Biological and Health Sciences

Investigating Mechanosensitive Signaling Pathways in Epithelial Cell-Cell Junction Remodeling and Barrier Function

Babli Adhikary (Barbour Scholar), Molecular, Cellular, and Developmental Biology

My dissertation research investigates how mechanosensitive signaling maintains epithelial integrity across multiple biological scales. Polarized epithelia experience constant mechanical challenges during development and homeostasis yet must preserve barrier function despite dynamic junction remodeling. My thesis focuses on three interconnected aims: (1) defining how the junctional kinase PAK4, recruited via Afadin, promotes vertex remodeling to resolve multicellular vertices and prevent barrier leaks; (2) characterizing “tricellular zipping,” a newly identified vertex remodeling process in Xenopus epithelium, and determining how pulsatile calcium and actomyosin dynamics drive this process to facilitate higher-order junction reorganization and tissue packing; and (3) elucidating how the mechanosensitive channel TRPV4 controls Rho flares-mediated tight junction remodeling to maintain barrier function in polarized epithelia. Together, these studies will provide a mechanistic framework for epithelial barrier maintenance in development, homeostasis, and disease and lay the foundation for targeting mechanosensitive pathways in barrier disorders.

Sleep, Chronotype, and Quality of Life Among Adults With Bipolar Disorder by Employment Status

Julianne Armijo, Nursing

Bipolar disorder (BD) is a chronic mood disorder that is frequently misdiagnosed as unipolar depression, contributing to delayed and inappropriate treatment and substantial disruptions to functioning and quality of life (QoL). Sleep disturbance is a core and persistent concern in BD, but sleep timing preferences and circadian phase, often operationalized as chronotype, may be equally important for understanding variability in symptoms, daily functioning, and recovery. Employment is a central life role that shapes financial security, identity, and routines, and it can either support or destabilize circadian regularity depending on job demands and scheduling. Paper 2 will provide the dissertation’s quantitative analytic foundation by examining longitudinally how sleep and chronotype relate to QoL in adults with clinically confirmed BD, and by testing whether employment status moderates these relationships using validated measures and established frameworks. Papers 1 and 3 will extend and contextualize these relationships in nursing, a profession characterized by shift work, irregular schedules, and high responsibility that can intensify circadian misalignment. In particular, Paper 3, a mixed-methods study, will emphasize chronotype as a clinically and occupationally relevant factor that may shape nurses’ vulnerability to sleep disruption, work transitions, and changes in QoL over time. Together, the three papers will connect population-level longitudinal patterns with profession-specific evidence and lived experience to inform targeted clinical and workplace interventions for nurses living with BD.

Paper 1 will be a scoping review exploring the unique psychological, professional, and interpersonal experiences of nurses living with BD. Guided by Joanna Briggs Institute methods, systematic searches will be conducted in PubMed, CINAHL, PsycINFO, Embase, Scopus, and Web of Science. Articles will be screened in Covidence, with data extraction and management in Covidence and NVivo 14. Review processes will be documented using a PRISMA diagram; approximately 2,887 records are expected to be screened. This review will identify themes relevant to wellness, including workplace factors, stigma, coping strategies, and relationships, and will generate recommendations for modifying personal and professional environments to better support nurses with BD. Findings will also highlight gaps and the need for diverse perspectives and robust qualitative and quantitative approaches.

Paper 2 will be a quantitative longitudinal study testing whether employment status moderates the association between sleep, chronotype, and QoL in adults with BD, drawing on concepts from Interpersonal and Social Rhythm Therapy and U.S. Bureau of Labor Statistics (BLS) employment definitions. Using data from the Heinz C. Prechter Longitudinal Study of Bipolar Disorder, the analytic sample will include adults with clinically confirmed BD diagnosed via structured interview. Validated self-report measures will assess sleep (Pittsburgh Sleep Quality Index, Epworth Sleepiness Scale), chronotype (Munich Chronotype Questionnaire), QoL (36-Item and 12-Item Short Form Surveys), and employment status (Life Functioning Questionnaire categorized using BLS definitions). Generalized estimating equations, mixed-effects models, and sensitivity analyses will address within-person change, missingness, comorbidities, and subgroup patterns. Results will be presented with effect sizes, 95 percent confidence intervals, and statistical significance set at p < .05.

