Search

Oligodendrocytes are the myelin-forming cells in the central nervous system (CNS). Myelin is a multilayered membrane sheath structure that surrounds neuronal cell axons. Oligodendrocyte lineage cells and myelin are important for various CNS functions, such as electrical signal propagation and axonal metabolic support. Recent studies revealed oligodendrocyte lineage cells and myelin also play important roles in motor skill learning, memory and the aging process. Inherited and acquired myelination deficits such as leukodystrophies and multiple sclerosis result in severe neurological dysfunctions. At present, molecular mechanisms controlling oligodendrocyte differentiation and myelination are not comprehensively understood. Recent evidence has shown that RNA modification is a dynamic and reversible process, which adds a new layer of understanding to epigenetic regulation in biological processes. N6-methyladenosine (m6A), the most abundant internal modification site in eukaryotic messenger RNA (mRNA), is the first example of reversible RNA methylation. The conservative enrichment of m6A in stop codons, 3’-untranslated regions, and long internal exons in mice and humans suggest its fundamental importance in RNA biology. Discovery of m6A methyltransferase protein “writers,” demethylase protein “erasers,” and “readers” suggests their dynamic regulatory roles in reversible RNA modification. Many in vitro and in vivo studies have shown that reversible methylation of mRNA is essential for cell survival and cell differentation. Compared to DNA and protein methylation, RNA modification has the poteintial to have rapid influence on mRNA metabolism and gene expression, and thereby contribute to rapid cellular responses to cellular and environmental cues. In the CNS, studies have found m6A mRNA methylation is critical for embryonic neruogenesis and neural repair. In addition, m6A mRNA methylation plays a regulatory role in various CNS functions such as learning and memory. As a glia population, oligodendrocyts share the same embryonic stem cell origin with neuorns, and actively interact with neurons to perform CNS functions. Therefore, it is highly important to understand whether m6A mRNA methylation plays a role in regulating oligodendrocytes lineage progression and function. This thesis will focus on understanding the potential role that m6A mRNA regulation plays in oligodendrocyte development, CNS myelination, and remyelination. Chapter 1 will provide an overview of the current knowledge of molecular mechanisms regulating oligodendrocyte lineage progression. Chapter 2 will describe our discovery of m6A mRNA modification mechanism in regulating oligodendrocyte lineage development and the myelination process. In mouse models that specifically inactivate an essential m6A writer component, METTL14, we found an abnormal myelination process and disrupted oligodendrocyte differentiation. We also discovered that Mettl14 deletion in oligodendrocyte lineage cells causes aberrant splicing of myriad RNA transcripts and disruption of gene expression, including an essential paranodal component neurofascin 155 (NF155). Chapter 3 will introduce the findings of m6A mRNA modification in the remyelination process in adult animals. We found that Mettl14 deletion in oligodendrocyte precursor cells (OPCs) prevents proper remyelination after the demyelination event. Chapter 4 will discuss how we can further enhance our understandings based on the findings in chapters 2 and 3, in order to gain a complete mechanistic picture of oligodendrocyte biology and promote therapeutic discoveries for myelin diseases.

This dissertation is devoted to the maximal L1-in-time regularity for a class of linear parabolic systems with variable coefficients. This theory can be applied to investigate the global-in-time well-posedness and stability issues for density-dependent viscous fluids, even if the initial fluctuation of the density is large. The results in Chapter 3 and most of the results in Chapter 4 have been addressed in the author's papers [56] and [57], respectively. The main result in Chapter 3 concerns the maximal L1 regularity and asymptotic behavior for solutions to the inhomogeneous incompressible Navier-Stokes equations under a scaling-invariant smallness assumption on the initial velocity. We obtain a new global L1-in-time estimate for the Lipschitz seminorm of the velocity field without any smallness assumption on the fluctuation of the initial density. In the derivation of this estimate, we study the maximal L1 regularity for a linear Stokes system with variable coefficients. The analysis tools are a use of the semigroup generated by a generalized Stokes operator to characterize some Besov norms and a new gradient estimate for a class of second-order elliptic equations of divergence form. In Chapter 4, we generalize the concrete maximal L1 regularity result obtained in Chapter 3 and establish an abstract one for a class of Cauchy problems associated with composite operators. Then we apply this abstract theory to study maximal L1 regularity for the Lame system with rough variable coefficients. To lower the regularity of the coefficients, we work in the Lp (in space) framework. For this, we use a classical method to establish Gaussian bounds of the fundamental matrix of a generalized parabolic Lame system with only bounded and measurable coefficients. As applications, we use a Lagrangian approach to study the global-in-time well-posedness of systems of compressible flows.

