Alopecia areata (AA) is a complex autoimmune disease in which autoreactive T cell-mediated attack of the hair follicle (HF) leads to non-scarring hair loss. Although AA is one of the most prevalent autoimmune diseases, the development of novel effective therapeutics has been limited. Standard of care remains observation for mild cases and steroids for moderate-to-severe cases, which have demonstrated only limited efficacy. The skin is a highly heterogeneous tissue at baseline, comprised of a diverse array of immune and non-immune cell types whose coordinated crosstalk is essential for homeostasis. The skin microenvironment becomes markedly altered as a result of disease-associated inflammation in AA. A pathognomonic histopathologic feature of AA is an intense lymphocytic infiltrate surrounding the lower portion of the HF in the growth phase of the hair cycle, known as anagen. We previously established that CD8+ T cells comprise the majority of this infiltrate in AA skin, and that they are necessary and sufficient to drive disease via JAK/STAT activation. While this discovery led to the pioneering use of JAK inhibitors as a novel class of therapeutics in AA, JAK inhibition is not a curative solution, since patients often experience relapse upon discontinuation of treatment. This not only underscores the continued need for translational drug discovery research in AA, but also reflects an incomplete understanding of the mechanisms that govern disease pathophysiology. Recent advances in single-cell RNA sequencing (scRNAseq) present an unprecedented opportunity to dissect the heterogeneity of complex tissues and disorders. Since its emergence, scRNAseq has proven to be a powerful tool for the discovery of rare cell types and novel therapeutic targets in a variety of contexts that range from cancer to autoimmunity. In this thesis, we leveraged scRNAseq to interrogate the cellular and molecular mechanisms underlying disease pathogenesis in AA at single-cell resolution, together with validation and functional experiments, with the goal of uncovering novel cell types and pathways that can guide the development of innovative therapeutic strategies. In Chapter 2, we performed scRNAseq of skin-infiltrating CD45+ immune cells to dissect lymphocyte heterogeneity in both murine and human AA. Our scRNAseq analyses informed a series of antibody-mediated cell depletion experiments that assessed the in vivo function of specific lymphocyte subsets in murine AA. Our results established CD8+ T cells as the predominant disease-driving cell type in AA. We identified shared mechanisms underlying CD8+ T cell heterogeneity in murine and human AA skin, in which CD8+ T cells form an “effectorness gradient” comprised of interrelated transcriptional states that culminate in increased expression of inflammatory cytokines and T cell effector function. We also demonstrated a role for CD4+ T helper cells in disease initiation, and determined that regulatory T cells possess intact immunosuppressive capacity in AA. In Chapter 3, we expanded upon the studies described in Chapter 2 and performed scRNAseq of skin-infiltrating CD45+ cells at various timepoints throughout disease course (from 3 to 24 weeks post-disease induction) in AA to analyze the temporal dynamics of lymphocyte heterogeneity in AA skin and skin-draining lymph nodes. In conjunction with scRNAseq, we also performed single-cell TCR sequencing to assess the dynamics of T cell clonality alongside changes in T cell transcriptional profiles. We observed a striking increase in CD8+ T cell clonal expansion during disease onset, which increased throughout disease progression and subsequently decreased in chronic AA, when the preclinical mouse model exhibits total body hair loss. Our single-cell analyses suggested that CD8+ T cell clonality and pathogenicity are closely linked, which we validated in vivo by demonstrating that a single expanded clonotypic population of CD8+ T cells is sufficient to induce disease in mice. In Chapter 4, we analyzed single-cell transcriptomic profiles obtained from full-thickness skin in mice with chronic AA to investigate the contributions of the HF and other non-T cell populations in disease. In this study, we also used a network biology-based approach to infer single-cell protein activity, which together with single-cell mRNA gene expression profiles uncovered a multitude of novel findings in AA. Our results revealed a role for necroptosis as a potential HF-intrinsic mechanism of pro-inflammatory signaling in AA, and also identified an MHC Class II signature specific to basal keratinocytes in AA skin. Furthermore, we uncovered a novel, rare population of disease-associated Arg1+ macrophages, which prompted us to revisit our immune-specific scRNAseq datasets described in Chapters 2 and 3 and perform an integrative analysis of this novel cell type in AA. Our preliminary in vivo studies suggested that targeting Arg1+ macrophages and/or arginine metabolism may ameliorate disease in AA. Taken together, this thesis presents a comprehensive, systematic interrogation of AA pathogenesis at single-cell resolution. Importantly, the validation and functional studies that were informed by our scRNAseq data demonstrate proof-of-concept of the use of single-cell technology to accelerate the discovery and translation of novel therapeutic targets in complex diseases. While we undertook a hypothesis-driven approach to design our studies, the data presented in this thesis was also profoundly hypothesis-generating, and has informed a number of ongoing projects in the laboratory with the shared goal of advancing our understanding of disease pathology in AA.

