Four Interdisciplinary Teams are Awarded $100,000 Seed grants to support projects focused on immunology, imaging, and informatics in precision immunomedicine (iPRIME) for Cardiovascular Disease by the University-funded iPRIME P2PE initiative.
Tick Bite Exposure and Coronary Artery Disease Risk
Loren Erickson, PhD, Department of Microbiology, Immunology, and Cancer Biology, Beirne B. Carter Center for Immunology Research, School of Medicine
Richard Flowers, MD, Department of Dermatology, School of Medicine
Jeff Wilson, MD, PhD, Department of Medicine, Allergy Division, Beirne B. Carter Center for Immunology Research, School of Medicine
Recent human studies provide evidence that humans with IgE sensitization to a novel oligosaccharide allergen present in mammalian meat (alpha-gal) have larger coronary artery plaques and unstable plaque features signifying increased coronary artery disease (CAD) compared to those without IgE to alpha-gal. Notably, there is overlap in the high prevalence of lone star ticks, alpha-gal IgE positivity, and atherosclerosis in the southeast region of the United States. Bites from the lone star tick induce IgE sensitization to alpha-gal. Affected individuals are clinically identified due to the presentation of severe allergic symptoms after ingestion of red meat, dairy, or other alpha-gal-containing foods. However, most individuals with IgE to alpha-gal do not manifest with delayed anaphylaxis and frequently have no outward identifying symptoms. As such, they continue to consume alpha-gal-containing food products (meat, dairy, etc.) which have the capacity to continue to stimulate IgE responses and inflammation in the vessel wall. The novel identification of the first potential allergen eliciting IgE responses linked to atherosclerosis holds promise for leading to strategies to identify and intervene in individuals at risk with allergen avoidance or other immunomodulatory approaches. Thus, there is a critical unmet clinical need to determine if alpha-gal IgE sensitization does indeed promote atherosclerotic plaque formation, discover cellular events that could serve to identify those at risk, and identify potential mechanisms which could lead to biomarker discovery and approaches to limit this silent inflammation. This research project will interrogate human immune cell phenotypes in skin and circulation in CAD patients following a natural tick bite to determine how these phenotypes associate with IgE to alpha-gal. This collaborative team-based project leverages the expertise of Dr. Loren Erickson (Department of Microbiology, Immunology, and Cancer Biology), Dr. Jeffrey Wilson (Department of Medicine, Allergy Division), and Dr. Richard Flowers (Department of Dermatology).
Ensuring Fairness and Robustness of Immunology Data
Aidong Zhang, PhD, Professor of Computer Science, Biomedical Engineering, and Data Science
Ani Manichaikul, PhD, Center for Public Health Genomics and Department of Public Health Sciences
This collaborative seed grant proposal is focused on developing Artificial Intelligence (AI) methods to address fairness and robustness of immune-based data sets, which are two important factors for making AI methods trustworthy, particularly during deployment or use of the methods. Our data sets (termed immunoDB) contain four different types of data: (1) whole genome sequence and/or genome-wide array data, (2) Immune cell data, (3) Imaging data, and (4) Clinical data, which include demographic information of patients, lab tests, clinical notes (text), and other information. To prepare the ImmunoDB data to be AI-ready, this project identifies and addresses potential unfairness through sample and feature de-biasing to improve the representativeness of the studied immunoDB data sets. We develop de-biasing techniques for the AI-based predictive approaches from the immunoDB datasets. This strategy will make us well-powered to study disparities and their effects on Cardiovascular Diseases. This project also addresses the limited coverage and distribution shift of the immunoDB datasets through data augmentation and domain generalization to learn a robust model that can directly generalize to similar but unseen data. We develop data augmentation methods to align the distribution of multiple similar source domains so we can learn a robust model to directly generalize well to similar but unseen domains. We will apply these methods to the ImmunoDB data. The data augmentation methods will not only accelerate the utilization of AI tools on a wide range of related datasets, but also open the possibility of many new applications. Our proposed immuno-based machine learning (ML) methods will offer fair and robust solutions that provide the immunomedicine community with a set of accessible tools for processing data optimized for downstream AI/ML-based analysis, as well as the tools that would enable such analysis to be done by both AI/ML experts as well as researchers without AI/ML expertise.