Network analysis has been playing an important role in bioinformatics analysis of transcriptomic data. By identifying differentially expressed genes that are “connected” with other differentially expressed genes, or genes that are not differentially expressed but “connected” with differentially expressed genes, it is meant to associate genes and phenotypes that simple differential expression analysis cannot do. Despite its popularity, the false positive discoveries in network analysis of transcriptomic data has not been systematically reviewed to date. In this study, we define false positive rate for each gene as its probability of being implicated in the null datasets, and seek to identify potential factors that could contribute to the inflated false positive rate for network analysis. By evaluating representative subnetwork detection, network propagation, and heterogeneous network analysis methods for transcriptomic data, we first demonstrated the inter-gene correlation, network node degree, and heterogeneous node association as the potential factors. We then proposed and evaluated two methods to reduce the bias caused by node degree and inter-gene correlation respectively and demonstrated their utility. The node degree and the inter-gene correlation are intrinsic factors of the biologic network and the transcriptomic data. Our solution to reduce the false positive rate caused by these factors are generalizable to other network analysis methods and transcriptomic data.

Arsenic, a ubiquitous environmental contaminant, is known to cause immunotoxicity. However, the mechanisms of arsenic induced immunotoxicity are still not clear. Arsenite (As+3), the trivalent inorganic form of arsenic, was found to increase DNA damage and inhibit PARP activity in thymus cells at 50 nM in vitro. Oxidative stress and double strand breaks (DSBs) formation occurred in thymus cells exposed to As+3 at 500 nM. The genotoxicity was found to be related to As+3 exporter expressions and reversed by zinc supplement. IL-7 signaling, the critical cytokine signal for the early DN thymic T cell proliferation, was found to be suppressed by As+3 at 500 nM and MMA+3 at 50 nM in vitro. The suppressive effects reduced the cell cycle gene expression. After 30 d in vivo exposure of up to 500 parts per billion (ppb) As+3, thymus cells were found to contain the higher levels of arsenic than bone marrow and spleen cells. Correlations between intracellular arsenic concentrations and the increase of DNA damage were found in both the bone marrow and thymus, but not the spleen. Monomethylarsonous acid (MMA+3) was also found to be more genotoxic than As+3 in vitro in bone marrow, spleen and thymus, and was demonstrated to be the most prevalent form of arsenic in bone marrow and thymus after in vivo exposures. Benzo(a) pyrene-7,8-dihydrodiol (BP-diol), the metabolite of benzo(a) pyrene (BaP) was found to interact with As+3 synergistically at low concentrations to increase the DNA damage and inhibit PARP activity in primary thymus cells. The interactive effects were also demonstrated to increase the expression of CYP1A1 and CYP1B1, two critical enzymes for the metabolism of BP-diol to the DNA adduct forming compound, benzo(a) pyrene-7,8-dihydrodiol-9,10-epoxide (BPDE). The differential sensitivity to As+3 of double negative (DN) and double positive (DP) thymus cell subsets was also evaluated and compared. Higher intracellular accumulation of As+3 was found in DN cells, with the higher DNA damage than DP cells after in vitro As+3 treatments. The expressions of two As+3 exporters, Mdr1 and Mrp1, were shown to be inversely correlated with the selective genotoxicity in DN cells. These studies contributed to our understanding of the mechanisms associated with the toxicity induced by environmentally relevant concentrations of arsenic and its potential interactions with other environmental contaminants.

Surfactant protein C (SP-C) is a major protein component of pulmonary surfactant. SP-C is exclusively expressed in Type II epithelial cells of the lung and is encoded by the SFPTC gene whose transcripts direct the synthesis of a 197-amino acid proprotein. Trafficking through the regulated secretory pathway results in proteolytic processing of the proprotein in multivesicular bodies and release of the 35-amino acid mature peptide. Proteins involved in recognition and internalization of SP-C into the multivesicular body have not been identified. The first hypothesis of this thesis is that the PY motif in the N-terminal propeptide of SP-C binds with the WW domains of Nedd4, and ubiquitination of the lysine residues on SP-C by the Hect domain of Nedd4 acts as the internalization signal. Although Nedd4 was detected in Type II cells, processing of a transfected PY mutant of SP-C was not affected. This result is contrary to the hypothesis; however, we cannot exclude the possibility that proprotein processing is independent of internalization into the multivesicular body. It will be necessary to assess SP-C secretion as the endpoint and this can only be done in type II cells. HEK cells transiently transfected with SP-C1-197 or SP-CΔexon4 showed distinct trafficking and processing patterns of the proteins, suggesting that HEK cells can discriminate between correctly folded and incorrectly folded SP-C. The second hypothesis of this thesis is that the quality control receptor for non-glycosylated proteins like SP-C can be identified by comparing SP-C binding proteins and SP-CΔexon4 binding proteins. Scale up failed to identify unique SP-CΔexon4 binding proteins suggesting that insufficient cell lysate was applied to the column. Radiolabeled SP-C with its binding partners was immunoprecipitated from stably transfected HEK 293 cells. Several binding proteins were uniquely associated with SPCΔexon4 thus providing proof of principle.