Alzheimer’s disease (AD) affects tens of millions of people worldwide. Its cause is unknown, with no cure, and disease-modifying treatment options have only recently become available. Emerging research has made a strong case for the involvement of immune cells, such as microglia and T cells, in modulating AD pathology. Newer technologies in RNA-sequencing have detailed specific phenotypic changes to microglia and T cells over the course of neurodegenerative disease. Some researchers have also used whole-genome sequencing to correlate genetic variants with changes in gene expression. However, no studies thus far have conducted this type of genotype-phenotype correlation in immune cells from aged individuals or AD patients. We have collected gene expression data from four sorted T cell subtypes in peripheral blood samples from 96 subjects in ROSMAP, a cohort of AD patients and age-matched controls. 78 of these subjects also have whole-genome sequencing data, which we used to detect genetic variants associated with changes in T cell gene expression. These are known as expression quantitative trait loci (eQTL). We found genes related to T cell cytotoxicity and immunosenescence in gene co-expression modules, among the eQTL, and in correlation with AD neuropathological traits or risk variants for several disease traits. We extended our findings related to disease association by calculating polygenic risk scores (PRSs) in our cohort from whole-genome sequencing data for 19 traits related to immune function and disease, including AD. Genes associated with the PRS for one or more disease traits often were in biological pathways related to downstream cytokine signaling, regulation of T cell receptor signaling, and T cell migration and trafficking. Overall, our findings indicate that the use of aged and AD patients in T cell genotype-phenotype correlation studies highlights genetic variants and differentially expressed genes that are not seen in studies using young, healthy individuals.

In eukaryotes, sequences that code for the amino acid structure of proteins represent a small fraction of the total sequence space in the genome. These are referred to as coding sequences, whereas the remaining majority of the genome is designated as noncoding. Studies of translation, the process in which a ribosome decodes a coding sequence to synthesize proteins, have primarily focused on coding sequences, mainly due to the belief that translation outside of canonical coding sequences occurs rarely and with little impact on a cell. However, recently developed techniques such as ribosome profiling have revealed pervasive translation in a diverse set of noncoding sequences, including long noncoding RNAs (lncRNAs), introns, and both the 5’ and 3’ UTRs of mRNAs. Although proteins with amino acid sequences derived partially or entirely from noncoding regions may be functional, they will often be nonfunctional or toxic to the cell and therefore need to be removed. Translation outside of canonical coding regions may further expose the noncoding genome to selective pressure at the protein level, leading to the generation of novel functional proteins over evolutionary timescales. Despite the potentially significant impact of these processes on the cell, the cellular mechanisms that function to detect and triage translation in diverse noncoding regions, as well as how peptides that escape triage may evolve into novel functional proteins, remain poorly understood.This thesis will describe novel findings that offer new insight into the process of noncoding translation mitigation revealed by a combination of high-throughput systems-based approaches and validated by biochemical and genetic approaches. Chapter 1 will discuss general concepts in the translation of noncoding sequences and the relevant cellular systems and impacts on human health. Chapter 2 will discuss the results of a high-throughput reporter assay investigating translation in thousands of noncoding sequences from diverse sources. The results discussed in this chapter revealed two factors involved in the mitigation of proteins derived from noncoding sequences: C-terminal hydrophobicity and proteasomal degradation. Chapter 3 will build on Chapter 2 and discuss the results of a genome-wide CRISPR/Cas9 knockout screen that identified the BAG6/TRC35/RNF126 membrane protein chaperone complex as a key cellular pathway in the detection and degradation of proteins with translated noncoding sequences. Having identified the BAG6 complex as targeting a specific reporter of translation of the 3’ UTR in the AMD1 gene, a series of knockout cell lines validated these results and demonstrated the participation of two additional genes, SGTA and UBL4A. Through coimmunoprecipitation western blots and rescue assays with flow cytometry as a readout, we confirmed physical interaction between BAG6 and the 3’ UTR of AMD1, and a similar experiment confirmed interaction between BAG6 and a readthrough mutant of the SMAD4 tumor suppressor gene. Finally, by combining our high-throughput reporter library with our BAG6 knockout cell line, we demonstrated that BAG6 targets hydrophobic C-terminal tails in many noncoding sequences of diverse origin. Finally, Chapter 4 will discuss the evolutionary perspective of noncoding translation through analyses of the sequence content of human and mouse genomes. The findings of this chapter demonstrate a significant trend for increased uracil content in noncoding regions of the genome, which frequently results in the translation of hydrophobic amino acids. We also find that many functional translated noncoding peptides localize to membranes, providing a theoretical link between the shuttling of translated noncoding sequences to a protein complex involved in membrane protein quality control and the emergence of newly evolving proteins from the noncoding genome.

