The CD8+ T cell response to the intracellular parasite Toxoplasma gondii varies dramatically between mouse strains, resulting in differences in control of the parasite. This difference is linked to MHC haplotype, and protection in BALB/c mice has been attributed to an unusually strong MHC-1 Ld-restricted CD8+ T cell response directed against a peptide derived from the parasite protein GRA6. Our lab has previously shown that the GRA6-Ld specific response is maintained by a proliferative intermediate population that gives rise to large numbers of armed effector T cells throughout chronic infection and shows no signs of functional exhaustion. To further investigate this robust CD8+ T cell response, we characterized the TCR repertoire of GRA6-Ld specific CD8+ T cells after T. gondii infection. We showed that the GRA6-Ld specific T cell response displays a TCR repertoire dominated by the Vβ2 variable chain, limited clonal diversity, a strongly selected CDR3β motif, and a public TCR. Additionally, we showed that another Ld-restricted CD8+ T cell response specific for the F2 peptide from Vaccinia virus also has a preference for Vβ2 and a similarly focused TCR repertoire.

The MHC-1 Ld molecule has limited peptide binding compared to conventional MHC molecules such as Kb or Db, which correlates with polymorphisms associated with “elite control” of HIV in humans. To investigate the link between this unusual MHC-1 molecule and the generation of “elite controller” CD8+ T cell responses, we compared the GRA6-Ld specific T cell response to the OVA-Kb specific response restricted to the conventional MHC-1 Kb molecule. GRA6-Ld specific T cells are significantly more protective and resistant to exhaustion in chronic T. gondii infection. Next, we used CRISPR/Cas9 to generate mice with a point mutation (W97R) in the peptide-binding groove of Ld that results in broader peptide binding and investigated the effect of this Ld W97R mutation on another robust Ld-restricted response against the IE1 peptide during Murine Cytomegalovirus (MCMV) infection. We observed an increase in exhaustion marker expression on the IE1-Ld specific CD8+ T cells during chronic MCMV infection. In both systems, T cells restricted to MHC-1 Ld exhibited resistance to exhaustion compared to T cells restricted to an MHC-1 with broader peptide binding, supporting a model in which limited peptide binding my MHC favors protective responses during chronic infection.

Production of armed effector CD8+ T cells during persistent infection requires a stable pool of stem-like cells that can give rise to short-lived effector T cells via a proliferative intermediate population. In chronic infection models marked by T cell exhaustion, this process can be transiently induced by check-point blockade, but it occurs spontaneously and continuously in the spleens of mice chronically infected with the protozoan intracellular parasite, Toxoplasma gondii, providing a unique opportunity to examine the steps in the differentiation process. Using this model, we observed distinct locations and gene expression patterns for stem-like memory cells, proliferative intermediate cells, and terminally differentiated effector cells, implying a link between differentiation and discrete anatomical niches in the spleen. Loss of the chemokine receptor CXCR3 on parasite-specific T cells did not impair their white pulp to red pulp migration but reduced their interactions with CXCR3 ligand-producing type 2 conventional DCs in the bridging channels and impaired their transition from memory to intermediate states, leading to a build up of memory T cells localized to the red pulp. These results demonstrate a critical role for CXCR3 during chronic infection by increasing T cell exposure to differentiation-inducing signals during red pulp migration, providing a dynamic and flexible mechanism for modulating effector differentiation in response to environmental signals. While mice chronically infected with T. gondii are able to effectively establish long-term control of the pathogen, they do not establish sterilizing immunity. Although the parasite is primarily restricted to the brain during the chronic phase of infection, CD8+ T cells in the periphery experience continuous antigenic stimulation, resulting in ongoing proliferation and effector differentiation. We investigated the role of IFNγ in regulating CD8+ T cell fate during chronic T. gondii infection. We observed that neutralization of IFNγ during chronic infection led to dramatic changes in CD8+ T cell fate, promoting robust proliferation at the expense of terminal effector differentiation. These changes in CD8+ T cell fate were dependent on CXCR3 and linked to increases in antigen presentation in the spleen, which occurred in the absence of widespread parasite reactivation. Finally, IFNγ neutralization resulted in the emergence of clusters of CD11c+ cells in the red pulp of the spleen that were robust producers of CXCR3 ligands and supported a proliferation rich-environment. These results highlight an ongoing role for IFNγ in mediating the clearance of T. gondii antigens during chronic infection and contribute to our understanding of the factors regulating CD8+ T cell fate.

