Breast cancer is the most common noncutaneous form of cancer among women in the US. In recent years, the overall mortality rate has declined, yet there still exists a significant racial disparity in the incidence and mortality between African American and Caucasian women. While numerous hypotheses have been proposed to explain this difference, few offer a biological explanation. There is a well established association between estrogens and breast cancer risk, and the ratio of two estrogen metabolites, 2-hydroxyestrone (2OHE1) and 16á-hydroxyestrone (16OHE1), has been implicated as a marker of breast cancer risk. Many studies have also assessed the relationship between endogenous estrogens and mammographic density. Mammographic density is one of the strongest predictors of breast cancer risk, but the mechanism by which it influences this risk remains unknown. Nonetheless, few have examined mammographic density in relation to the 2OHE1:16OHE1 estrogen metabolite ratio (EMR). Research suggests that the Cytochrome P450 1B1 (CYP1B1) gene may also mediate breast cancer risk, as this gene is very active in estrogen metabolism. In fact, the Leu432Val polymorphism has reportedly been associated with urinary levels of the 2OHE1:16OHE1 EMR. The objective of this study was to investigate some of the relationships found among the 2OHE1: 16OHE1 EMR, CYP1B1 Leu432Val polymorphism, mammographic density, race, and breast cancer risk. The 2OHE1:16OHE1 EMR was associated with both breast cancer risk and the CYP1B1 Leu432Val polymorphism, yet, no association with breast cancer risk and this polymorphism was observed. This suggests that if the CYP1B1 Leu432Val polymorphism alters breast cancer risk, it does so through variations in the 2OHE1:16OHE1 EMR. When taking race into account, no association between mammographic density and the 2OHE1:16OHE1 EMR was observed. In culture, evidence was found to suggest that the 2OHE1:16OHE1 EMR is influenced by subcellular effects or other intrinsic factors (i.e. genetic variation), as passage number was the only significant contributor to the 2OHE1:16OHE1 EMR. The results of this study have great public health significance, as it provides a better understanding of the risk factors, including racial differences, and etiology of breast cancer, which will ultimately lead to better prevention and treatment for all women.

Cancers and infectious diseases are becoming a growing public health problem in the world today. The importance of my work for public health is in detecting these diseases earlier and more accurately, potentially leading to better therapies and higher survival rates for patients. Current diagnostic techniques focus on detecting antibodies from serum, gene expression and miRNA profiles of tumor tissues and, more recently, in the bodily fluids of patients. This dissertation shows a novel technique that makes use of small microvesicles called exosomes and the microRNAs (miRNAs) they carry for the potential diagnosis of cancer. Exosomes are small (40-100 nm) membrane-bound vesicles that are created from the inverse budding of the multivesicular endosome and originate from a variety of tumor types. Exosomes can be easily purified from cell cultures and serum of patients and have recently been shown to carry small non-coding RNAs called miRNAs. In the first part of this study, I developed techniques that enabled us to increase the amount of our exosome and total RNA starting material before proceeding to use these as potential diagnostics for head and neck cancers. Although the use of exosomes to diagnose diseases is not novel, the use of miRNAs present in tumor-derived exosomes is a new approach. In the final two chapters, I discuss the use of exosomes to diagnose KSHV viral infections as well as head and neck cancer. Increasing the accuracy and reducing the amount of starting material needed for these studies would provide a non-invasive technique to detect viral infections and cancers. This would help in providing earlier therapeutic treatments and help to increase the longevity and quality of life of patients.