Glioblastoma (GBM) is the most common and aggressive primary brain tumor and is comprised of transcriptionally heterogeneous cells and a complex microenvironment. Despite decades of research effort, few treatments significantly benefit clinical outcomes. This may be, in part, due to the lack of tools to directly measure functional responses of these heterogeneous cell types under therapy. This thesis aims to advance the understanding of cell type-specific therapeutic response by the development and application of integrated single cell imaging and RNA sequencing technology. Chapter 1 provides an overview of GBM and its heterogeneity, how investigation of cell type-specific phenotypes would benefit the development of GBM treatments, and current sequencing and imaging technologies to examine cell phenotypes with single-cell resolution. Chapter 2 presents a new microfluidic technology for joint single cell imaging and RNA sequencing that can link imaging-based phenotypes and transcriptional identity of the same individual cells with high throughput, molecular capture efficiency, linking accuracy, and user-friendliness. Chapters 3 and 4 present applications of this technology in investigating cell type-specificities of GBM treatments. Chapter 3 focuses on the specificities of 5-aminolevulinic acid (5-ALA), an FDA approved fluorogenic agent, used in fluorescence guided surgery and reveals 5-ALA labeling is not specific to transformed glioma cells, which encourages further studies to systematically compare its performance with potential alternatives. Chapter 4 focuses on the specificities of drug responses by presenting a functional drug screening approach that directly links cell states measured by apoptosis indicators with transcriptional states, which greatly enhances the interpretability of single cell-resolved drug perturbation assays.

Cell type-specific alternative splicing (AS) of pre-mRNA regulated by RNA-binding proteins (RBPs) is widespread, but particularly prominent in the brain, driving gene isoform differences between a diverse range of neuron types. While several AS programs have been shown to be critical to the function of particular neuron types, previous studies have usually been limited to one or a few RBPs and cell types, resulting in a piecemeal understanding of these regulatory patterns. Towards a comprehensive view of the neuron type-specific AS regulatory landscape, we apply current computational and experimental methods to survey neuronal AS, infer its regulation by hundreds of RBPs, and experimentally validate regulatory predictions. In Chapter 1, we examine AS in 133 transcriptomic cell types of mouse cortical neurons defined by single-cell RNA sequencing (scRNA-seq) and define neuron type-specific exons and some of their likely regulators. In Chapter 2, we leverage the rich transcriptomic dynamics of the cortical neuron dataset to systematically infer splicing regulatory network and predict RBP activity on the cell type level. We use the information theory-based method ARACNe to reverse engineer RBP-target regulatory networks and VIPER to infer differential RBP activity across neuron types in a workflow we call Master Regulator analysis of Alternative Splicing (MR-AS). RBP regulons predicted by MR-AS are consistent with high-confidence lists of RBP targets and are supported by motif and CLIP read distribution analyses. Estimation of cell type-specific RBP activity using the predicted regulons shows the expected decreases in RBP KO samples. Chapter 3 focuses on two neuron type-specific AS regulatory programs as case studies, which we validate in vitro using embryonic stem cell (ESC)-derived neuron types. Elavl2 was predicted to drive neurons towards an MGE interneuron-specific AS profile. Elavl2 knockout in ESC-derived MGE interneurons causes modulation of exon inclusion consistent with the predicted regulation of MGE interneuron AS, shifting their splicing profiles towards those of CGE interneurons. We also identified a module of exons that show consistent AS between long- and short-projection neurons across multiple neuronal classes, which are shifted in the expected direction when ESC-derived interneurons are transcriptionally reprogrammed to reflect a long-axon globus pallidus-like neuronal identity. In Chapter 4, we use the RBP regulons to predict RBP activity on a single-cell level and examine its variability, leading us to identify both neuron type-specific AS programs and a neuron type-orthogonal gradient of activity (NTOG). Exons associated with responses to neuronal depolarization and long-term potentiation show a gradient of inclusion across the NTOG, suggesting it may reflect differential activation of activity-dependent AS programs of the assayed neurons. Together, the results described in this thesis demonstrate the validity and broad utility of the inferred AS regulatory networks as a resource for elucidating RBP splicing regulation differences and their functional impact across neuron types.