T cells are one of the main effectors of the adaptive immune response and as such, their development must be tightly regulated. Despite being well studied, the processes that produce mature T cells are not fully understood. The ultimate goal in T cell development biology is to understand the in vivo cues well enough to be able to recapitulate development in vitro. Such knowledge would unleash the therapeutic potential of stem cell biology. Although there are many gaps in our current knowledge, I have chosen to focus on two extreme ends of the developmental spectrum, from the earliest determinants in hematopoietic fate in human embryonic stem cells to the final processes of T cell development, positive and negative selection, which are ultimately responsible for the generation of a useful and self-tolerant T cell repertoire.

Although T cells can be derived in vitro from cord blood and fetal liver, a robust development from human embryonic stem cells into the T lineage has not yet been achieved in vitro. It's known that there are different developmental potentials among human embryonic stem cell lines. Therefore, we performed a comparison of the T lineage developmental potential of a panel of six independently derived human embryonic stem cell lines for their ability to differentiate into the known T lineage progenitors, CD34+ hemangioblasts or CD34+CD45+ hematopoietic precursors by embryoid body formation or OP9 co-cultures. Not only did we find indeed, there are differences in hematopoietic potentials among lines, but the passage conditions through which the stem cell lines were maintained also dramatically affected hematopoietic potential. None of the stem cell-derived progenitors could develop into T cells, unlike those found in cord blood. Differences between individual lines should allow for the future study of genetic or epigenetic factors that determine lineage commitment as a cell transitions from a stem cell to a T cell.

Second, I focused on the final stages of T cell development, specifically, the processes of positive and negative selection. Most of our current understanding of signaling behavior comes from in vitro studies that fail to fully recapitulate thymic selection events. Therefore, we expanded upon a relatively new model of selection, the thymic slice system, which supports positive selection and can be used to introduce a synchronized population of thymocytes. By adding a cognate peptide antigen or overlaying slices with peptide-loaded DCs, we were able to induce negative selection. Using the thymic slice system, we visualized the migration and signaling behavior of thymocytes undergoing positive versus negative selection. Additionally, we characterized the effect of the affinity of the peptide-TCR interaction on migration, signaling, and developmental outcome. We found that thymocytes undergoing positive selection formed transient, serial interactions that lasted only a few minutes while thymocytes undergoing negative selection formed long lasting, monogamous contacts. The APC type plays an important role in selection outcome as dendritic cells and not epithelial cell cells presenting OVA most efficiently promoted negative selection and sustained calcium levels. Co-stimulatory and adhesion molecules did not appear to be required for selection outcome but contributed to sustained calcium signaling during negative selection and may therefore play a role in determining the threshold of affinity. We also found that thymocytes undergoing positive selection accumulate TCR signals over time, increase their speed, and appear to change their responsiveness to a TCR-induced stop signal over time.

Central tolerance, or the establishment of self-tolerance during T cell development in the thymus, is critical for the prevention of autoimmunity. Developing T cells (thymocytes) are exposed to a wide array of self-peptides presented by a variety of thymic resident cells. Thymocytes bearing T cell receptors (TCR) that strongly engage self-ligands are deleted from the T cell repertoire during negative selection. Alternately, strong TCR signals can direct the development of thymocytes into unconventional T cell lineages that play regulatory or immunosuppressive roles, a process termed agonist selection. Given that strong TCR signaling can result in either of these divergent thymocyte fates, the factors that govern autoreactive thymocyte fate remain unclear. Here, we used an experimental system, in which thymocytes of a defined specificity are overlaid onto thymic slices and undergo a synchronized wave of development, to address this question.

First, we focused on factors that promote autoreactive thymocyte death. We have previously observed that thymocyte death and phagocytosis are concurrent in a ubiquitous model of negative selection. Here, we observed that these interactions are also prevalent in an additional model of negative selection, indicating that the close coupling of cell death and phagocytosis is a general feature of autoreactive thymocyte death in situ. We further explored the role of phagocytes in negative selection, and demonstrated that phagocytosis is critical for the induction of autoreactive thymocyte death. We also found that negative selection is most efficient when peptide is presented by phagocytic cells. Taken together, our data demonstrate that phagocytes play a more active role in negative selection than previously appreciated, both as peptide-presenting cells and as active inducers of autoreactive thymocyte death. Moreover, our findings indicate that the phagocytic ability of a peptide-presenting cell is an important determinant of autoreactive thymocyte fate.

