Christy L. Avery, Ph.D. - US grants

ABSTRACT This is a NIH Pathway to Independence Award (K99/R00) grant proposal, intended to promote the career of Dr. Christy Avery, PhD, a post-doctoral fellow at the University of North Carolina, into a path of independent research. The Candidate is a trained epidemiologist with a significant track record of research in the field of cardiovascular disease epidemiology. Her goal is to integrate these skills with focused training in clinical research to become an independent scientist at the interface of basic science, clinical practice, and population research. The topical area for the proposed experiential development is translational research in heart failure. During the mentored K99 phase the Candidate will expand her clinical research skills by interacting with clinical specialists who are productive scientists in heart failure research, engaging in formal didactics, attending multidisciplinary seminars, journal clubs, and scientific meetings, participating study activities with her mentors, and by leading a project that develops novel heart failure classification tools applicable to clinical practice and population research. These activities will be supervised by mentor Dr. Gerardo Heiss MD PhD, Kenan Professor of Epidemiology and co-mentor Dr. Patricia Chang, MD MHS FACC, Assistant Professor of Medicine. In addition, a team of collaborators with complementary areas of expertise will supplement Dr. Avery's training in specific areas. Together, the mentor, co-mentor, and collaborators are fully committed to assisting the Candidate reach her research training and career development goals and to ensuring the Candidate's successful transition from postdoctoral fellow to independent researcher. During the independent R00 award phase, the Candidate will apply the tools developed during the K99 component in the cohort of the ongoing Atherosclerosis Risk in Communities (ARIC) Study. For this purpose the ARIC study investigators are making available data on the cohort's follow-up calls, their physician visits, emergency department visits and hospitalizations, and the associated diagnoses. This information base will be combined with outpatient and inpatient Medicare claims data and reports by participant's treating physicians to estimate the burden of heart failure and its patterns of diagnosis and care in ARIC Study participants. The Candidate also will characterize the transition from heart failure managed in outpatient settings to acute/decompensated heart failure events requiring hospitalization. The proposed research is both novel and of major potential for translation, with possible relevance for early diagnosis and proactive management of heart failure as well as for health services allocation and program evaluation. The short term benefits from this study will be a contribution to the under- studied area of quantifying the outpatient burden of HF and its secular trends in communities.

DESCRIPTION (provided by applicant): Ideal cardiovascular health (CVH) is a novel concept recently defined and adopted by the American Heart Association to set national goals for health promotion and disease reduction through 2020 and beyond. CVH emphasizes the maintenance of a low-risk profile throughout the life course and is operationalized through measurement and classification (ideal; intermediate; poor) of seven CVH metrics (nonsmoking, body weight, physical activity, diet, total cholesterol, blood pressure, and blood glucose). Although maintenance of ideal CVH across the lifespan may eliminate ¿ 70% of the U.S. cardiovascular disease (CVD) epidemic, fewer than 1% of Americans successfully maintain ideal levels of all seven CVH metrics. Suboptimal CVH profiles are especially prevalent among U.S. minority groups, who attain intermediate and poor levels of CVH metrics earlier in life than their non-minority counterparts. However, the ages at which members of minority populations transition between ideal, intermediate and poor levels of CVH metrics have not been described. The fact that early transitions from ideal to intermediate and poor levels of CVH metrics have been strongly and consistently associated with cardiovascular morbidity and mortality lends some urgency to this line of health disparities research. A parallel need is estimation of the populatio-wide benefits of modest, but achievable increases in the probability of maintaining ideal levels of CVH metrics across the life course, as the very high prevalence of intermediate or poor CVH metric levels in U.S. populations makes complete elimination of these two classifications highly unlikely. However, these two knowledge gaps remain unaddressed, likely reflecting a lack of large, contemporary, and longitudinal population-based studies that include the major U.S. race/ethnic groups and cover the entire at-risk age epochs that span adolescence to late adulthood. In this application we respond to these challenges by leveraging extant cross-sectional data and state of the art Markov models to estimate the race/ethnic- and sex-stratified age-specific probabilities of transitioning between ideal, intermediate and poor levels of CVH metrics among approximately 35,000 African American, Hispanic/Latino, American Indian, and Caucasian U.S. participants. The estimated transition probabilities will then be merged with results from published intervention and longitudinal cohort studies to calculate race/ethnic- and sex- specific reductions in CVD associated with interventions that increase the proportion of the U.S. population maintaining ideal CVH. Together, results from this study will inform public health and intervention programs aimed at promoting ideal CVH in the age groups most at risk for its loss. A focus on African American, Hispanic/Latinos, and American Indian populations adds further significance to this innovative proposal, as these populations attain intermediate and poor CVH profiles earlier than European Americans and shoulder elevated burdens of the downstream cardio-, renal- and cerebrovascular disease manifestations.

