No abstract available. / Thesis (M.Med.)--University of KwaZulu-Natal, Durban, 2014.

Streptococcus pyogenes infections can give rise to a diverse array of long-term secondary sequelae, including those in the brain characterized by both motor and neuropsychiatric disorders: Sydenham’s chorea and Pediatric Autoimmune Neuropsychiatric Disorders Associated with Streptococcus infections (PANDAS). These conditions are thought to be mediated by neuroinflammatory responses and autoantibody entry into the brain, but the mechanisms are not well understood.Previous work by our laboratory has demonstrated that recurrent intranasal S. pyogenes (Group A Streptococcus, or GAS) inoculations in mice cause infiltration of CD4 T cells into the anterior brain, disruption of the blood-brain barrier (BBB), increased numbers of activated myeloid cells and degradation of excitatory synapses leading to neural circuitry deficits. However, the molecular mechanisms underlying these phenotypes have not been fully explored. To understand how the neurovasculature and myeloid cells respond to recurrent GAS infections at the transcriptome level, I profiled cells from mouse olfactory bulb (OB) and nasal lymphoid tissue by single-cell RNA sequencing (scRNAseq). I found marked shifts in both endothelial cell and microglia populations at the transcriptome level after GAS infections, including downregulation of BBB-associated transcripts by endothelial cells (ECs), and increased production of inflammatory cytokines and chemokines, type I interferon response, and antigen presentation genes by microglia (Chapter 3). I validated several differentially expressed genes using flow cytometry, immunosorbant assays, RNA fluorescence in situ hybridization (FISH), and multiplexed error-robust FISH (MERFISH). Single-cell spatial transcriptomics of the OB revealed regional heterogeneity among microglial responses to GAS, possibly driven by proximity to infiltrating T cells. Analysis of transgenic CX3CR1/TMEM119 dual myeloid reporter mice confirmed that perivascular and meningeal macrophage numbers increase in response to GAS, but, unlike in other neuroinflammatory diseases, few macrophages infiltrate the brain parenchyma. Our laboratory has previously shown that Th17 cells are critical for BBB damage and activated microglia in response to repeated intranasal GAS infections, but the contribution of T helper (Th) 17 cell-derived cytokines in this process, as well as the transcriptional effects of Th17 cells on endothelial cells and microglia are unknown. To expand on these findings, I performed scRNAseq on retinoic acid-related orphan receptor γt (RORγt) mutant mice (Chapter 4) which showed a significant rescue in BBB-associated genes (e.g. Mfsd2a, Itm2a and Itih5) in endothelial cells. Chemokine production and type I interferon gene expression by microglia was also significantly rescued in RORγt mutants; surprisingly antigen presentation by microglia in response to GAS was exacerbated, at both the gene and protein level. Interleukin (IL)-17A is a major cytokine produced by Th17 cells. To examine the role of IL-17A in disease pathogenesis, I treated wild-type mice with an IL-17A neutralizing antibody during the course of GAS infections (Chapter 4). This treatment was sufficient to recapitulate the transcriptional effects on microglia and endothelial cells, as well as rescue BBB permeability previously found in RORγt mutants, indicating that IL-17A may play a critical role in transcriptional responses of endothelial cells and microglia to recurrent GAS infections in vivo. However, IL-17A did not disrupt tight junctions or induce transcytosis on ECs in vitro, suggesting that its effects on ECs in vivo are indirect. Th17 cells are capable of considerable phenotypic plasticity in response to chronic inflammation. To understand this process during recurrent GAS infections, I performed a time course analysis of CD4 T cell subsets after two, three, four and five infections (Chapter 5). This analysis revealed that proportions of “pathogenic” interferon γ-expressing Th17 cells increased over time, as did the number of CD4 T cells expressing granulocyte-macrophage colony stimulating factor (GM-CSF), a cytokine with pleiotropic effects on autoimmunity. Moreover, I determined that RORγt mutants have decreased proportions of GM-CSF+ CD4 T cells in their nasal mucosa, raising the question of whether GM-CSF may also contribute to CNS pathology (BBB permeability or microglial activation) in addition to IL-17A. To address this question, I generated mice deficient for GM-CSF in T cells and found that conditional deletion of GM-CSF in CD4+ cells partly rescued type I interferon and antigen presentation responses in microglia by scRNAseq, but did not rescue BBB leakage, suggesting that GM-CSF and IL-17A have distinct roles in the neurovascular and neuroinflammatory responses to GAS. To relate the findings in mice to the human disease, in Chapter 6 we performed cytokine profiling in sera from PANDAS/PANS patients at the acute phase of the disease using a multiplex bead-based immunoassay. We found that many chemokines and cytokines produced by activated microglia or macrophages in the mouse model were also highly elevated in the sera of PANDAS/PANS patients. These findings suggest an important link to the human disorder both to understand disease mechanisms in humans and to use them as future clinical biomarkers for diagnosis and treatment monitoring.