Human microglia play a pivotal role in neurological diseases, but few targeted therapies that directly modulate microglial state or function exist due to an incomplete understanding of microglial heterogeneity. This thesis aims to advance our understanding of microglial heterogeneity by using single-cell RNA sequencing to profile live human microglia from autopsies or surgical resections across diverse neurological diseases and using computational tools to infer chemical and genetic regulators of specific microglial substates. Chapter 1 provides an overview of microglial ontogeny, function, and known heterogeneity, especially in disease contexts. It also describes the steadily increasing disease burden seen in neurological disease as well as the lack of efficacious treatments and future directions for microglia-targeted therapies. Chapter 2 focuses on microglial heterogeneity in an understudied disease, ALS, describing population structure shifts seen in ALS across cortex and spinal cord. Chapter 3 instead focuses on exploring underlying cross-disease microglial population structure, identifying subsets with metabolic and functional properties, as well as subsets enriched in susceptibility genes for neurodegenerative disease. We then demonstrate applications of this type of data by using our resource to annotate other datasets. Chapter 4 leverages this data in another way, by identifying and validating candidates for chemically and genetically inducing subtype-specific states in vitro. Notably, we show that Camptothecin downregulates the transcriptional signature of disease-enriched subsets and upregulates a signature previously shown to be depleted in Alzheimer’s. Finally, I review our findings and discuss future directions for the field.

Transcriptome studies disentangle functional mechanisms of gene expression regulation and may lend key insights into disease mechanisms. However, the cost of RNA-sequencing and types of tissues currently assayed pose major limitations to study expansion and disease-relevant discovery. This thesis develops methods for sampling noninvasive biospecimens for transcriptome studies, investigating their technical and biological characteristics, and assessing the feasibility of using noninvasive samples in transcriptomic and clinical applications. Chapter 1 explores the technical and biological features of four potential noninvasive sample types (buccal swabs, hair follicles, saliva, and urine cell pellets) in a pilot study of 19 individuals whereby four separate collections of each tissue were performed (i.e. 76 samples/tissue, 304 samples in total). From this data, consistency of library preparation, cell type content, replication of GTEx cis-eQTLs, and disease applications were assessed. In all, hair follicles and urine cell pellets were found to be most promising for future applications. Chapter 2 investigates the scaling potential of noninvasive sampling in SPIROMICS, a COPD clinical cohort. To do so, 140 hair follicle and 110 buccal swab samples were collected from seven different clinical sites. Consistency of sample quality was observed to be high for hair follicles, and hair cell type abundance estimates were consistent within SPIROMICS and compared to the 19 subject pilot study. Mapping of cis-eQTLs in hair revealed 339 associations not identified in any prior study. These cis-eQTLs show higher replication in GTEx tissues that share cell types with hair follicles, indicating hair follicles may indeed capture gene expression regulatory mechanisms found in more invasive tissue types of the body. This thesis suggests future use of noninvasive sampling will facilitate discovery by increasing sample sizes in more diverse populations and in tissues with greater cell type diversity and biological relatedness to disease mechanisms. Moreover, the nature of noninvasive sampling enables complex, longitudinal study designs with greater ability to capture context-dependent mechanisms of genetic regulation not currently able to be interrogated.