ABSTRACT Genetic susceptibility underlies a majority of cardiovascular diseases (CVD) and their antecedents, underscored by genome-wide association studies (GWAS) that identified >1,500 loci to-date. Each GWAS-identified locus potentially provides novel mechanistic insight, yet translation of study findings remains largely incomplete, representing a critical barrier to progress. Pleiotropy, a variant that affects multiple phenotypes, is a long- described and pervasive, but largely uncharacterized avenue to advance genomic medicine. Specifically, studies of pleiotropy have the potential to clarify molecular functions, identify mechanistic ?common denominators, inform diagnosis and treatment, and prioritize variants for functional interrogation. Systematic and comprehensive interrogation of pleiotropy is particularly relevant for CVD phenotypes, as decades of human and animal studies support a shared genetic architecture that collectively affects downstream clinical disease. Yet, few studies have comprehensively and systematically evaluated pleiotropy within or across cardiovascular phenotypes or extended investigations to examine how pleiotropic variants affect clinical disease. Further, many CVDs and their antecedents disproportionately affect African Americans (AA) and Hispanic/Latinos (HL). However, the majority (>80%) of participants included in GWAS to-date are of European (EU) ancestry. This research disparity creates a biased view of human variation, fails to leverage the unique genetic architecture of AAs and HLs for fine-mapping, and hinders translation of genetic findings into clinical and public health applications relevant for broad populations. We respond to these gaps by leveraging high-quality, harmonized, and centrally available phenotype and genotype data from the Population Architecture Using Genomics in Epidemiology (PAGE) consortium and the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study (n=100,917; 35% AA; 32% EU; 24% HL) as well as cutting edge statistical methods to comprehensively identify loci with potential evidence of pleiotropy within and across blood pressure, cholesterol, cardiac conduction, glycemic, inflammatory, and obesity cardiovascular domains as well as incident MI and stroke (Aim 1). At known and novel loci with strong evidence of potential pleiotropy, we will leverage population structure, haplotypic architecture, and phenotype correlation through multi-ethnic, multi-phenotype fine-mapping to prioritize variants for further interrogation (Aim 2). Finally, we will leverage longitudinal data and pathway models to disaggregate variants displaying evidence of biological pleiotropy (i.e. variant affects multiple phenotypes due to shared biology) from variants displaying evidence of mediated pleiotropy (e.g. variant influences one phenotype and this phenotype influences a second phenotype) (Aim 3). We hypothesize that CVD phenotypes and clinical disease may be more accurately characterized as variations in clinical expression, with common biological mechanisms. By investigating pleiotropy, we hope to clarify these mechanisms, which has the potential to inform phenotype classification, drug development and repurposing, and CVD prevention.

