Carl D. Novina - US grants

DESCRIPTION (provided by applicant): Translation of mRNAs into proteins must increase to sustain the malignant progression of cancers. Increased translation and increased expression of translation machinery genes in cancers has been generally considered to be a consequence of cellular transformation. However, several recent observations challenge this concept. First, over-expression of the rate-limiting translation initiation factor eIF4E can transform multiple cell types. Second, microRNA (miRNA) repression is frequently reduced in cancers. Because miRNAs often repress oncogenes, reduced miRNA repression increases oncogene expression. Third, eIF4E has been implicated in miRNA-mediated repression and robust increases in eIF4F (a complex of eIF4E, eIF4A and eIF4G) inhibit miRNA function. Therefore, up- regulation of eIF4E may not be a passive participant in cellular transformation but may instead play an active role during oncogenesis by increasing general translation and inhibiting miRNA repression. Human melanomas are optimally suited to test the roles of up-regulated eIF4E in oncogenesis. Human melanomas frequently up-regulate eIF4E, down-regulate miRNAs, down-regulate miRNA- associated proteins including Dicer and Argonautes (Agos), and therefore up-regulate miRNA-targeted oncogenes. Additionally, we have access to a collection of 55 melanoma short-term cultures (MSTCs) that have been annotated by SNP, mRNA, and miRNA expression. Moreover, MSTCs are experimentally-tractable, grow readily ex vivo without requiring immortilization, and thus accurately reflect tumor biology and genetics operant in vivo. To define the roles of up-regulated eIF4E in melanomagenesis, we will knockdown or over-express eIF4E in MSTCs and test proliferation rates, cell-cycle progression, escape from replicative senescence, colony formation in soft agar, and invasive activity. We will then determine the effects eIF4E perturbation on general translation, miRNA repression, or both by performing metabolic labeling and miRNA reporter assays. To comprehensively identify all genes affected by eIF4E perturbation, we will perform microarray analysis of MSTCs before and after eIF4E perturbation. To identify miRNA- repressed mRNAs affected by eIF4E perturbation, we will perform microarray analysis of Ago2 immunopreciptiates. Over-expression of eIF4E should increased expression of miRNA-targeted mRNAs but should decrease Ago2 association of those mRNAs. We will functionally test eIF4E- responsive genes by knocking them down in MSTCs over-expressing eIF4E and re-testing cancer- relevant phenotypes. Dissecting the pathways that link up-regulated eIF4E expression with increased translation of untargeted and miRNA-targeted mRNAs during oncogenesis will elucidate fundamental properties of the malignant progression of melanomas and potentially uncover new therapeutic targets. PUBLIC HEALTH RELEVANCE: Translation must be up-regulated to meet the increased metabolic demands of tumor formation. Experiments in this proposal challenge the notion that increased translation is merely a consequence of cellular transformation. These studies will reveal how increased expression of eIF4E affects translation of all mRNAs, including those selectively repressed by microRNAs, to promote oncogenesis.

DESCRIPTION (provided by applicant): Diamond Blackfan Anemia (DBA) is an inherited bone marrow failure syndrome characterized by reduced erythropoiesis and increased cancer risk. Congential mutations in ribosomal protein genes (RPGs) are most closely linked with this syndrome. It is unclear why decreased RPG expression selectively decreases erythropoiesis and predisposes to cancers that typically require increased protein synthesis necessary for rapid growth. Our recent data may provide a molecular connection between reduced RPG expression and the clinical phenotypes observed in DBA. A high-throughput screen to discover effectors and regulators of human microRNA (miRNA) function identified RPGs as a novel class of genes regulating miRNA activity. Specifically, knockdown of every RPG decreased miRNA activity and thus increased expression of miRNA-target mRNAs. In this application, we similarly propose that in DBA, RPG mutations reduce miRNA activity which alters expression of developmentally-regulated genes and oncogenes (both enriched in miRNA target sites), leading to reduced erythropoiesis and increased cancer risk. To test this mechanism of decreased erythropoiesis in DBA, we will obtain DBA patient and normal donor bone marrow and tissue samples. We will generate induced pluripotent stem (iPS) cells from both sources. We will compare erythropoiesis from bone marrows and iPS from normal donors with and without RPG knockdown to bone marrows and iPS from DBA patients. We will characterize and isolate discreet erythroid progenitors from each of these sources. We will then isolate total RNA, monosomal RNA, and polysomal RNA from each progenitor and perform miRNA (Luminex bead) and mRNA (RNA-seq) expression profiling. Transcriptome and translatome maps will be generated and compared to each other and to miRNA expression profiles from matched control samples. We will use these data sets as inputs for unbiased gene network analyses to identify affected (miRNA-regulated) pathways when RPG expression is reduced. We will then confirm the roles of genes in the predicted networks by manipulating their expression in hopes of reversing erythropoiesis phenotypes in RPG knockdown and DBA patient samples. We will use siRNAs or specific miRNA mimics to knockdown genes whose mRNA levels or translation robustly increased when RPG expression was reduced. Conversely, we will use cDNAs to reconstitute genes whose total mRNA levels or translation robustly decreased when RPG expression was reduced. Our studies will establish novel connections between RPG expression, miRNA function, miRNA-targeted gene expression, and erythropoiesis, and may reveal new targets for DBA therapy.