One example of an agonist-selected lineage is TCRαβ CD8αα intraepithelial lymphocytes (CD8αα IEL), which reside in the small intestine. CD8αα IELs have been shown to play a protective role in colitis, suggesting that they are important for maintaining tolerance and homeostasis in the gut. However, the cellular and molecular interactions required to mediate CD8αα IEL development are not well defined. Here, we demonstrate that the selection of CD8αα IELs requires higher affinity TCR signals than negative selection. Additionally, we demonstrate that access to cytokines influences autoreactive thymocyte fate, as the cytokine IL-15 impairs negative selection, while IL-15 and a related cytokine, IL-2, both promote CD8αα IEL development. Furthermore, our data indicate that peptide presentation by hematopoietic cells is required for efficient CD8αα IEL development.

Taken together, these studies define factors that play important roles in determining autoreactive thymocyte fate, including the strength of TCR signal, phagocytic ability of the peptide-presenting cell, and access to cytokines. Our data contributes to mounting evidence that the context in which a thymocyte encounters peptide, shaped largely by the characteristics of the peptide-presenting cell, is a key determinant of thymocyte fate.

During positive selection, thymocytes avoid death by sensing and adjusting to homeostatic signals from self-peptide:major histocompatibility complexes (MHCs) in a process known as functional tuning. As a result, thymocytes mature into a diverse repertoire of T cells with different degrees of self reactivity, and cells with high and low self reactivity exhibit differences in effector function in the periphery. CD5 is a known tuning molecule for thymocytes with high self reactivity. However, the molecules involved in the functional tuning of cells with low self reactivity during positive selection are yet to be identified, and overall little is known about the positive selection of T cells with low self reactivity. To investigate this, we identified TG6, a functionally relevant, MHCI-restricted TCR transgenic (tg), as a novel model for studying CD8 T cells with low self reactivity. Using this model, we show that cells with low self reactivity have relatively brief TCR signals, not associated with a migratory pause, during positive selection, and that these truncated signals are correlated with delayed progress through positive selection. Cells with low self reactivity can take a week or more to mature—compared to just 3 days for cells with high self reactivity. Though all polyclonal thymocytes have increased TCR sensitivity at the beginning of positive selection, we provide evidence that cells with low self reactivity retain a gene expression signature consistent with increased TCR sensitivity late into development. Furthermore, we identify two genes, Kcna2 and Tmie, that have high expression in cells with low self reactivity and may increase TCR sensitivity. The expression of Kcna2 and Tmie is set in cells with low self reactivity during positive selection and maintained in the periphery—a characteristic expression pattern of functional tuning genes. Our results fit with a model where self reactivity directs the rate of TCR signal accumulation, TCR sensitivity, and the time to reach maturation. These differences in positive selection may be important for appropriate functional tuning of T cells, based on their self reactivity.

Abstract

Immunity to Toxoplasma gondii

by

Harshita Satija Grover

Doctor of Philosophy in Molecular and Cellular Biology

University of California, Berkeley

Professor Nilabh Shastri, Co-chair

Professor Ellen Robey, Co-chair

Toxoplasma gondii (T. gondii) is an obligate intracellular protozoan parasite, capable of infecting all warm-blooded animals, and can cause a severe disease in immunocompromised individuals. Protection to T. gondii is largely mediated by CD8 T cells, although CD4 T cells have also been shown to be important in immunity to T. gondii. However, the natural antigens recognized by these CD8 T cells, and how they mediate immune protection in vivo is poorly understood. Recently, it has been discovered that in mice with H-2^d MHC molecules, there is an immunodominant CD8 T cell response to GRA6, a dense granule T. gondii protein. In the absence of the T cell response to GRA6, such as in ERAAP-/- animals, mice succumb to infection. On the contrary, C57BL/6 (H-2^{b) mice do not mount a response to GRA6. Studying antigen presentation dependent on MHC haplotype is important because the response to T. gondii in mice is controlled by genes in the H-2 region, wherein mice with the H-2b haplotype are more susceptible to infection than mice with H-2d MHC molecules. In addition, C57BL/6 (B6) mice are good models because a large number of genetic mutations in the innate and adaptive immune system have been established in this strain.}

To better understand T cell responses in the susceptible strain, we immunized B6 mice with T. gondii, and found that it leads to potent CD4 T cell but weak CD8 T cell response. To identify the CD4 T cell stimulating antigens, we generated a T. gondii-specific, lacZ inducible, CD4 T cell hybridoma and used it as a sensitive probe to screen a T. gondii-cDNA library. We isolated a cDNA encoding a putative secreted protein of unknown function that we named CD4Ag28m and identified the minimal peptide, AS15 that was presented by MHC class II molecules to the CD4 T cells. Immunization of mice with the AS15 peptide provided protection to subsequent parasite challenge, resulting in a lower parasite burden and cyst load.

