally. At the chronic phase of infection, confocal microscopy was used to visualize (A) cleaved caspase-8 (magenta) with DAPI (blue), (B) cleaved caspase-8 (magenta) with ZsGreen-expressing CD8⁺ T cells (green), and (C) ZsGreen-expressing CD8⁺ T cells (green) alone. Scale bars, 20 μm. Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice (Cd8α^Cre) and Casp8^fl/flRipk3^−/−Ai6 mice (control) were infected with 10 cysts of the Me49 strain of T. gondii intraperitoneally, and cyst burden was analyzed at (D to F) 4, 6, and 8 wpi. (G) Survival curve of infected Casp8^fl/flRipk3^−/−Ai6 control and Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice. Immune cells in the brain were analyzed by flow cytometry at 6 wpi: (H) CD8⁺ T cells, (I) CD8⁺IFN-γ⁺ T cells, and (J) CD8⁺SIINFEKL⁺ T cells, (K) CD4⁺ T cells, (L) CD4⁺IFN-γ⁺ T cells, and (M) inflammatory monocytes (CD45^hiCD11b⁺Ly6G⁻Ly6C^hiNOS⁺). Gene expression was measured from whole-brain homogenate by RT-qPCR: (N) Ifng and (O) Nos2. (P to R) Infection of CD8⁺ T cells in the brain was observed by IHC in Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice. Parasite staining is shown in magenta, and ZsGreen-expressing CD8⁺ T cells in green. Scale bars, 20 μm. Statistical significance determined by randomized block ANOVA and least-squares means: (D) Casp8^fl/flRipk3^−/−Ai6 (n = 16) and Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 (n = 19) (three experiments); (E, H, K, and M) Casp8^fl/flRipk3^−/−Ai6 (n = 12) and Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 (n = 18) (three experiments); (F) Casp8^fl/flRipk3^−/−Ai6 (n = 16) and Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 (n = 11) (two experiments); (I and L) Casp8^fl/flRipk3^−/−Ai6 (n = 6) and Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 (n = 8) (two experiments); (N and O) Casp8^fl/flRipk3^−/−Ai6 (n = 5) and Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 (n = 14) (two experiments). Statistical significance determined by log-rank (Mantel-Cox) test: (G) Casp8^fl/flRipk3^−/−Ai6 (n = 13) and Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 (n = 11) (two experiments). Statistical significance determined by unpaired t test: (J) Casp8^fl/flRipk3^−/−Ai6 (n = 4) and Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 (n = 4) (one experiment). Data are presented as mean ± SEM; *P < 0.05, **P < 0.01, and ***P < 0.001.

We next examined immune responses in the Cd8α^CreCasp8^fl/fl-Ripk3^−/−Ai6 mice. Similar to mice that completely lack Casp8 and Ripk3, we saw a significant increase in the number of CD8⁺ T cells and CD8⁺IFN-γ⁺ T cells in the brains of the Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice at 6 wpi (Fig. 5, H and I). To confirm that the expanded CD8⁺ T cell population included parasite-specific T cells, we infected the Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice with Pru–ovalbumin (OVA) parasites. At 6 wpi, we found an increase in the number SIINFEKL-specific CD8⁺ T cells (Fig. 5J). In the mice infected with Me49, at 6 wpi, there was no significant difference in the number of CD4⁺ T cells but a significant decrease in the number of CD4⁺IFN-γ⁺ T cells (Fig. 5, K and L). Although we observed a decrease in CD4⁺IFN-γ⁺ T cells, this did not affect activation of the inflammatory monocyte population; there was an increase in the number of CD11b⁺Ly6C^hiiNOS⁺ inflammatory monocytes (Fig. 5M). Measuring cytokine expression in whole-brain homogenate by RT-qPCR, we found equivalent expression levels of Ifng and Nos2 in brains of Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice compared to control mice (Fig. 5, N and O). Together, these findings underscore the critical role that caspase-8 expression in CD8⁺ T cells plays in parasite restriction during chronic brain infection.

The results from the Cre-secreting parasite infection (Fig. 3M) and the increase in brain parasite burden in the Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice (Fig. 5, E and F) strongly suggest that CD8⁺ T cells would be more heavily infected in the absence of Casp8 expression (24, 75). To visualize whether CD8⁺ T cells deficient in Casp8 could sustain infection, we imaged brains at 6 wpi to identify parasite-infected CD8⁺ T cells. In examining the cortex, we readily observed parasite-infected CD8⁺ T cells in the Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice (n = 4 mice per two tissue slices per mouse, approximately two to four infected cells per 50-μm-thick brain slice) (Fig. 5, P to R). To our surprise, we also identified infected CD8⁺ T cells that harbored multiple parasites in a single parasitophorous vacuole, suggesting that the parasite survives in a Casp8-deficient CD8⁺ T cell long enough to undergo replication (fig. S10, A to L). While most of the infected CD8⁺ T cells only harbored a single tachyzoite, 17% harbored a vacuole containing two parasites, and 13% harbored a vacuole containing four parasites (fig. S10M). We observed that the infected CD8⁺ T cells were typically in a region devoid of other CD8⁺ T cells and were not found in immune clusters near sites of reactivation (defined as an average of >4 CD8⁺ T cells within a 20-μm-radius circle on a single plane, with individual parasites present) (fig. S10N). As expected, in control Cd8α^CreAi6 mice, we did not identify any infected CD8⁺ T cells at 6 wpi (n = 3 mice per two tissue slices, 50-μm brain slices per mouse) (fig. S10, O to Q). Thus, we find that with a lack of caspase-8 in CD8⁺ T cells, the parasite can infect and survive in CD8⁺ T cells despite the presence of a strong IFN-γ response. This suggests a role for caspase-8–deficient CD8⁺ T cells in parasite dissemination throughout the brain. From these results, we demonstrate a previously unknown feature of CD8⁺ T cell parasite control via caspase-8 as a protective mechanism of T. gondii restriction in the mouse brain.