Neurovascular injuries are leading causes of disability implicated in neurological dysfunction. Much of the Central Nervous System (CNS) homeostasis depends on concerted signaling between neurons, glial cells, and vasculature–the neurovascular unit (NVU). Neurovascular injuries disrupt the NVU causing hypoxia, ischemia, neuroinflammation, and neuronal death. Much of the neuroinflammatory responses associated with neurovascular injuries have been characterized, but the contribution of specific signaling pathways from the injured endothelium to inflammatory response remains to be established. To understand vascular-glial communication in the context of vascular injury, the Troy lab has used a mouse model of retinal vascular injury, retinal vein occlusion (RVO). The retina is a CNS enclosed tissue that allows live visualization of vascular and neuronal condition upon injury, genotype, and/or treatment. Previous studies in the laboratory determined that non-apoptotic expression of endothelial caspase-9 (EC Casp9) was key for the development of retinal edema, capillary ischemia, and neuronal death. Caspases are known for their role in mediating cell death, but how and if glial cells orchestrated outcomes remain unknown. This thesis work aimed to investigate the role of caspase-9 signaling in vascular-glial communication and its contribution to pro-inflammatory cytokine levels and neurodegeneration in neurovascular injury. To answer this, we first optimized the mouse model of RVO and profiled the levels of caspases in RVO retinas treated or untreated with a caspase-9 inhibitor using immunohistochemistry. Then, we used tamoxifen inducible endothelial and astroglial caspase-9 KO lines, subjected them to RVO and measured glial changes, cytokine levels, capillary ischemia, retinal edema, neuronal death, and vision dysfunction. We first found that RVO induces a range of cell-specific levels of caspases and that inhibition of caspase-9 specifically modulated the levels of endothelial caspase-9 and 8, neuronal caspase-9, 7, and 6, astroglial caspase-6, and leukocytic caspase-9 and 7. Our studies also suggest that endothelial caspase-9 induces a decrease in reactive microglia, inflammatory cytokines, cleaved- caspase-6 and GFAP cleavage in astrocytes. EC Casp9 deletion also altered changes in GFAP, nestin and AQP4 levels in Müller glia. Through investigating an astroglial caspase-9 KO, we discovered that astroglial caspase-9 could be upstream of astroglia caspase-6. Additionally, we found that astroglial caspase-9 loss protected hypoxic retinas from capillary ischemia but not from retinal edema nor neuronal death. Lastly, we used an optokinetic test to study the potential role of endothelial and astroglial caspase-9 in RVO-induced vision disfunction. Our results indicate that removing caspase-9 from endothelial cells or astrocytes protected contrast sensitivity damage in visual function one day post-RVO. In sum, the present thesis work demonstrates that endothelial and astroglial caspase-9 signaling can lead to inflammation and worsening of visual function in neurovascular injury.

Central nervous system (CNS) tissues are highly metabolically active which makes them particularly susceptible to vascular injury. Disruption to the supply of oxygen and nutrients by damaged vasculature can result in neurodegeneration in both the eye and brain. The retina is an accessible part of the CNS that can be taken advantage of to study neurovascular diseases through live, non-invasive visualization of vascular and neuronal conditions upon injury. Retinal vein occlusion (RVO) is a common neurovascular disease of the eye and is the second leading cause of blindness in working age adults. While pathophysiology is well described and can be determined by retinal edema, breakdown of the blood-retina-barrier (BRB), inflammation, and neurodegeneration, the underlying signaling pathways behind the pathology is not well understood. To understand the mechanism of disease in RVO, the Troy lab has employed a mouse model to investigate pathways. Previous studies in the lab determined that as early as 1 hour post RVO, there was a large induction of caspase-9, a known cell death protease, in endothelial cells. When further investigated, it was confirmed that these cells were not dying despite the high expression of caspase-9, implying a non-apoptotic role. Deletion of endothelial caspase-9 was sufficient to protect against the development of retinal edema, capillary ischemia, and neuronal death, indicating caspase-9 is a key player in the mechanism of disease. This thesis work aims to investigate which signaling events drive non-apoptotic endothelial caspase-9 signaling by investigating upstream and downstream mechanisms of endothelial caspase-9. To interrogate this question, the mouse model of RVO was optimized, limiting the variability previously observed to ensure accurate and reproducible results. Then, we used a tamoxifen inducible endothelial cell Apaf-1 (apoptosis protease activating factor-1) knock out (Apaf-1 iECKO) mouse line in order to investigate the contribution of upstream activation of non-apoptotic endothelial caspase-9 signaling. Apaf-1 iECKO mice and WT littermates were subjected to RVO. Then, expression of caspase-9 and -7, retinal edema, capillary ischemia, neuronal death, vision dysfunction, and BRB integrity were measured. The deletion of endothelial Apaf-1 resulted in reduced expression of cl-caspase-9 and caspase-7, indicating endothelial caspase-9 was activated by Apaf-1. Apaf-1 deletion also resulted in protection against some of the pathologies seen after RVO including retinal edema, capillary ischemia, and neurodegeneration. Lastly, in order to elucidate the signaling pathway further, experiments using endothelial cell-specific AAVs (adeno-associated virus) packaged with a downstream caspase-7 inhibitor were proposed and described. In sum, this thesis work reveals that endothelial caspase-9 is canonically activated by Apaf-1, but still leads to non-apoptotic signaling, indicating downstream caspase-9 substrates could be the source for non-apoptotic function within endothelial cells.