Browsing by Author "Wechsler, Daniel S"

The dissertation will focus on deciphering novel roles for PICALM in cellular biology. PICALM (Phosphatidyl Inositol Clathrin Assembly Lymphoid Myeloid Protein) is a ubiquitously expressed protein that was initially identified as a partner for AF10 in a chromosomal translocation in a lymphoma cell line. Since its identification, PICALM has been shown to act as an accessory adaptor protein in clathrin-mediated endocytosis and to regulate the internalization of proteins involved in vesicular trafficking (SNARE proteins). In addition, mutations in the PICALM gene have been shown to be linked to the development of leukemia and Alzheimer’s Disease. As a result of our studies, we have determined that PICALM is involved in two previously unappreciated cellular processes: macroautophagy and cellular cholesterol metabolism. This dissertation will address each of these processes in turn.

The thesis begins with an introduction to PICALM, including a description of PICALM’s known cellular functions and its relationship to disease. In addition, general aspects of macroautophagy and cellular cholesterol metabolism will be introduced (Chapter 1). Chapter 2 will describe the materials and methods that were used in the experimental analysis.

Chapter 3 describes our observation of a novel role for PICALM in macroautophagy. PICALM regulates SNARE protein internalization and localization. Intriguingly, SNARE proteins (VAMP3 and VAMP8) are involved in vesicular trafficking and macroautophagy. Thus, we sought to determine a role for PICALM in regulating macroautophagy by experimentally reducing or overexpressing PICALM. Our studies show that both reduction and overexpression of PICALM can modulate macroautophagy. In addition, our work indicates that PICALM modulates macroautophagy by altering autophagosome breakdown, without having an effect on autophagosome formation. This section of the thesis concludes with a possible mechanism by which PICALM may modulate macroautophagy. A substantial portion of this Chapter appeared in Moreau et al, Nature Communications, 2014 (1).

Chapter 4 focuses on PICALM’s ability to modulate cellular cholesterol homeostasis. We initially performed a microarray experiment using picalm-deficient and PICALM-expressing cells in order to obtain biological insight into possible novel roles for PICALM. This study suggested that modulating the level of PICALM expression alters cellular cholesterol homeostasis. We went on to demonstrate that PICALM reduction and overexpression result in altered cholesterol metabolism gene expression. In addition, we examined the effect of PICALM deficiency on cholesterol flux, and unexpectedly showed that PICALM reduction results in elevated cholesterol internalization, and cellular cholesterol levels. The LDL receptor is the primary route by which cholesterol is internalized. Thus, we measured LDL receptor internalization by flow cytometry. We showed that internalization of the LDL receptor is elevated in the absence of PICALM. This portion of the thesis concludes with a possible mechanism by which PICALM alters cellular cholesterol metabolism. The majority of this Chapter appeared in Mercer et al, PLoS ONE, 2015 (2).

Finally, Chapter 5 summarizes our observations and discusses the relationship among PICALM, macroautophagy and cellular cholesterol metabolism. In addition, future directions of these projects and how these studies are relevant to disease will be discussed.

The t(10;11) chromosomal translocation gives rise to the CALM-AF10 fusion gene and is found in patients with aggressive and difficult-to-treat hematopoietic malignancies. CALM-AF10-driven leukemias are characterized by a perturbed epigenetic and transcriptional state. Specifically, the HOXA cluster genes are hypermethylated on Histone H3 lysine 79 (H3K79), which corresponds with their transcriptional upregulation. Conversely, CALM-AF10 cells display global H3K79 hypomethylation. DOT1L, the H3K79 histone methyltransferase, interacts with the OM-LZ domain of AF10, and the AF10 OM-LZ domain has been shown to be necessary and sufficient for CALM-AF10-mediated transformation. These data have suggested a critical role for the AF10-DOT1L interaction in CALM-AF10 leukemias. However, the mechanism(s) by which DOT1L-mediated epigenetics are perturbed and the precise role of CALM in leukemogenesis have remained unclear.

In this dissertation, we examine the contribution of CALM to CALM-AF10-mediated leukemogenesis. We determine that CALM contains a functional nuclear export signal (NES) that mediates steady-state cytoplasmic localization of CALM-AF10. An NES is a highly conserved leucine-rich amino acid sequence that is recognized by the nuclear export receptor, CRM1. Classically, CRM1 binds to NES-containing proteins and mediates their export from the nucleus to the cytoplasm through the nuclear pore complex. Through structure-function analyses, we determine that the CALM-derived NES is necessary and sufficient for CALM-AF10-dependent leukemogenesis. In addition, fusions of NES motifs from heterologous proteins (ABL1, Rev, PKIA, and APC) in-frame with AF10 are sufficient to immortalize murine hematopoietic progenitors in vitro. From these data, we conclude that a CRM1-dependent NES represents the functional contribution of CALM for CALM-AF10-mediated leukemogenesis.