Fas deficiency increased brain parasite burden and decreased animal survival during chronic T. gondii infection

Caspase-8 can be activated by a number of mechanisms, including death receptors in the TNF receptor superfamily (TNFRSF) (76–78). Previous work has shown that disruption of TNFR signaling either through genetic deletion or TNF neutralization causes the mice to succumb to infection—a process that appears to be driven by an inflammatory monocyte and iNOS deficiency in the brain (54, 79). These mice do not have a lymphoproliferative phenotype such as Casp8^−/−Ripk3^−/− mice, suggesting that TNF does not drive caspase-8–mediated cell death in CD8⁺ T cells. Fas is a well-known activator of caspase-8 that is highly expressed in T cells, and Fas-deficient mice develop lymphoproliferation and splenomegaly similar to the Casp8^−/−Ripk3^−/− mice (71). Thus, we explored the role of Fas during chronic T. gondii infection. At 4 wpi, there were no differences in cyst parasite burden in the brains of C57BL/6 WT and Fas^lpr mice; however, there was a significant increase in parasite burden in the Fas^lpr mice at 6 wpi compared to littermate Fas^lpr+/− controls (Fig. 6, A and B). Fas^lpr mice began to succumb to infection around 11 wpi, while C57BL/6 WT mice survived for several months (Fig. 6C). At 6 wpi in Fas^lpr mice, there was a significant increase in CD8⁺ T cells (Fig. 6, D and E) and no difference in CD4⁺ T cells and CD11b⁺Ly6C^hiiNOS⁺ inflammatory monocytes compared to littermate controls (Fig. 6, F and G). We found no difference in the expression of Ifng or Nos2 in brains of Fas^lpr mice compared to littermate controls (Fig. 6, H and I). From this work, we conclude that Fas, an activator of caspase-8, is critical for limiting the expansion of the CD8⁺ T cell population and restricting T. gondii in the brain.

Fig. 6. Fas deficiency led to increased brain parasite burden and decreased survival.

Fas^lpr mice and control mice were infected with 10 cysts of the Me49 strain of T. gondii intraperitoneally, and cyst burden was analyzed. (A) Parasite burden at 4 wpi comparing C57BL/6 WT control and Fas^lpr mice. (B) Parasite burden at 6 wpi comparing Fas^lpr+/− (littermate control) and Fas^lpr mice. (C) Survival curve of infected C57BL/6 WT control and Fas^lpr mice. Brain immune cell populations were quantified by spectral flow cytometry comparing Fas^lpr+/− (littermate control) and Fas^lpr mice at 6 wpi: (D and E) CD8⁺ T cells, (F) CD4⁺ T cells, and (G) inflammatory monocytes (CD45^hiCD11b⁺Ly6G⁻Ly6C^hiNOS⁺). Gene expression was measured from whole-brain homogenate by RT-qPCR: (H) Ifng and (I) Nos2. Statistical significance determined by randomized block ANOVA and least-squares means: (A) WT (n = 14) and Fas^lpr (n = 13) (three experiments); (B, E, F, and G) Fas^lpr+/− (n = 16) and Fas^lpr (n = 11) (three experiments); (H and I) Fas^lpr+/− (n = 8) and Fas^lpr (n = 7) (two experiments). Statistical significance determined by log-rank (Mantel-Cox) test: (C) WT (n = 8) and Fas^lpr (n = 12) (two experiments). Data are presented as mean ± SEM; *P < 0.05, **P < 0.01, and ***P < 0.001.