Pluripotent stem cells have the capability to undergo unlimited self-renewal and differentiation into all somatic cell types. They have acquired specific adjustments in the cell cycle structure that allow them to rapidly proliferate, including cell cycle independent expression of cell cycle regulators and lax G1 to S phase transition. However, due to the developmental role of embryonic stem cells (ES) it is essential to maintain genomic integrity and prevent acquisition of mutations that would be transmitted to multiple cell lineages. Here we show that several modifications in DNA damage response of ES cells accommodate dynamic cycling and preservation of genetic information. ATM-dependent checkpoint signaling cascade is activated after irradiation of ES cells, and induces G2/M, but not G1/S cell cycle arrest. The absence of a G1/S cell cycle arrest promotes apoptotic response of damaged cells before DNA changes can be fixed in the form of mutation during the S phase, while G2/M cell cycle arrest allows repair of damaged DNA following replication. Human ES cells express higher level of DNA repair proteins, and rely on homologous recombination to repair double strand breaks. Radiation does not lead to long-term loss of pluripotency, since irradiated ES cells show transient decrease in the level of pluripotency factor transcripts, while protein levels remains stable. One week after irradiation, ES cells retain capacity to differentiate into three germ layers and form teratomas in immunocompromised mice. Similarly to ES cells, induced pluripotent stem (iPS) cells are poised to proliferate and exhibit extreme sensitivity to DNA damage, lack of G1/S cell cycle arrest, and express high level of DNA repair genes, suggesting that DNA damage responses are controlled by developmental state of the cell. Public health significance of this study originates in great promise that human ES and iPS cells hold in cell replacement therapies. Since human ES, and particularly iPS, cells represent potential source of cells for clinical and pharmaceutical applications, the DNA damage response pathways that maintain genomic integrity need to be studied in greater detail.

Coronary heart disease (CHD) is a major public health problem in western countries as it continues to be a leading cause of premature mortality and morbidity. Several risk factors contribute to CHD risk, including dyslipidemia with low high density lipoprotein cholesterol (HDL-C) and high low density lipoprotein cholesterol (LDL-C). Meta analysis of genome wide linkage analysis in families with diverse ethnicity has revealed a strong linkage with different lipid traits on chromosome 19q13.2. There are several candidate genes present under this linkage region, including APOE/C1/C4/C2 gene cluster. With the exception of APOE, other genes in this cluster have not been extensively evaluated in relation to lipid profile. Therefore, identifying APOC4 genetic variants that modulate HDL-C level is a great public health importance. In this study we focused on the APOC4 gene and hypothesized that rare and common variants in this gene could affect plasma lipid levels. Integration of common variants common disease (CVCD) and rare variants common disease (RVCD) hypotheses has been conducted in a limited number of studies. The aim of this study was to identify both common and rare variants in APOC4 by sequencing individuals having extreme low and high HDL-C levels from U.S. non-Hispanic Whites (NHWs) and African Blacks, and to examine their effects on HDL-C and correlated lipid levels. In the sequencing analysis, a total of 65 variants were identified in NHWs and African Blacks. Of these 26 were present in NHWs and 51 in Blacks. Among NHWs, 31% of the low HDL-C group had rare or less common variants versus 10% of the high HDL-C group. On the other hand, reverse trend was observed in the Black sample (46% of the low HDL-C group versus 54% of the high HDL-C). Screening of these observed rare and common variants in the complete NHWs and Blacks dataset would provide more information about their association with plasma HDL-C and correlated lipid traits.