ABSTRACT Genetic susceptibility underlies a majority of cardiovascular diseases (CVD) and their antecedents, underscored by genome-wide association studies (GWAS) that identified >1,500 loci to-date. Each GWAS-identified locus potentially provides novel mechanistic insight, yet translation of study findings remains largely incomplete, representing a critical barrier to progress. Pleiotropy, a variant that affects multiple phenotypes, is a long- described and pervasive, but largely uncharacterized avenue to advance genomic medicine. Specifically, studies of pleiotropy have the potential to clarify molecular functions, identify mechanistic ?common denominators", inform diagnosis and treatment, and prioritize variants for functional interrogation. Systematic and comprehensive interrogation of pleiotropy is particularly relevant for CVD phenotypes, as decades of human and animal studies support a shared genetic architecture that collectively affects downstream clinical disease. Yet, few studies have comprehensively and systematically evaluated pleiotropy within or across cardiovascular phenotypes or extended investigations to examine how pleiotropic variants affect clinical disease. Further, many CVDs and their antecedents disproportionately affect African Americans (AA) and Hispanic/Latinos (HL). However, the majority (>80%) of participants included in GWAS to-date are of European (EU) ancestry. This research disparity creates a biased view of human variation, fails to leverage the unique genetic architecture of AAs and HLs for fine-mapping, and hinders translation of genetic findings into clinical and public health applications relevant for broad populations. We respond to these gaps by leveraging high-quality, harmonized, and centrally available phenotype and genotype data from the Population Architecture Using Genomics in Epidemiology (PAGE) consortium and the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study (n=100,917; 35% AA; 32% EU; 24% HL) as well as cutting edge statistical methods to comprehensively identify loci with potential evidence of pleiotropy within and across blood pressure, cholesterol, cardiac conduction, glycemic, inflammatory, and obesity cardiovascular domains as well as incident MI and stroke (Aim 1). At known and novel loci with strong evidence of potential pleiotropy, we will leverage population structure, haplotypic architecture, and phenotype correlation through multi-ethnic, multi-phenotype fine-mapping to prioritize variants for further interrogation (Aim 2). Finally, we will leverage longitudinal data and pathway models to disaggregate variants displaying evidence of biological pleiotropy (i.e. variant affects multiple phenotypes due to shared biology) from variants displaying evidence of mediated pleiotropy (e.g. variant influences one phenotype and this phenotype influences a second phenotype) (Aim 3). We hypothesize that CVD phenotypes and clinical disease may be more accurately characterized as variations in clinical expression, with common biological mechanisms. By investigating pleiotropy, we hope to clarify these mechanisms, which has the potential to inform phenotype classification, drug development and repurposing, and CVD prevention.

ABSTRACT Cardiovascular diseases (CVD) remain leading causes of morbidity, mortality, and early disability, and are exacerbated by obesity. It is well known that obesity stresses metabolic pathways, thereby accelerating CVD risk. Yet, the specific biologic mechanisms remain poorly understood. Metabolites are biologically active small- molecule intermediates and byproducts of metabolism that lie along pathways linking genetic susceptibility with CVD and are responsive to obesity, related health behaviors, and CVD risk factors. Thus, metabolites can be powerful disease biomarkers and therapeutic targets and may provide targetable ?mechanistic bridges? linking genome-wide association study (GWAS) findings with CVD risk factors and clinical disease. We hypothesize that: (1) genetic susceptibility influences CVD risk along specific metabolic pathways; (2) that metabolites on these pathways (i) affect and (ii) are affected by CVD risk factors to (3) increase clinical disease risk; and that (4) obesity modifies a subset of metabolite effects. Yet, the majority of metabolomics studies to-date have been largely cross-sectional or clinical efforts in older, European-ancestry populations, with inconsistent control of confounders, including diet, and they have ignored plausible modifiers, including obesity. To address these major research gaps, we will generate longitudinal untargeted and targeted metabolomics profiles in the biracial (47% African American) CARDIA study (n=5,115; 18-30 years in 1985-86; n~3,270 in 2020-21). The CARDIA study is uniquely suited to test the proposed study hypotheses, with 35 years of longitudinal data collected over the key your adult lifecycle period when CVD risk accelerates in concert with increasing obesity. We will develop and employ cutting-edge metabolomics and statistical methods to characterize known and unknown metabolite signals. Longitudinal data, Mendelian randomization, and pathway-based modeling enable assessment of (i) metabolic perturbations that influence CVD and (ii) CVD risk factors that influence metabolic perturbations, (iii) overall and in the context of a growing obesity burden. We address the following specific aims: 1) identify metabolites and major metabolic pathways that influence metabolic CVD risk factors (cholesterol, blood pressure, and glycemic phenotypes); 2) identify metabolic CVD risk factors that influence metabolites and major metabolic pathways; 3) leverage statistical innovations and existing `omics, phenotype, and covariate data for causal inference, to evaluate mechanistic frameworks, and characterize novel metabolites; and 4) test metabolites identified in the CARDIA study for evidence of association with CVD risk factors and clinical endpoints (coronary heart disease, heart failure, and stroke) in the biracial Atherosclerosis Risk in Communities (ARIC) study. We anticipate that the proposed project, prepared by a multi-disciplinary team with expertise in CVD and metabolic epidemiology, nutritional biochemistry, metabolomics, bioinformatics, biostatistics, and genetics, will inform disease mechanisms, with strong potential for identifying biomarkers of CVD risk. Together, our innovations will help identify novel therapeutic and nutritional targets to reduce the global burden of CVD.