Prostate cancer is consistently the most frequently diagnosed cancer in American males as well as the second leading cause of cancer-related mortality. This underscores the dire need to understand the healthy prostate and how it can transform into a diseased state. Therefore, I have sought to investigate the heterogeneity and ontogeny of the mouse and human prostate. To do this, I employed single-cell RNA-sequencing, electron microscopy, immunofluorescence, and immunohistochemical analyses to identify specific cell populations, as well as lineage tracing, organoid culture, and tissue recombination assays to assess the function and origin of these populations. I discovered a profound level of heterogeneity uniquely within the luminal epithelial compartment of the prostate, including several novel populations. These luminal populations differ in distribution between mouse prostate lobes and along the proximal-distal axis within each lobe. These populations demonstrated significant differences in progenitor behavior in both organoid culture and tissue recombination assays, as shown by their differential abilities to proliferate, generate patterned structures, and differentiate into distinct cell types. Comparisons of the mouse prostate cell populations to cells from several benign human prostate samples showed that there is also luminal heterogeneity in the human prostate, and that several mouse populations have substantial gene expression overlap with human prostate populations. The observed luminal heterogeneity as well as the functional differences were consistent across several different published studies of the mouse prostate, and cross-species transcriptional similarities between populations were maintained across additional human samples, indicating that these findings are robust. My findings suggest that the luminal compartment of the mouse prostate contains distinct populations of cells that may act as reserve progenitors, and that their distribution across the prostate lobes could be functional. Additionally, if these populations can be cells of origin for prostate cancer, then their differences in progenitor behavior could contribute to the heterogeneity observed in prostate cancer prognosis and treatment response, which could have substantial clinical implications for patients.

Advances in sequencing technologies have sparked the discovery of new genetic etiologies for neurological and neurodevelopmental disorders. As new disease-causing mutations are unveiled, questions into the specific mechanisms of pathogenicity and potential therapeutic approaches arise. To address these questions, in vivo and in vitro models have been generated and analyzed; but how best to utilize these models, and how well they recapitulate the human brain, is still not fully understood. Within the work discussed in this thesis, we address this problem through the transcriptomic and functional interrogation of these models in the context of neurodevelopment and disease. In Chapter 2 of this thesis, we describe the use of single-cell RNA-sequencing to examine the longitudinal transcriptomic profiles of neuronal network establishment and maturation in ex vivo mouse cortex- and hippocampus-derived cultures. Our data highlights unique developmental transcriptomic profiles for individual genes, disease gene subclasses, and biological processes, and discusses cell population-specific divergent transcriptomic profiles between genes associated with neurological diseases, focusing on epilepsy and autism spectrum disorder. We also compared the data from our ex vivo system to transcriptomic data collected from in vivo neonatal and adult mouse brains and human cortical organoids, highlighting the importance of the generation and consideration of system-specific transcriptomic datasets when looking into a gene, disease, or biological process of interest, and serves as a vital resource for researchers. In Chapter 3, we propose a high-throughput drug discovery paradigm utilizing the application of transcriptome reversal for neurodevelopmental disorder-associated genes that affect the transcriptome. This approach describes the idea that if gene dysregulation is causal for the pathogenicity of a disease, then correcting the transcriptional signature should have a therapeutic effect. We demonstrated that small-molecule induced gene expression changes vary between both cell lines and neural cell populations, and highlight both the importance of selecting the appropriate model of disease and creating cell population-specific signatures for compounds and disease. In Chapter 4, we focus on the utilization of multi-electrode arrays for the electrophysiological characterization of primary cortical networks derived from mouse models of epileptic encephalopathy. This technique allows for the analysis of numerous neuronal and network synchronization metrics for spontaneous longitudinal activity and responses to external stimuli in the form of electrical stimulation and compound addition. In particular, mouse models with mutations in the genes Grin2a, Gnb1, and Scn1a were analyzed. We discovered significant hyperexcitability, bursting, and synchrony phenotypes, and discuss how acute and chronic compound addition can be used to interrogate biological pathways and reverse disease activity signatures.

No description available.