In the United States, coronary heart disease (CHD) is the most common cause of death and number one killer of American males and females. Several epidemiological studies have identified risk factors for CHD, like low high-density lipoprotein cholesterol (HDL-C), elevated total cholesterol and low-density lipoprotein (LDL) cholesterol, and high triglycerides (TGs), but underlying genetic variations that cause predisposition to these traits still remain unclear. Lipoprotein lipase (LPL) is one of the major genes involved in lipid metabolism and its gene sequence variation has already been reported to be associated with the risk of CHD and risk of other complex diseases like dyslipidemia, type 2 diabetes, essential hypertension, and Alzheimer's disease. Unraveling the unknown genetic variation in the LPL gene in relation to HDL-C and correlated lipid traits is critically important for public health because identification of genetic markers may lead to promising future public health interventions, like prognostic tools and therapeutic approaches to alleviate the burden of CHD in the U.S. In this study, we investigated the role of common and rare variation in LPL by resequencing individuals having extremely low (n=48) and high (n=47) HDL-C levels selected from a population-based non-Hispanic white (NHW) sample of 623 individuals. A total of 179 variants were identified in 95 individuals by resequencing the entire LPL gene, including 91 uncommon or rare variants [minor allele frequency (MAF) <0.05)] and 88 common variants (MAF ≥ 0.05). Of the 91 relatively uncommon or rare variants, 21 were present only in the low-HDL group and 25 were present only in the high HDL-C group. Overall, the prevalence of uncommon or rare variants was higher in the high HDL-C than the low HDL-C group. Thirty two of the 88 common variants demonstrated significant association (P-value <0.05) between the high and low HDL-C groups. We also examined 12 common variants (MAF ≥ 0.05) in the total NHW sample and identified 7 variants to be significantly associated with lipid levels. In conclusion, our comprehensive resequencing of the LPL gene confirms that both common and rare variants in this gene are associated with interindividual variation in plasma lipid profile.

Chromosome termini form nucleoprotein structures called telomeres that consist of tandem repeats of TTAGGG DNA sequences (mammals) and telomeric proteins. Telomeres play a critical role in cell survival and genomic stability. Biochemical studies showed that the G-rich strand of telomeres can fold into secondary DNA structures called G-quadruplexes (G4-DNA), which are thought to impact telomere length regulation and telomeric DNA stability. G4 DNA structures are capable of interfering with DNA synthesis by blocking DNA polymerases in vitro and are proposed to hinder replication in vivo. We cloned telomeric repeats into reporter cassettes on shuttle vectors and replicated them in normal human somatic cells to determine if telomeric repeats induce mutations and deletions due to their ability to fold into G4 DNA structures. We demonstrated for the first time that G-rich telomeric repeats, in spite of their G4 DNA forming ability are stable upon replication in normal human cells. In contrast, ciliate telomeric sequences that form more stable G4 DNA than human telomeric sequences, induce more mutations. Stochastic telomere loss is seen in the premature aging disorder Werner Syndrome, which is caused by loss of the RecQ helicase protein WRN. We hypothesized that WRN deficiency leads to replication fork stalling and collapse due to G4 DNA formed by telomeric repeats resulting in deletions of DNA sequence. Shuttle vectors with a telomeric or control sequence were replicated in U2OS cells deficient or proficient for WRN. Replication of shuttle vectors in normal cells did not influence shuttle vector mutant frequencies, while WRN depleted cells exhibited elevated mutant frequencies for both telomeric and control vectors but the increase was significantly higher for the telomeric vector. We demonstrated that WRN is involved in suppressing mutagenesis in shuttle vectors with telomeric sequences. We are also testing DNA synthesis in plasmids through regions of single stranded DNA containing telomere repeats in WRN proficient and deficient cells. Public health significance: Shortened telomeres are associated with age related diseases such as heart disease, cancer and premature aging disorders. These assays will help us investigate factors that cause accelerated telomere loss with the goal of preventing or delaying disease.