Paper 3 uses a mixed-methods design to extend these questions specifically to self-identified nursing professionals within the Prechter cohort (n = 16), with an emphasis on how chronotype and circadian disruption intersect with work schedules and employment transitions. Quantitative measures of sleep, chronotype, QoL, and employment status will be complemented by semi-structured LIFE interviews to capture mood episodes and work transitions through event-sequenced qualitative analysis. Triangulation will clarify how chronotype-linked sleep patterns and occupational change co-occur with shifts in QoL among nurses with BD.

Collectively, these three papers aim to inform clinically and occupationally relevant interventions, such as circadian-stabilizing routines, chronotype-informed scheduling and shift accommodations, sleep hygiene supports, flexible work arrangements, and stigma reduction, to promote sustained wellness, retention, and safe, high-quality healthcare delivery.

Integrative Approaches to Identify Cell-Type-Specific Gene-Regulatory Interactions Underlying Disease Mechanisms

Elysia Chou, Bioinformatics

Heritable, complex diseases often involve gene dysregulation stemming from variation in non-protein-coding DNA, which makes up approximately 98.5 percent of our genome, highlighting the need to investigate how DNA elements regulate genes and how variation in these elements impacts their function. Mapping these regulatory elements to their target genes is the first step to uncovering the molecular mechanisms that underlie diseases. Such efforts guide experimental validation and inform new diagnostic and therapeutic approaches. Here, we develop two integrative methods to study element-gene interactions. Our first method links disease-associated genetic variants to their target genes using a novel data fusion approach. Our second method improves upon existing element-gene linking models by incorporating both single-cell multimodal data and information about relevant DNA-binding proteins. We apply our models to study psoriasis and schizophrenia and pinpoint hundreds of disease-specific gene-regulatory interactions among hundreds of thousands of possibilities, all within a cell-type-specific context.

The Effects of Glucocorticoid Receptor Knockdown in a Corticostriatal Pathway on Inhibitory Control and Stress Responsivity in Rodents

Princess Felix, Neuroscience

Environmental cues guide adaptive behavior but can also elicit maladaptive responses, as observed in substance use disorder (SUD), where drug-associated cues trigger craving and relapse. This vulnerability arises when cues acquire excessive incentive salience, a process that varies across individuals. Individuals prone to attributing incentive salience to reward cues are also vulnerable to psychiatric disorders characterized by a lack of inhibitory control over behavior, or impulsivity. However, the neural mechanisms that underlie inhibitory control remain unclear. Here we show that glucocorticoid receptors (GRs) within a prelimbic cortex-dorsal striatum pathway mediate incentive salience attribution. Pathway specific GR knockdown in transgenic rats increased cue-directed behavior and elevated corticosterone following cue-reward learning without altering stress-induced corticosterone. These findings identify a GR-dependent cortico-striatal mechanism driving cue-motivated behavior independently of systemic stress responsivity, offering new insights into the neurobiology of impulsivity and potential strategies for novel therapeutics targeting SUD and related psychiatric conditions.

Visible Stewardship as Urban Sustainability Signal: Large Vision-Language Models Reveal Spatial and Temporal Neighborhood Dynamics

Jianxing Guan (Barbour Scholar), Environment and Sustainability

Urban landscapes are dynamic socio-environmental systems where change emerges first through micro-scale, visible transformations reflecting human stewardship, evolving land patterns, and legacy effects. In pursuit of urban sustainability amid environmental and social challenges, understanding how micro-scale landscape transformations diffuse and accumulate to shape human perception, maintenance, and urban dynamics across space and time has become increasingly critical. In this dissertation, I employ large vision-language models applied to multi-year street-level imagery to detect visible signs of stewardship—micro-scale landscape elements that connote human intention and care. By innovatively integrating AI-derived visual indicators with household survey data, I examine how visible stewardship influences residents’ maintenance behaviors within and beyond their home landscapes and how these changes shape neighborhood vacancy and occupancy patterns over time. My results will bridge micro-scale landscape changes and broader urban dynamics, advancing both theoretical and computational understanding of human-environment interactions and informing pathways toward more sustainable urban futures.