Impact of sodium-glucose cotransporter 2 inhibitors on systemic inflammation in women with coronary microvascular disease
Patricia Rodriguez-Lozano, MD, Director of Women’s Heart Health Program Department of Medicine, Cardiology
Mete Civelek, PhD, Department of Biomedical Engineering, Center for Public Health Genomics
Heart disease is the leading cause of death among women and men. The scientific community has strived for a “one size fits all” approach to reduce our understanding of atherosclerosis to a single unifying model. However, significant differences between sexes in the underlying pathology of atherosclerosis and its gene regulation exist. Sex-specific mechanisms warrant attention considering the alarming trend of increasing coronary artery disease (CAD) in young women. Ischemic heart disease in women is different from that of men. Women have a unique phenotype with fewer calcified lesions and more non-obstructive disease. Recent studies also identified differences between sexes in coronary endothelial and/or microvascular dysfunction, with women with more prevalence of coronary microvascular disease. Sodium-glucose cotransporter 2 inhibitors (SGLT2i), a drug class approved for the treatment of diabetes, has been shown to reduce atherosclerotic events significantly, hospitalization from heart failure (HF), cardiovascular and total mortality, and progression of chronic kidney disease. New recommendations suggest using SGLT2i in symptomatic patients with chronic HF with reduced ejection fraction to reduce hospitalization and cardiovascular mortality, regardless of type 2 diabetes. In preclinical mouse models of myocardial infarction, SGLT2i improved coronary microvascular function and increased cardiac output. Recent studies show that SGLT2 inhibition has possible mechanisms of benefit that are unlikely to be related to improved glycemic control, such as reduced inflammation. This iPRIME-funded project is an ancillary study to a clinical trial that is currently testing the hypothesis that SGLT2i treatment will improve coronary microvascular dysfunction in women. We will study the impact of SGLT2i treatment on markers of systemic inflammation, specifically inflammasome activity in macrophages in study participants. These studies will significantly advance our understanding of SGLT2 inhibitor’s beneficial mechanisms in microvascular disease in women with non-obstructive CAD.
Integrative single-nucleus and spatial profiling analysis to identify novel biomarkers of chronic inflammatory dilated cardiomyopathy
Clint L. Miller, PhD, Center for Public Health Genomics, Department of Public Health Sciences, School of Medicine
Sula Mazimba, MD, Heart and Vascular Center, Division of Cardiovascular Medicine, School of Medicine
Chronic inflammatory DCM involves complex host interactions with a pathogen (viral, bacterial, or fungal) as well as dysregulated host immune responses in high-risk patients with a genetic predisposition. Dr. Miller recently established protocols to capture chromatin accessibility landscape in coronary artery tissues from explanted and donor hearts and Dr. Mazimba has established a study to procure and analyze heart transplant recipient myocardial and vascular tissues to be integrated with advanced multi-modal cardiac imaging (e.g., echocardiography, cardiac CT, cardiac MRI, PET, etc) and genetic datasets. Finally, collaborator Dr. McNamara has collected single-cell proteomics data from circulating blood and derived PBMC samples in patients with various stages of inflammatory heart disease. By integrating these multi-modal datasets and tissues (immune cells, cardiac and perivascular tissues) there is an exciting opportunity to identify novel cross-tissue immune modulators underlying the initiation and/or progression of inflammatory DCM. We propose the following aims: Aim 1. Perform snATAC profiling of chronic inflammatory DCM and age-matched myocardial tissues to define cell-type specific regulatory mechanisms. Aim 2. Perform spatial transcriptomics profiling of chronic inflammatory DCM and age-matched myocardial heart tissues to identify region and stage-specific disease mechanisms. Aim 3. Perform integrative multi-modal bioinformatics and statistical analyses to identify shared and tissue specific inflammatory biomarkers associated with clinical outcomes related to HFrEF. Using this multi-dimensional, data-driven approach we expect to resolve both tissue resident and infiltrating immune markers of distinct pathological hallmarks of chronic inflammatory DCM. We expect to identify markers that may precede the adverse cardiac remodeling processes and improve the benefit/risk ratio for both follow-up preclinical and translational studies.