ABSTRACT Genetic studies leveraging large-scale genotyping (i.e., ?GWAS data?) are increasingly ubiquitous, as demonstrated by the >56,000 unique single nucleotide polymorphism (SNP)-trait associations identified to-date by genome-wide association studies (GWAS). GWAS data also are being used to understand social-genetic effects, control for genetic predispositions in population health and social science studies, and examine genetic correlation between traits. Despite growing adoption, studies leveraging GWAS data remain largely limited to adult populations of European ancestry and tend to ignore the physical and social environment. Studies with GWAS data combined with rich, longitudinal environmental and phenotype data are therefore needed to permit dynamic, multilevel, integrative research approaches to health that capture bidirectional biological and contextual contributions and their interactions over time. The National Longitudinal Study of Adolescent to Adult Health (Add Health) is an ongoing, nationally representative, multiethnic longitudinal study of the social, behavioral, and biological linkages in health and developmental trajectories from early adolescence into adulthood. As the only nationally representative longitudinal study of young adults that contains multilevel social, behavioral, environmental, and biological data (including recently available GWAS data through dbGaP, a NIH-sanctioned repository), Add Health is well positioned to address these research gaps. However, numerous and persistent challenges prevent broad usage of GWAS data by the research community. Specifically, Add Health users may be ill-prepared for conceptualizing, accessing, storing, understanding, analyzing, and interpreting high dimensional (i.e. >30 million SNPs) GWAS data, an impression supported by our recent survey of users. To enable widespread use of this valuable resource, this application aims to: (1) develop resources to aid users in accessing, understanding, analyzing, and interpreting Add Health GWAS data; and (2) initiate and support a scientific community of Add Health GWAS data users. The proposed application builds upon a 25-year commitment of Add Health investigators to user support and data dissemination, which has resulted in prolific research production with unparalleled disciplinary breadth. We are confident that the proposed resources, which will not be developed without dedicated funding, will expedite access to and facilitate high-quality studies of the Add Health GWAS data by a new group of investigators who may have little-to-no experience with GWAS data. Ultimately, we anticipate that this application will multiply the impact of Add Health sociogenomics research throughout the scientific community and provide a stimulus for new scientific discovery.

ABSTRACT Cardiometabolic risk factors and type 2 diabetes (T2D) impart a substantial and growing morbidity and mortality burden that disproportionally affects racial/ethnic minorities, including Hispanic/Latinos (H/L). The population burden, established disparities, and limited availability of T2D treatments to reverse progression or prevent long- term complications underscore an urgent need to clarify mechanistic pathways that may serve as novel targets for prevention and treatment. Chronic low-grade inflammation is a widely recognized common pathological feature underling cardiometabolic risk factors and T2D, particularly in H/L when compared to other racial/ethnic groups; identifying specific mediators of chronic low-grade inflammation could greatly enhance efforts to tailor existing agents or develop of novel therapies, especially in populations at highest risk. Prior attempts to examine specific mediators of chronic low-grade inflammation have been limited by a focus on downstream markers, including C-reactive protein, which are less likely to be causal or are difficult to reliably measure. Upstream regulation of systemic inflammation is in turn mediated by fatty acid derived lipid mediators termed eicosanoids. Although select eicosanoids have been associated with cardiometabolic risk factors and T2D, prior studies have only assessed a handful of the most abundant eicosanoids in humans. We propose to address this major research gap by leveraging advances in analytical mass spectrometry (MS) that now enable the rapid and accurate quantification of >150 eicosanoids spanning major biosynthetic pathways. Eicosanoids will be assayed in the deeply-phenotyped population-based Hispanic Community Health Study/Study of Latinos (SOL) cohort, enabling cost-effective testing of study hypotheses in a H/L population with established cardiometabolic risk factor and T2D disparities. Specifically, we will identify known and novel eicosanoids associated with cardiometabolic risk factors and T2D, as well as leverage existing genomics data to conduct causal inference studies and evaluate mechanistic frameworks for key eicosanoids. This work will shed insight into the mechanisms underlying cardiometabolic disease in H/L, identify potential sources of health disparities in a genetically admixed cohort, and provide an essential foundation for future studies of inflammatory-modulating therapies aimed at reducing the burden of cardiometabolic disease in the population at large.