DISCUSSION

The results of this study underscore a previously unknown role for caspase-8 in controlling T. gondii infection in the brain and animal survival. Prior studies highlight a role of caspase-8 in promoting immune responses to pathogen infection, particularly through enhanced cytokine production (10, 11, 13, 80). Contrastingly, this study revealed a robust T_H1 cell immune response and associated cytokine production in caspase-8–deficient mice. Because caspase-8 is an initiator of extrinsic apoptosis, our work suggests that regulation of cell death may be a means by which caspase-8 contributes to parasite restriction in the brain (76). In the Casp8^−/−Ripk3^−/− mice, we found a high parasite load and a large population of parasite-interacted cells in comparison to WT and Casp8^+/−Ripk3^−/− mice. Because of this, we hypothesized that multiple cell types would use caspase-8 to restrict parasite replication in the brain. In the Casp8^−/−Ripk3^−/− mice, we found high numbers of parasite-interacted neurons and astrocytes; however, these cell populations did not use Casp8 as a mechanism of parasite control. Neurons are the cell population that harbor T. gondii in WT mice, while astrocytes have been shown to have increased parasite interactions in mice with high parasite loads (24). The increase of neuron and astrocyte-interacted cells in the Casp8^−/−Ripk3^−/− mice was likely driven by the high parasite burden in the Casp8^−/−Ripk3^−/− mice. The Casp8^−/−Ripk3^−/− mice had a high parasite burden in the periphery during the acute infection that contributed to more parasite seeding into the brain at 8 dpi. This suggests that a peripheral cell type uses caspase-8 as a mechanism of parasite control during the acute infection. In this study, we examined Aldh1l1-expressing, LysM-targeted, and Cd8α-expressing cells during the acute phase of infection and found no difference in parasite burden.

During the chronic phase of infection, we found that caspase-8 was particularly important in CD8⁺ T cells. During T. gondii infection, CD8⁺ T cells play an integral role in parasite restriction via cytokine production and cytotoxicity. It was unexpected that they would also use caspase-8 to restrict the parasite; however, in Cd8α^CreCasp8^fl/flRipk3^−/−Ai6 mice, we observed infected CD8⁺ T cells and evidence of parasite replication within these infected cells. In addition, these mice were more susceptible to infection and had an increased brain parasite burden likely due to parasite survival within CD8⁺ T cells. Further demonstrating the importance of caspase-8 in CD8⁺ T cells, we identified CD8⁺ T cells with active caspase-8 during the chronic phase of infection. Using the TF-Pru infection in the Casp8^−/−Ripk3^−/− mice, we observed many CD8⁺ T cells that had encountered T. gondii but did not remain infected. In C57BL/6 WT mice, however, we did see a parasite-interacted population to a much smaller degree than in the Casp8^−/−Ripk3^−/− mice. Overall, our work suggests that in C57BL/6 WT mice, if CD8⁺ T cells become infected with T. gondii, then they can undergo caspase-8-mediated apoptosis, preventing parasite spread.

Prior studies have shown that CD8⁺ T cells are vulnerable to infection with T. gondii. In lymph nodes of T. gondii–infected mice, CD8⁺ T cells can become infected and have been proposed as a possible means of parasite spread to other regions of the body (81). We hypothesize that, in the brain during parasite cyst reactivation events, CD8⁺ T cells can become infected and potentially spread parasite within the brain (75, 82). Our current study has revealed the importance of caspase-8 in limiting parasite spread within T cells in the brain, as the lack of caspase-8 likely allows CD8⁺ T cells to live longer and allow for parasite survival, replication, and spread. Unexpectedly, we did not find infected CD8⁺ T cells at sites of reactivation. The highly mobile nature of CD8⁺ T cells could allow for infected CD8⁺ T cells to migrate away from a site of reactivation. Supporting our hypothesis, Shallberg et al. (83) visualized an actively infected CD8⁺ T cell in a C57BL/6 WT mouse carrying the parasite through the brain parenchyma during the chronic stage of infection. Exploring mechanisms of T. gondii dissemination in the brain during early brain infection, Schneider et al. (75) identified infected CD8⁺ T cells, but not CD4⁺ T cells, infected infiltrating myeloid cells, and microglia. Together, these studies and our own demonstrate that in C57BL/6 mice, CD8⁺ T cells are exposed to T. gondii infection in the brain and could be a means for parasite dissemination. Building on this, our work emphasizes the importance of caspase-8–mediated cell death of infected CD8⁺ T cells as a crucial mechanism to limit T. gondii spread.