ABSTRACT Melanoma is the sixth most common cancer in the United States, causing almost 8000 deaths annually. While melanoma is treatable when caught in the early stages, once the disease has progressed and becomes metastatic it is almost universally fatal within 5 years, so new therapeutic options are clearly needed. Circulating tumor cells (CTCs) are cancer cells shed from a primary or metastatic tumor and eventually seed new cancerous sites. In melanoma, CTCs have been found to be responsible for much of the aggressive nature of metastatic disease, including resistance to molecularly-targeted therapies. Because the virulence of individual CTCs varies, we hypothesize that unique gene expression signatures in individual CTCs is the source of drug resistance. Specifically, we hypothesize that the BRAF inhibitor vemurafenib imposes a selective pressure on melanoma CTCs which results in resistance. Unraveling the underlying mechanisms of vemurafenib resistance will require comprehensive analysis of individual CTCs during the evolution of resistance. We recently advanced technology to the point where we can robustly isolate rare melanoma CTCs and transcriptionally profile individual CTCs by single cell RNA-seq. With our innovative approaches, we can identify distinct genetic programs in melanoma CTC each with unique potential to acquire resistance to molecularly-targeted therapies. We propose to precisely annotate specific gene expression signatures in individual CTCs to the evolution of vemurafenib resistance. Melanomas are both genetically heterogeneous and clinically accessible and therefore represent an optimal cancer type for the development of analytic methods to assess the effects heterogeneous transcriptional outputs on the evolution of drug resistance. Our investigative team is ideally suited to carry out the technology development and data generation required. Historically, melanoma CTCs have been difficult to isolate and therefore a challenge to study. However, we have overcome this technical hurdle and recently determined optimal conditions for collecting melanoma CTCs from the bloodstream. For this proposal, we seek to leverage our experience, resources and capabilities to develop and implement transcriptional profiling by RNA-seq from individual CTCs isolated from melanoma patients before and after vemurafenib treatment. We will thus define transcriptional outputs in response to vemurafenib treatment, experimentally validate their roles in establishing vemurafenib resistance and thereby identify key genes as potential targets for novel vemurafenib-resistance therapy.

DESCRIPTION (provided by applicant): The advent of custom built DNA binding domains fused to restriction enzymes changed the landscape of biomedical research and heralded a new age in gene therapy. This fact is highlighted by Phase I clinical trials in which zinc finger nucleases are being used to knockout the CCR5 receptor in CD34+ T cells of HIV infected individuals to generate HIV-resistant T cells (clinical trial #NCT00842634). In addition to targeted gene knockout, other biomedical applications of the designer nucleases are gene correction and gene activation. However, gene correction requires homologous recombination following a DNA cut - a very inefficient process - and current strategies for stable gene activation require constitutive expression of an ectopic protein, limitations that are likely to hinder therapeutic efficacy. In contrast to endogenous gene editing or genetic activation, I propose to develop a platform for epigenetic reprogramming of any locus of interest. As a proof-of-concept for this technology, I propose to engineer site-specific DNA binding module fusions with DNA demethylating enzymes for epigenetic induction of fetal hemoglobin (HbF) for therapy of sickle cell disease (SCD). SCD is caused by mutation of the adult b-globin gene which forces red blood cells to sickle and occlude small blood vessels leading to exquisite pain and a wide range of medical complications. Fetal hemoglobin is normally silenced at birth by DNA methylation of the g-globin locus as adult hemoglobin increases. Because small increases in HbF can cure this disease, SCD is an optimal application for developing epigenetic re-programming technologies. The major advantage of epigenetic engineering is that successful DNA demethylation of CpGs on both strands will lead to durable HbF-induction, not requiring continuous expression of ectopic proteins. This highly-innovative strategy has no precedent in the SCD literature. Development of this technology will lead to potent and durable