In the second part of this dissertation, we examine the mechanism(s) by which the CALM NES imparts transformation potential to AF10. We determine that the CALM NES is essential for CALM-AF10-dependent Hoxa gene upregulation and aberrant H3K79 methylation. Using co-immunofluorescence microscopy, we observe increased cytoplasmic localization of DOT1L in the presence of CALM-AF10, suggesting that mislocalization of DOT1L may lead to a global loss of H3K79 methylation. In addition to mediating nuclear export, we find that the CALM-CRM1 interaction is critical for targeting CALM-AF10 to the Hoxa locus. Inhibition of CRM1 with Leptomycin B prevents transcription of Hoxa genes in CALM-AF10 leukemia cells. These findings uncover a novel mechanism of leukemogenesis mediated by the nuclear export pathway and support further investigation of the utility of CRM1 inhibitors as therapeutic agents for patients with CALM-AF10 leukemias.

Brain tumors are an incredibly diverse group of neoplasms, as evidenced by their varied locations in the brain, histological characteristics, and genetic alterations. Brain tumor heterogeneity can be potentially explained by distinct oncogenic events or cells-of-origin, or by region-specific intrinsic or extrinsic factors. Brainstem Glioma (BSG) is a particularly deadly brain tumor, afflicting 200-300 children in the United States each year. High-grade BSG (also known as Diffuse Intrinsic Pontine Glioma, DIPG) cannot be surgically removed, and the standard treatment of radiation therapy provides only temporary relief from symptoms. The past 5 years has witnessed a dramatic increase in knowledge regarding the biological basis of this disease along with the realization that BSG is distinct from other more common types of glioma, such as cerebral cortex glioma (CG). It was the goal of this study to investigate the regional differences in gliomas arising in the brainstem versus the cerebral cortex, using mice as a model system, and to begin to understand the contributions of the various possible sources of heterogeneity.

In doing so, we have uncovered region-specific gene expression patterns in these two types of pediatric gliomas that are apparent even when the initiating genetic alterations and cell-of-origin are kept constant. Focusing on the paired box 3 (Pax3) gene, which is expressed at higher levels in BSG than CG, we have found that Pax3 expression not only characterizes mouse BSGs driven by PDGF signaling, Ink4aARF-loss, p53-loss, and H3.3-K27M expression, but also identifies a novel subset of human BSGs that are associated with PDGFRA alterations and wild type ACVR1 and that commonly harbor TP53 alterations and the H3.3-K27M mutation.

As Pax3 plays a pro-tumorigenic role in other types of cancer, we hypothesized that Pax3 expression contributes to the brainstem gliomagenesis process as well. By utilizing mouse models, we found that Pax3 inhibits apoptosis and promotes proliferation of Nestin-expressing brainstem progenitor cells in vitro and enhances PDGF-B-driven BSG in vivo. Furthermore, we speculate that Pax3 expression may be a marker for Wnt pathway activation in BSG, which is targetable via pharmacologic agents. Indeed, a subset of Wnt inhibitors tested effectively slowed the growth of BSG cells in vitro, however cross talk with the Shh pathway might indicate that dual Wnt and Shh inhibition is necessary.

In addition, the regional expression pattern of Pax3 in gliomas correlates with its expression in normal murine brain development, leading us to hypothesize that Pax3 progenitor cells in the neonatal brainstem can serve as a cell-of-origin for BSG. We discovered that targeting Pax3 progenitors with PDGF-B overexpression and Ink4aARF- or p53-loss induces high-grade BSG that physiologically resemble the human disease. This novel and distinct model of BSG may be utilized in the future for preclinical studies.

The identification of Pax3 as a regional marker of mouse and human BSG has led to the discovery of a novel subset of the human disease, the identification of a novel oncogene contributing to pathogenesis, and the characterization of a novel cell-of-origin with the potential to give rise to the disease. This information contributes significantly to the current understanding of the mechanisms and cellular origins of BSG, and will hopefully instruct future investigations into how to better treat this disease.

Diffuse Intrinsic Pontine Glioma (DIPG) is a highly aggressive pediatric brainstem tumor recently found to contain high incidence of H3.3K27M mutations. These mutations reprogram the H3K27me3 epigenetic landscape of DIPG by inhibiting the H3K27-specific histone methyltransferase EZH2. This leads to global reduction with focal gains of the repressive H3K27me3, a mark responsible for cell fate decisions. To date the tumor-driving effects of H3.3K27M remain largely unknown. We show H3.3K27M cannot form tumors alone, however it cooperates with PDGF signaling in vivo, enhancing gliomagenesis and reducing survival of p53 WT and knockout murine models of DIPG. We find H3.3K27M expression drives increased proliferation of tumor-derived murine neurospheres, suggesting cell cycle deregulation may contribute to increased malignancy in mutant tumors. RNA-Seq on tumor tissue from H3.3K27M expressing mice showed global upregulation of PRC2 target genes, and a subset of newly repressed genes enriched in regulators of development and cell proliferation. Strikingly, H3.3K27M induces targeted repression of the p16/ink4a locus, a critical regulator of the G0/G1 to S phase transition. We find increased levels of H3K27me3 at the p16 promoter, however pharmacological reduction of this promoter methylation does not rescue p16 expression. While DNA methylation is also present at this promoter, it is not K27M-dependent. Intriguingly, inhibition of DNA methylation restores p16 levels and is cytotoxic against murine tumor cells. Importantly, we show that H3.3K27M-mediated p16 repression is an important mechanism underlying the proliferation of H3.3K27M tumor cells as in vivo cdkn2a knockout eliminates the survival difference between H3.3K27M and H3.3WT tumor-bearing mice.