CD8⁺ T cells have long been known to be crucial for restricting T. gondii during the chronic phase of infection (30, 84). Here, we illuminate a previously unknown function of CD8⁺ T cells in pathogen restriction. The established mechanism CD8⁺ T cells use for parasite restriction are predominantly through production of IFN-γ and killing of infected cells through perforin-mediated cell lysis (28–31, 48, 49). Typically, IFN-induced mechanisms of pathogen control are adequate for limiting intracellular infection (32, 35). Our work demonstrates that CD8⁺ T cells can remain infected with T. gondii in the presence of a robust IFN-γ response. Here, we have begun to explore previously overlooked questions of CD8⁺ T cell biology: What happens to a CD8⁺ T cell when it becomes infected? Can CD8⁺ T cells intrinsically eliminate pathogens through an IFN-γ–stimulated response? It is tempting to speculate that caspase-8 could be required in T cells to limit other intracellular pathogens. However, aside from T. gondii, few pathogens have been reported to directly infect CD8⁺ T cells. There have been reports of CD8⁺ T cell infection with measles virus or lymphocytic choriomeningitis virus. CD8⁺ T cell infection with Epstein-Barr virus has also been reported in immunocompromised individuals where virus spreads beyond the typical B cell reservoir (85–88). Theileria parva, a parasite that infects cattle, has also been shown to infect CD8⁺ T cells and induce acute lymphoproliferative disease (89). In addition, human T cell lymphotropic virus type 1 (HTLV-1) can infect T cells and induces adult T cell leukemia/lymphoma in ~5% of patients (90). T. parva– and HTLV-1–infected T cells are resistant to Fas-mediated apoptosis (91, 92). Pathogen-induced lymphoproliferation and resistance to Fas signaling in these infections strongly suggest that intracellular infection of CD8⁺ T cells is associated with pathogen-mediated interference with caspase-8 signaling pathways. HTLV-1 and T. parva infection induces cellular FLICE-inhibitory protein (c-FLIP) expression to prevent caspase-8 activation (91, 93). It has been well documented that numerous viruses and bacteria have evolved mechanisms to inhibit caspase-8 activity (94, 95). In addition, T. gondii has been reported to inhibit caspase-8 activation in vitro (96). However, our work strongly suggests that caspase-8 remains functional in vivo and restricts parasite replication in WT mice. In addition, pathogen evasion of caspase-8 typically induces RIPK3-mediated necroptosis (97, 98). If T. gondii inhibited caspase-8, then we would potentially expect to see a role for RIPK3 in parasite restriction, which we did not find. Overall, this work raises the question of why, in the case of HTLV-1 and T. parva infection, CD8⁺ T cells fail to mount other successful pathogen restriction mechanisms, i.e., RIPK3-mediated necroptosis or interferon-dependent programs.

Caspase-8 plays a multifaceted role in host defense, influencing both inflammatory signaling and cell death pathways. The catalytic activity of caspase-8 promotes apoptosis while suppressing RIPK3-mediated necroptosis, balancing cell fate decisions during infection (11, 80). In addition, caspase-8 functions as a scaffold within oligomeric complexes, driving inflammatory responses (65). In bacterial infections and Plasmodium berghei ANKA infection, caspase-8 enhances cytokine production (13–15, 18, 99). The effects of caspase-8 deficiency vary by pathogen: During Yersinia infection, a lack of caspase-8 reduces cytokine production and increases bacterial load, whereas in Plasmodium infection, a lack of caspase-8 provides protection from immune-mediated pathology (13, 15). In the context of murine cytomegalovirus, caspase-8–deficient mice exhibit heightened T cell cytokine production, but infection is unaffected in terms of viral titers and survival (19, 100). Caspase-8 deficiency during T. gondii infection is fatal—primarily due to an overwhelming brain parasite burden (14). These findings underscore the complex nature of caspase-8 in pathogen defense: It can initiate both proinflammatory and cell death pathways, with outcomes that depend on the specific pathogen and cell type involved. This raises critical questions about how individual cell types leverage caspase-8 for host defense and whether its functions extend to other inflammatory contexts.

Another critical question this study raises is the following: What is the activator of caspase-8 in CD8⁺ T cells? Death receptors, including TNFR, Fas, and TNF-related apoptosis-inducing ligand (TRAIL; DR4/5), among others, have been well described as activators of caspase-8. However, numerous other pattern recognition receptors including Toll-like receptor 4 (TLR4), Z-DNA binding protein 1 (ZBP1), and NLR Family Pyrin Domain Containing 3 (NLRP3) have also been shown to be able to activate caspase-8 (6). In the MERFISH data, we found that Tnfrsf10b, Tnfrsf21, Tnfrsf25, and Tlr4 are very lowly expressed, while Tnfrsf1a, Fas, and Zbp1 had a moderate level of expression in CD8⁺ T cells. Extensive work has been done exploring a role for TNF in T. gondii infection. Upon TNF signaling disruption, the main immune impairment consists of a monocyte and iNOS deficiency in the brain (54, 79). In addition, TNFR-deficient mice do not develop a lymphoproliferative phenotype such as Casp8^−/−Ripk3^−/− mice. This would suggest that TNF likely does not drive caspase-8 activity in CD8⁺ T cells during T. gondii infection. Fas is a well-known activator of caspase-8 that is highly expressed in T cells, and Fas-deficient mice develop lymphoproliferation and splenomegaly similar to the Casp8^−/−Ripk3^−/− mice (71). Early work demonstrated that Fas (lpr) and Fasl (gld)–deficient mice had worse outcomes despite robust IFN-γ responses during ocular T. gondii infection (52, 101). Here, we demonstrated that the Fas (lpr) mice had increased brain parasite burden and succumb earlier than WT controls despite an expanded CD8⁺ T cell population. One potential mechanism of caspase-8 activation in CD8⁺ T cells during chronic T. gondii infection could be through CD8⁺ T cell FasL, suggesting that activated T cells eliminate infected cells through caspase-8 in addition to canonical cytolysis.