Furthermore, in order to characterize the CD8 T cell response to T. gondii in B6 mice, we restimulated T cells from mice immunized with T. gondii, with MHC Class II-/- bone marrow dendritic cells (BMDCs) to allow for the proliferation of CD8 T cells. This allowed us to generate a CD8 T cell hybridoma that was T. gondii specific and MHC Class I H-2D^b restricted. This hybridoma was used as a probe to screen the T. gondii cDNA library, which revealed that it recognizes ROP5, a rhoptry protein from T. gondii. We identified that YAL9 is the minimal antigenic peptide recognized by CD8 T cells. In chronically infected mice, response towards YAL9 in the brain and spleen is minor but detectable, and unlike the protective GRA6 response in H-2^d mice, immunization of B6 mice with YAL9 does not provide protection in B6 mice challenged with lethal dose of T. gondii. Most importantly, we found that altering the expression and trafficking of ROP5 from rhoptries to dense granules in parasites enhances the immunogenicity of ROP5 in B6 mice and now protects the mice from lethal challenge with transgenic parasites after peptide immunization.

Identification of the antigens in the H-2^b mice will enable us to study antigenic specific T cell responses, including effector functions such as clonal expansion and memory response at different stages of infection. We can now generate T cell receptor transgenic mice and use the antigen specific T cells from these mice to visualize and measure the dynamics of T cell interaction with antigen presenting cells during infection. Most significantly, understanding the nature of how these antigens are presented to T cells will allow us to design new and improved vaccine candidates against T. gondii and other apicomplexan parasites.

Aging was once thought to be an inexorable process, driven by wear and tear that steadily accumulates in cells. However, we have now come to the realization that aging is subject to modulation by genetic pathways, including the insulin/IGF-1 pathway, the mTOR pathway, and sirtuins. A commonality among these pathways is that they exert their pro-longevity and pro-health effects through responding to stressors, such as diet, oxidative stress, or protein-folding stress, and initiating a self-protective program in the cell. Sirtuins have been shown to play a major role in the cellular response to calorie restriction and oxidative stress. However, their role in protein homeostasis and in particular, the response to stress from unfolded proteins has not been well-characterized. We explored this topic by studying the role of SIRT7, a nuclear member of the mammalian sirtuin family, in responding to protein-folding stress in the endoplasmic reticulum (ER) and in mitochondria. We found that SIRT7 plays a pivotal role in the response to ER stress and mitochondrial stress by inhibiting the transcription of ribosomal subunits and mitochondrial ribosomal subunits, which results in a reduction in translation, allowing more time for proper protein folding. We found that SIRT7 mediates its effects by binding to the transcription factors Myc and NRF-1, which are master regulators of ribosomal subunits and mitochondrial biogenesis respectively. Our study suggests that the stress response mediated by SIRT7 at the cellular level may have important implications for obesity-associated diseases and hematopoietic stem cell function.

CD8 T cells restricted to the nonclassical MHC 1 (MHC 1b) molecule MHC-E show great promise in vaccine settings, but the development and specificity of MHC-E restricted T cells remains poorly understood. Here we focus on a CD8 T cell population reactive to a self-peptide (FL9) bound to mouse MHC-E (Qa1b) that is presented in response to loss of the classical (MHC 1a) processing enzyme ERAAP, termed QFL T cells. We find that mature QFL thymocytes are predominantly CD8+CD4-, show signs of agonist selection, and give rise to both CD8 and CD8 intraepithelial lymphocytes (IEL), as well as circulating memory phenotype CD8 T cells. QFL T cells require the MHC 1 subunit -2 microgloblin (2m), but do not require Qa1b or epithelial MHC 1 for positive selection. However, QFL thymocytes do require Qa1b for agonist selection and full functionality. Our data highlight the relaxed requirements for positive selection of an MHC-E restricted T cell population, and suggest a CD8+CD4- pathway for development of CD8 IELs.