Cardiovascular disease (CVD) is a major public health concern in the U.S., and is the leading cause of death for both men and women. Abnormal plasma lipoprotein levels, especially low high-density lipoprotein cholesterol (HDL-C) levels, are among major factors that influence the CVD risk. Functional and candidate gene association studies and recent genome-wide association studies (GWAS) have identified several genes as being potentially significant for HDL-C and other lipid (total cholesterol [TC], low-density lipoprotein cholesterol [LDL-C], and triglycerides [TG]) levels. Our group has been comprehensively investigating several HDL-C levels-associated genes using sequencing and genotyping methods to test both common and rare variant hypotheses. In this study, we sought to replicate the GWAS signals from other genes that have not been targeted by our sequencing effort in three epidemiological samples, U.S. non-Hispanic Whites, U.S. Hispanics, and African Blacks. We selected 40 SNPs (primarily those influencing the HDL-C levels) for analysis and genotyped each SNP in all populations when present at sufficient frequency (6 SNPs were not analyzed in Blacks). For 25 SNPs, we were able to replicate the genome-wide significant associations with the same lipid trait in the same direction in at least one ethnic group studied: at nominal significance (p<0.05) for 14 SNPs, with marginal p-values (0.05-0.10) for 3 SNPs, and with trend for association with p-values between 0.10-0.20 for 8 SNPs. Similarly, we were able to replicate 18 of 37 SNPs for HDL-C, 5 of 6 SNPs for TC, 5 of 7 SNPs for TG, and 2 SNPs for LDL-C levels. Two SNPs showed significant (p<0.05) but discordant results for association with the HDL-C levels as compared to those reported in the original GWAS. For 10 SNPs, we observed significant associations with lipid traits other than those reported as genome-wide significant in the original GWAS. Identification and replication of genetic associations with plasma lipid levels is relevant to public health as it may lead to improvements in prevention and treatment of dyslipidemia, which is a major risk factor for heart disease.

Ovarian cancer, the most lethal of all gynecological malignancies, represents a significant public health burden to women worldwide. The current challenges associated with ovarian cancer stem from a lack of effective screening strategies, an inability to detect the disease at a treatable stage, and the disappointing impact of treatment regimens over the entire disease course. A multi-faceted evaluation of circulating biomarkers of ovarian cancer was conducted in order to identify specific biomarkers and combinations which might serve as effective tools in the screening, triage, and therapeutic targeting of ovarian cancer patients. Ovarian epithelial carcinoma (OEC) represents a heterogeneous disease characterized by several histological subtypes displaying divergent etiology, pathology, and treatment responsiveness. Serum biomarkers were identified which displayed subtype-specific alterations in a comparison of OEC patients and benign controls. These results suggest that circulating biomarkers may assist in the selection of patients for targeted therapies. The efficient triage of women diagnosed with a pelvic mass based on risk of malignancy is known to result in a significant improvement in outcome for ovarian cancer patients and also a significant reduction in morbidity and anxiety for women with benign masses. Several multimarker panels, including the optimal combination of CA 125 and HE4, were capable of discriminating benign from malignant pelvic masses. Based on current and previous findings, this biomarker panel may represent a novel diagnostic tool in this clinical setting. Urine may offer several distinct advantages over serum as an analytical biofluid based on its low complexity, high stability, and lack of invasivity. An analysis of urine biomarkers revealed that several previously identified ovarian cancer biomarkers offer higher diagnostic performance in urine versus serum. Urine multimarker panels were effective in discriminating ovarian cancer cases from controls while a combination of urine and serum biomarkers resulted in the highest performance. The current study provides compelling evidence for the use of circulating biomarkers in several capacities within the setting of ovarian cancer. The collective impact of biomarker research on the clinical management of ovarian cancer has the potential to significantly improve overall public health.