In this study, we found an unexpected role for caspase-8 in CD8⁺ T cells as a parasite restriction mechanism in the brain. We hypothesize that caspase-8 is driving apoptosis to restrict parasite survival and spread in CD8⁺ T cells. This study was limited by the use of genetic models that completely deleted Casp8 versus a manipulation of the enzymatic activity of caspase-8. Inhibition of caspase-8 enzymatic activity could be used to confirm caspase-8–driven apoptosis as the mechanism of parasite restriction in CD8⁺ T cells. Visualization of cleaved caspase-8 in CD8⁺ T cells in the brain during chronic infection proved to be challenging due to the rapid nature of apoptosis. Moreover, caspase-8 activity in an actively infected CD8⁺ T cell was not demonstrated because of limitations in the number of infected cells and the ability to visualize both parasite and cleaved caspase-8 via IHC. Utilization of genetic models that inactivate caspase-8 enzymatic activity specifically in CD8⁺ T cells would allow for further confirmation of our hypothesis of caspase-8–mediated apoptosis being critical for restriction of T. gondii in CD8⁺ T cells in vivo. Overall, our work contributes to the growing body of literature related to cell death as a pathogen restriction mechanism. Here, we highlight the protective role of caspase-8 specifically in CD8⁺ T cells, where it serves as a key mechanism for restricting T. gondii infection.

MATERIALS AND METHODS

Animals

WT (C57BL/6J), ROSA26^Ai6/Ai6 (#007906) (67), Syn1^Cre (#003966), Aldh1l1^CreERT2 (#031008), Cx3cr1^CreERT2 (#020940), LysM^Cre (#004781), Cd8α^Cre (#008766), Fas^lpr (#000482), and CBA/J (#000656) strains were obtained from the Jackson Laboratory and maintained within our animal facility. Swiss Webster (#024) strains were purchased from Taconic. Casp8^−/−Ripk3^−/−, Casp8^+/−Ripk3^−/−, and Casp8^fl/flRipk3^−/− mice were provided by J. R. Lukens (102, 103). Each cre line was cross-bred with Casp8^fl/flRipk3^{−/−Ai6+/+} mice to generate [X]^cre/+ × Casp8^fl/flRipk3^{−/−Ai6+/+}. All procedures involving animal care and use were approved by and conducted in accordance with the University of Virginia’s Institutional Animal Care and Use Committee under protocol number 3968.

Tamoxifen treatment

To induce cre expression and drive excision of Casp8 within the Aldh1l1^CreERT2 and Cx3cr1^CreERT2 mouse lines, tamoxifen (Sigma-Aldrich, catalog no. T5645) was dissolved in corn oil (Sigma-Aldrich, catalog no. C8267) and filtered through a 0.45-μm filter (Millipore, catalog no. SLGSM33SS). At 4 weeks of age, age- and sex-matched mice were intraperitoneally injected with tamoxifen (200 mg/kg) every other day for 10 days (total of five injections). Mice were then allowed to recover for 4 weeks before parasite infection.

Parasite strains

Mice were infected with type II strains of T. gondii. The Me49 strain was maintained in chronically infected (2 to 6 months) Swiss Webster mice and passaged through CBA/J mice. The transgenic Prugniaud strains of T. gondii–expressing cre fused to TF-Pru (24) and OVA (amino acids 140 to 386) with TdTomato (Pru-OVA) (104) were generously provided by A. Koshy (University of Arizona) and maintained by serial passage through human foreskin fibroblast (HFF) monolayers in parasite culture medium [Dulbecco’s modified Eagle’s medium (Gibco, catalog no. 10313-021), 20% Medium 199 (Gibco, catalog no. 11150-059), 10% fetal bovine serum (FBS; Gibco, catalog no. 10082-147), 1% penicillin/streptomycin (Gibco, catalog no. 15140-122), and gentamicin (10 μg/ml; Sigma-Aldrich, catalog no. G1272)].

Infection

For experimental infections with the Me49 strain, tissue cysts were prepared from homogenized brains of chronically infected (4 to 8 weeks) CBA/J mice. Mice were inoculated intraperitoneally with 10 tissue cysts of Me49 in 200 μl of 1× PBS (Gibco, catalog no. 14190144). For infections with the TF-Pru or the Pru-OVA strain, tachyzoites were purified from HFF cultures via needle passage of scraped cells and filtered through a 5.0-μm filter (Millipore, catalog no. SLSV025LS). Mice were then inoculated intraperitoneally with 1000 tachyzoites in 200 μl of 1× PBS. Mice used for endpoint studies were euthanized by CO₂ asphyxiation if they showed weight loss greater than 20% of their preinfection body weight.