Project Summary Shwachman-Diamond Syndrome (SDS) is an inherited bone marrow failure syndrome affecting multiple organ systems. However, the hematological defects in SDS remain poorly characterized. We used single cell RNA sequencing to identify increased TGF? signaling specifically in hematopoietic stem cells (HSCs) and multipotent progenitors (MPPs), but not lineage-committed progenitors. We also identified significantly reduced expression of lncRNAs that normally down-regulate TGF? signaling. Our data implicate lncRNAs in SDS bone marrow failure. To define how lncRNAs function in normal myelopoiesis and how their dysregulation is pathogenic in SDS, we will use a recently developed yeast three hybrid (Y3H) assay to systematically define lncRNA-protein interactions for 100 lncRNAs identified in normal and disease hematopoiesis. Y3H is capable of detecting low- abundance and low-affinity interactions lncRNA-protein interactions. Overlaying the lncRNA-protein interacting onto existing PPI networks will define lncRNA-regulated pathways of normal myelopoiesis and their dysregulation in SDS. Because we will screen of overlapping lncRNA fragments of 100 individual lncRNAs, these studies will enable systematic lncRNA structure-function analyses. The information obtained through this structure-function analysis will be critical for establishing sequence and/or structural parameters which may enable prediction of lncRNA-protein binding activity, thereby advancing our understanding of the biology of lncRNAs, even those not included in this initial screen. We will then validate roles for lncRNAs in hematopoiesis and altered TGF? activation. Specifically, we will use CD34+ cells depleted of SBDS and increase (by ectopic expression) or decrease (by knockdown) lncRNAs and their interacting proteins and test these effects in self renewal, myeloid lineage commitment, proliferation, cell death, oxidative stress, inflammation and TGF? activation. To further test roles for specific lncRNAs in TGF? signaling, we will also test the ability of TGF? inhibitors to rescue normal hematopoietic function. We will identify TGF?-responsive genes in affected subpopulations that could be targeted for novel SDS therapies, and test their function in SDS hematopoiesis. Thus, these studies will (1) identify lncRNAs, proteins, and lncRNA- protein interactions directing normal and SDS pathogenic myeloid function; (2) develop a platform for systemically studying lncRNAs; and (3) provide pre-clinical evidence that could support a clinical trial targeting lncRNA-directed pathways for future SDS therapy.

ABSTRACT Approximately 160 million American men, women and children are overweight or obese. Almost 75% of men, 60% of women and 30% of boys and girls under the age of 20 are obese or overweight. Obesity is associated with numerous co-morbidities including diabetes, cardiovascular disease, and cancers. There are two types of fat. Brown fat is common in newborns and small mammals. It is useful for generating heat, especially in small bodies that tend to lose heat. White fat increases as people age and is useful for storing energy and insulating against heat loss. Increased white fat is associated with increased obesity-related disease whereas increased brown fat is associated with reduced obesity-related disease. Several recent studies of adipose tissue have identified transcriptional signatures associated with brown fat adipogenesis and specifically thermogenic programs. Some of these studies have also identified long non- coding RNAs (lncRNAs) associated with white fat and brown fat identity and function. However, the mechanism(s) of how these lncRNAs affect transcriptional programs and determine brown fat identity and function are poorly understood. We propose to use our recently developed lncRNA-based yeast three hybrid (Y3H) assay to systematically define lncRNA-protein interactions for 318 conserved lncRNAs implicated in brown fat adipogenesis. Our Y3H system is capable of detecting lncRNA-protein interactions that are otherwise difficult to detect due to issues of low-abundance and low-affinity interactions. Overlaying lncRNA-protein interactions onto existing protein- protein interaction networks will define lncRNA-regulated pathways of normal brown fat adipogenesis and will also identify lncRNA-regulated pathways that might be manipulated to increase thermogenesis. Because we will screen overlapping lncRNA fragments of these 318 individual lncRNAs, these studies will enable systematic lncRNA structure-function analyses. The information obtained through this structure-function analysis will be critical for establishing sequence and/or structural parameters which may enable prediction of lncRNA-protein binding activity, thereby advancing our understanding of the biology of lncRNAs, even those not included in this initial screen. For a set of lncRNA-protein interactions, we will validate roles in mouse models of brown fat adipogenesis by increasing (by ectopic expression) or decreasing (by knockdown) lncRNAs and their associated proteins in brown fat progenitor cells. We will assess the effects of these interactions by testing obesity, brown fat lineage commitment (by FACS analysis) and thermogenesis (by transcriptional profiling). Thus, these studies will (1) identify lncRNA-protein interactions directing adipogenesis and thermogenesis; (2) develop a platform for systemically studying lncRNAs; and (3) provide pre-clinical evidence support clinical trials targeting lncRNA- directed pathways for obesity-related diseases.