Osteoporosis, sarcopenia and changes in fat distribution with age increase risk of fractures, affect quality of life, and are of major public health significance. Investigations into the genetic architecture of endophenotypes of these conditions could lead to better prediction of who is at greatest risk as well as revealing targets for therapies to delay disease onset or diminish their effects on afflicted individuals. Covariation among these conditions may be due to pleiotropy, although little is known about the specific genes involved. I explored relationships among twenty-two measures of arm and leg bone mineral density and geometry, arm and leg lean mass and arm and leg fat mass using data from two populations of Afro-Caribbeans from the island of Tobago: a sample of 1,937 unrelated men aged ≥ 40 years and a set of 470 men and women aged ≥ 18 years in seven extended pedigrees (mean family size = 67). I also performed genomewide association (GWA) studies of lumber spine and femoral neck bone mineral density (BMD) and fractures in an older (aged ≥ 70 years) population of European and African Americans (n = 1,663 and 1,139 respectively). Hierarchical and principal component (PC) analysis revealed three clusters: (1) a &#x201C;geometry group&#x201D; that comprises mostly bone geometry traits and lean mass (PC1); (2) a &#x201C;density group&#x201D; that comprises mostly BMD traits (PC2); and (3) a &#x201C;fat mass group&#x201D; that comprises measures of fat mass (PC3). Estimates of residual heritability ranged from 0.206 to 0.763 (p &#x3C; 0.007 for all traits). Linkage analysis revealed significant evidence (LOD &#x3E; 3.3) for quantitative trait loci (QTLs) on two chromosomes: 10q for PC1 and tibial periosteal circumference and 21q for PC3 and arm fat mass. GWA analyses of BMD and fractures in European and African Americans revealed several dozen potential candidate loci with suggestive levels of significance (p ≤ 5 &#xD7; 10&#x207B;⁶), the most promising of which is SLC4A7 on 3p24.1, a sodium bicarbonate cotransporter expressed in osteoclasts. Thus, I present evidence for specific QTLs with pleiotropic effects on multiple body composition traits, as well as loci associated with areal BMD and fracture risk. Additional analyses of these regions could reveal genes that jointly influence susceptibility to osteoporosis, sarcopenia and obesity.

Fatty acid β-oxidation disorders (FAODs), the most frequent group of inborn errors of metabolism, are clinically heterogeneous and clear genotype-phenotype correlations have not been described. Very long chain acyl-CoA dehydrogenase (VLCAD) deficiency and short chain acyl-CoA dehydrogenase (SCAD) deficiency result from mutations on the ACADVL and ACADS genes, respectively. Multiple mutations have been described in both disorders. Acute symptoms are often induced by physiological stress such as fasting, but the pathophysiologic mechanism underlying disease symptoms and phenotypic heterogeneity remains unknown. The aim of this dissertation was to explore the biological changes induced by genetic mutations and the gene-gene interactions and environmental effects on these defects. Proteomic changes in SCAD and VLCAD deficient mice, as well as the changes induced by fasting in VLCAD deficiency, were measured quantitatively using a combination of proteomics techniques. Broad mitochondrial dysfunction and derangements in multiple energy metabolism related proteins were altered in SCAD deficiency, indicating a complex mechanism for development of symptoms. Overall, a pattern associated with hepatotoxicity implicated in mitochondrial dysfunction and alteration of fatty acid metabolism was identified. Affected pathways converge on disorders with neurologic symptoms, suggesting that even asymptomatic individuals with SCAD deficiency may be at risk to develop more severe symptoms. Several candidate biomarkers were suggested through Ingenuity Pathway Analysis. Numerous proteomic changes were characterized in VLCAD deficient mice in both fed and fasting states, and relevant biological pathway were identified. Fasting in both deficient and wild type animals induced alterations in several proteins. The pattern of alterations induced by fasting was different in VLCAD deficient mice from that in wild type animals. Mitochondrial chaperonins HSP60 and HSP10 altered differently in VLCAD deficient mice depending on the feeding state. Fasting altered an apparent compensatory increase in oxidative phosphorylation seen in the fed state. Thus, environmental factors and gene-environment interaction play important roles in the pathogenesis of VLCAD deficiency. The diversity of protein changes in variable pathways due to deficiencies in SCAD and VLCAD may help explain the phenotypic heterogeneity in patients. These proteomic studies present new paradigm for exploring the mechanisms of disease, gene-environment interactions, and their contribution to gene-gene interactions in FAODs, one group of the mandatory target diseases in newborn screening program. The results advanced the knowledge about FAODs and offered the benefits for the improvement in the diagnosis and treatment of theses disorders of great public health interest. Future characterization of functionally interactive genes and association studies in humans will provide further insight into disease mechanism. Further studies on candidate biomarkers are necessary to identify novel markers for prognosis prediction, adjunct diagnosis and therapy guidance in patients.