Tissue processing

At specified time points, postinfection mice were euthanized by CO₂ asphyxiation. Peritoneal lavage fluid was collected by injecting and removing 5 ml of cold 1× PBS through the intact peritoneal membrane with a 27-gauge needle (BD, catalog no. 305109). PECs were washed and resuspended in complete RPMI 1640 medium (cRPMI) [10% FBS, 1% penicillin/streptomycin, 1% sodium pyruvate (Gibco, catalog no. 11360-070), 1% nonessential amino acids (Gibco, catalog no. 11140-050), and 0.1% 2-mercaptoethanol (Gibco, catalog no. 21985-023)]. Blood was collected by cardiac puncture using a 27-gauge needle. Blood was stored at 4°C overnight and then centrifuged at ×20,000g for 20 min at 4°C to separate serum (Eppendorf, 5425 R). For experiments collecting brain and spleen, transcardiac perfusion with 20 ml of cold 1× PBS was performed. Brains and spleens were harvested and put into cRPMI. Brains were passed through an 18-gauge needle (BD, catalog no. 305195), and 500 μl was stored at −20°C for RNA and DNA extraction and cyst counting. Remaining brain homogenate was then enzymatically digested with collagenase/dispase (0.227 mg/ml; Sigma/Aldrich, catalog no. 10269638001) and deoxyribonuclease (50 U/ml; Sigma-Aldrich, catalog no. 11284932001) at 37°C for 45 min with intermittent inverting. Postdigestion, brain homogenate was passed through a 70-μm filter (Corning, catalog no. 352350) and washed with cRPMI. Filtered brain homogenate was then resuspended with 20 ml of 40% Percoll (Cytiva, catalog no. 17-0891-02) and centrifuged at 650g for 25 min to remove myelin via aspiration of upper layer. Brain samples were then washed and resuspended with cRPMI. Spleens were manually homogenized and passed through a 40-μm filter (Corning, catalog no. 352340). Samples were then resuspended in 2 ml of red blood cell lysis buffer (0.16 M NH₄Cl) for 2 min. Cells were then washed with cRPMI and resuspended. For intracellular cytokine staining, cells were resuspended in 200 μl of cRPMI with Brefeldin A (20 μg/ml) (Selleckchem, catalog no. S7046) or Brefeldin A, Phorbol 12-myristate 13-acetate (200 ng/ml) (Sigma Aldrich, catalog no. P1585), and ionomycin (1 μg/ml) (Sigma Aldrich, catalog no. I0634) for 5 hours at 37°C. Cells suspensions were enumerated by diluting 1:10 in 0.4% trypan blue solution (Sigma-Aldrich, catalog no. T8154) and counted on a hemocytometer (Hausser Scientific, catalog no. 3110) using a DM2000 LED bright-field microscope (Leica).

Parasite burden quantification

To measure parasite burden by qPCR, genomic DNA was isolated from mouse brain, heart, lung, and liver. Tissues were first homogenized in PBS using the Omni TH tissue homogenizer (Omni International). Parasite genomic DNA was isolated using the Isolate II Genomic DNA Kit (Bioline, catalog no. BIO-52067). Amplification of T. gondii 529–base pair repeat region using the SensiFAST Probe No-ROX Kit (Bioline, catalog no. BIO-86005) and CFX384 Real-Time System (Bio-Rad Laboratories) was performed as previously described (105). A total of 500 ng of DNA was loaded into each reaction. T. gondii isolated from HFFs was used to make a serial standard curve from 3 to 300,000 genome copies and determine the number of T. gondii genomes per mircograms of tissue DNA. To measure brain parasite burden by cyst counts, whole brains were passed through an 18-gauge and then 22-gauge (BD, catalog no. 305155) needle to mechanically homogenize the tissue. A total of 30 μl of brain homogenate was mounted on a microscope slide. T. gondii cysts were counted visually using a DM2000 LED bright-field microscope.

Flow cytometry

Single-cell suspensions were resuspended in Fc Block {fluorescence-activated cell sorting (FACS) buffer [PBS, 0.2% bovine serum albumin (BSA) (Bioworld, catalog no. 9048-46-8), and 2 mM EDTA (Invitrogen, catalog no. AM9260G]} with 2.4G2 antibody (0.1 μg/ml; BioXCell, catalog no. CUS-HB-197) and 0.1% rat gamma globulin (Jackson ImmunoResearch, catalog no. 012-000-002) into a 96-well plate for 10 min at room temperature. To measure SIINFEKL-specific CD8⁺ T cells, cells were preincubated with phycoerythrin (PE)–conjugated H-2Kb/OVA (SIINFEKL) tetramer (National Institutes of Health Tetramer Core Facility) for 15 min at room temperature before cell surface staining. Cells were stained for surface markers and fixable live/dead viability dye for 30 min at 4°C. Cells were then washed twice with 50 μl of FACS buffer. For intracellular staining, cells were fixed with fixation/permeabilization solution (eBioscience, catalog nos. 00-5123-43 and 00-5223-56) for 20 min at room temperature or at 4°C overnight. Cells were then washed twice with 50 μl of 1× permeabilization buffer (eBioscience, catalog no. 00-8333-56) and stained for intracellular markers in 1× permeabilization buffer for 30 min at 4°C. Last, cells were resuspended in 200 μl of FACS buffer and acquired using a three-laser Cytek Aurora Flow Cytometry System or Gallios flow cytometer (Beckman Coulter). Data were analyzed using FlowJo software V10.9.0. The Fixable Viability Dye–eFluor 780 (Thermo Fisher Scientific, catalog no. 50-112-9035) was used at 1:800. The following antibodies were used at a dilution of 1:200: CD45–Alexa Fluor 700 (BioLegend, catalog no. 103128), CD11b–PerCP Cy5.5 (Thermo Fisher Scientific, catalog no. 45-0112-80), Ly6G-BV711 (BioLegend, catalog no. 127643), Ly6C-PE (Thermo Fisher Scientific, catalog no. 12-5932-82), iNOS–allophycocyanin (APC; Thermo Fisher Scientific, catalog no. 17-5920-82), MHC II Super Bright 780 (Thermo Fisher Scientific, catalog no. 78-5321-82), CD11c–eFluor 506 (Thermo Fisher Scientific, catalog no. 69-0114-80), CD3e–PE Cy7 (Thermo Fisher Scientific, catalog no. 25-0031-81), CD4-BV650 (Thermo Fisher Scientific, catalog no. 563232), CD8-BV421 (Thermo Fisher Scientific, catalog no. 563898), Foxp3-eFlour 450 (Thermo Fisher Scientific, catalog no. 48-5773-82), B220-PECy5 (Thermo Fisher Scientific, catalog no. 553091), IFN-γ–PerCPCy5.5 (Thermo Fisher Scientific, catalog no. 45-7311-82), CD178 (FasL)-APC (Thermo Fisher Scientific, catalog no. 17-5911-82), β T cell receptor–APC (Thermo Fisher Scientific, catalog no. 17-5961-81), CD3e-fluorescein isothiocyanate (FITC; Thermo Fisher Scientific, catalog no. 11-0031-85), NK1.1-FITC (Thermo Fisher Scientific, catalog no. 11-0031-85), CD19-FITC (Thermo Fisher Scientific, catalog no. 11-0193-82), B220-FITC (Thermo Fisher Scientific, catalog no. 53-0452-82), IL-12p40-APC (BioLegend, catalog no. 505205), CD64–PE Cy7 (BioLegend, catalog no. 139314), Granzyme B–APC–eFluor 780 (Thermo Fisher Scientific, catalog no. 47-8898-82), and TNFα-PE (Thermo Fisher Scientific, catalog no. 12-7321-81).

Multiplexed error-robust fluorescence in situ hybridization

C57BL/6 mice were infected intraperitoneally with Me49 or saline and euthanized at 4 wpi. Dissected brains were fresh frozen in optimal cutting temperature (OCT) freezing medium in isopentane chilled with liquid nitrogen. Brains were then sectioned at 10-μm thickness (Leica, CM1950) directly onto MERSCOPE slides (Vizgen). Slides were prepared according to MERSCOPE User Guide for Fresh and Fixed Frozen Tissue Sample Preparation Rev. D with amendments made on the basis of MERSCOPE Quick Guide Modified Fixation for Fresh & Fixed Frozen Sample Preparation Rev. A and MERSCOPE Technical Note Improving Tissue Adherence Rev. A. Cell segmentation was performed using the Vizgen Postprocessing Tool (VPT; v1.1.2), configured to use the Cellpose cyto2 model with cylindrical segmentation based on the 4′,6-diamidino-2-phenylindole⁺ (DAPI⁺) nuclear signal from the middle (third) z plane. The segmentation output files generated by VPT were imported into a Python 3 environment using Squidpy (v1.2.2) to construct an AnnData (v0.10.4) cell-by-gene expression matrix. Quality control thresholds were applied to filter out low-quality or spurious cell profiles. Nuclei were retained if they contained more than 20 total transcripts and if their total transcript counts, number of detected genes, and cell volume fell within the following thresholds: total transcript counts and detected genes greater than 20 and below the 99th percentile and nuclear volume within the first to 99th percentiles. Samples were integrated into a shared latent space using scVI-tools (v1.0.4). High-quality cells were then annotated through Leiden clustering and differential expression of canonical cell type marker genes analysis using the Wilcoxon rank sum test in Scanpy (v1.9.6). Figures were produced using the Vizgen MERSCOPE Visualizer software. Information regarding gene panel and metadata of samples is in sheets S1 and S2.

Caspase-3 inhibition

To measure active caspase-8, mice were injected intraperitoneally with 3 mg/kg of Ac-DEVD-CHO (Enzo Life Sciences, catalog no. ALX260030M005). We began treatment at 4 wpi and gave five injections over the course of 2 weeks. The day after the final treatment mice were euthanized by CO₂, asphyxiation and brains were prepared for IHC.

Histology

For H&E staining, brains were bisected along the sagittal midline and fixed in 4% formalin before staining and slide mounting. For IHC, brains were bisected along the sagittal midline and fixed in cold 4% paraformaldehyde (EMS, catalog no. 15710-S) for 24 hours at 4°C. Brains were then cryoprotected in 30% sucrose for 24 hours at 4°C, embedded in OCT (Tissue-Tek, catalog no. 25608-930), and frozen on dry ice. Fifty-micrometer sections were then prepared using a CM1950 cryostat (Leica) and stored in 1× PBS as free-floating sections. Antigen retrieval was performed to stain for cleaved caspase-8. Free-floating sections were incubated in 10 mM sodium citrate at 95°C for 30 min. Sections were then cooled for 30 min at room temperature. To immunostain free-floating brain sections, the fixed tissues were incubated in a blocking solution [2% normal donkey serum (Jackson ImmunoResearch, catalog no. 017-000-121), 1% BSA, 0.05% Tween 20 (Thermo Fisher Scientific, catalog no. BP337), and 0.5% Triton X-100 (Sigma-Aldrich, catalog no. 028SK001) in 1× PBS] at room temperature for 1 hour. Then, tissue was stained for 1 hour at room temperature or overnight at 4°C with primary antibodies in blocking solution. Samples were washed three times in 0.05% Tween 20 solution and stained with secondary antibodies for 1 hour at room temperature in blocking solution. Last, tissues were washed three times, mounted onto glass slides using AquaMount (Polysciences, catalog no. 18606), and coverslipped (Globe Scientific, catalog no. 1419). In some experiments, the tissue was counterstained with DAPI (Thermo Fisher Scientific, catalog no. 62248) and washed just before mounting onto slides. Primary antibodies included the following: anti-Me49 (1:10,000 dilution) (gift from F. Araujo), NeuN (1:500 dilution) (EMD Millipore, catalog no. ABN91), aldehyde dehydrogenase 1 family member L1 (ALDH1L1; 1:200 dilution) (Abcam, catalog no. ab87117), GFAP (1:500 dilution) (DAKO, catalog no. Z0334), S100β (1:200 dilution) (Abcam, catalog no. ab52642), and cleaved caspase-8 (Asp³⁸⁷) (D5B2) (Cell Signaling Technology, catalog no. 8592T) (1:400). Secondary antibodies were used at 1:500 dilution. To stain Me49, ALDH1L1, GFAP, S100β, and cleaved caspase-8: donkey anti-rabbit Alexa Fluor 594 (Invitrogen, catalog no. R37119); to stain NeuN: donkey anti-chicken Alexa Fluor 647 (Invitrogen, catalog no. A78952). Images were acquired using a Leica STELLARIS 5 confocal microscope and analyzed using Fiji software (106).

Enzyme-linked immunosorbent assay

Serum cytokine levels were detected according to the manufacturer’s instructions using Quantikine enzyme-linked immunosorbent assay (ELISA) (Mouse IL-12 p40 allele specific, R&D Systems, catalog no. M1240) and DuoSet ELISA (Mouse IFN-gamma, R&D Systems, catalog no. DY485).

Cell isolation

Purified cell populations isolated to measure excision of Casp8 were collected according to the manufacturer’s instructions: neurons (Miltenyi Biotech, catalog no. 130-115-390), ACSA2⁺ astrocytes (Miltenyi Biotech, catalog no. 130-097-678), CD11b⁺ microglia (Miltenyi Biotech, catalog no. 130-093-634), monocytes (Miltenyi Biotech, catalog no. 130-100-629), and CD8α⁺ T cells (Miltenyi Biotech, catalog no. 130-104-075).

Reverse transcription quantitative polymerase chain reaction

Brain homogenate was homogenized in TRIzol (Thermo Fisher Scientific, catalog no. 15-596-026), and RNA was extracted according to the manufacturer’s protocol. cDNA was then generated using a High-Capacity Reverse Transcription Kit (Applied Biosystems, catalog no. 4374967). qPCR was performed using 2× Taq-based Master Mix (Bioline, catalog no. 21105) and Taq Man gene expression assays (Thermo Fisher Scientific, catalog no. 4331182). Samples were run on a CFX384 Real-Time System thermocycler (Bio-Rad Laboratories). Genes were normalized to murine Hprt. The 2 ^(–ΔΔCt) method was used to report relative expression (107). The following Thermo Fisher Scientific mouse gene probes were used: Hprt (Mm00446968_m1), Ifng (Mm01168134_m1), Tnf (Mm00443258_m1), Nos2 (Mm00440502_m1), and Casp8 (Mm01255716_m1).

Statistical analysis

All data were graphed in GraphPad Prism 9. To account for the biological variability between infections, data from experimental replicates were analyzed in R using a randomized block analysis of variance (ANOVA) and least-squares means paired with a post hoc analysis to adjust P values for multiple comparisons using the Tukey method. This analysis models experimental groups as a fixed effect and experimental day as a random effect (108). Survival curves were analyzed by log-rank (Mantel-Cox) test. Experiments with only one experimental replicate were analyzed using GraphPad Prism 9 and one-way ANOVA or unpaired t test. Outliers were identified and removed using ROUT method with a Q value of 1. P values are indicated, with *P < 0.05, **P < 0.01, and ***P < 0.001.

Supplementary Materials

The PDF file includes:

Figs. S1 to S10

Legends for sheets S1 and S2

Other Supplementary Material for this manuscript includes the following:

Sheets S1 and S2