Norovirus Alters Host Metabolism for Efficient Replication

Adam Hafner , Microbiology and Immunology

Human noroviruses (HNoVs) are the leading cause of acute non-bacterial gastroenteritis worldwide. Despite the devastating public health impact of HNoV infections, neither vaccines nor therapeutics exist, underscoring the need for further investigations to better understand NoV biology and develop intervention strategies. Viruses commonly hijack host metabolic pathways to create more favorable intracellular environments to ensure successful propagation. However, the extent to which NoVs can reprogram host metabolism is unclear. This project demonstrates that NoV infection significantly upregulates glycolysis and glutaminolysis, thereby commandeering host resources to promote viral replication. Additionally, we have discovered that multiple metabolic pathways are essential for successful NoV reproduction, and that NoV non-structural proteins can directly alter the activity of metabolic enzymes. These findings provide important insights into the mechanisms NoVs utilize to successfully replicate, deepening our understanding of virus-host interactions and potentially uncovering novel therapeutic targets.

Subcellular Compartmentalization Regulates Timing and Selectivity of the Cellular Stress Response

Noah Helton, Genetics and Genomics

Cells exposed to environmental stress must rapidly reprogram gene expression to ensure survival. While transcriptional and translational responses are well-documented, the spatial regulation of mRNAs during stress remains poorly understood. This dissertation investigates how subcellular compartmentalization regulates the timing and selectivity of gene expression during stress and recovery. First, using single-molecule imaging and chemical genetics, I demonstrate that ribosome association on upstream open reading frames (uORFs) prevents key stress-induced gene mRNAs (e.g., ATF4, GADD34) from being sequestered into repressive cytoplasmic stress granules. Second, I reveal that nuclear export of stress-induced transcripts is regulated, where a subset of mRNAs (e.g., HSPA1A) are retained in the nucleus during acute stress and exported during the recovery phase after stress. Collectively, these findings reveal that subcellular compartmentalization of mRNAs is a fundamental regulatory layer of the central dogma, ensuring a precise and coordinated response to stress.

Leveraging Multi-Omics to Characterize the Human β Cell Cytokine Response

Belle Henry, Immunology

Type 1 diabetes (T1D) is an autoimmune disease characterized by the destruction of insulin-producing pancreatic β cells. T1D is caused by a combination of β cell fragility and immune system dysregulation leading to β cell demise. There is an urgent need for disease-modifying therapies to enhance β cell resiliency in the face of inflammatory stress in T1D. Published approaches to understanding β cell inflammatory stress have relied on transcriptomics, and there is a lack of corresponding proteome and phosphoproteome data. Thus, we exposed human β cells to pro-inflammatory cytokines and quantified changes in the transcriptome, proteome, and phosphoproteome. Analysis of the phosphoproteome revealed cytokine-induced alterations in kinase activity that were not reflected in changes in mRNA or protein abundance. We demonstrated that inhibition of one of those kinases, LRRK2, protected β cells against inflammatory toxicity. Further work will utilize this dataset to identify additional kinases driving β cell cytokine responses.

Using Mixed Methods to Develop an Intervention that Optimizes Oral Anti-Cancer Agent Self-Management among Arab Americans

Mohamad Ismail (Roblin Fellow), Clinical Pharmacy Translational Science

Oral anti-cancer agents (OAAs) shift cancer treatment from clinic to home, requiring patients to manage medication use and symptoms independently. This study aims to identify and address factors influencing OAA self-management, particularly among Arab Americans. Using a mixed methods approach, we will first conduct semi-structured interviews and photovoice to explore patient experiences and identify key barriers and facilitators. Findings will inform the development of a cross-sectional survey to quantify these factors. Results from both phases will be integrated to inform the co-design of culturally tailored intervention components. Finally, we will engage key stakeholders, including patients, providers, and community partners, to assess the acceptability, feasibility, and appropriateness of the proposed intervention. This study will generate actionable insights to support the development of patient-centered strategies that improve medication self-management and reduce disparities in cancer care.

Maternal Glyphosate Exposure in Relation to Endogenous Hormones, Lactation Duration, and Early Childhood Neurodevelopment

Haley Jenkins, Environmental Health Sciences

Glyphosate-based herbicides are widely used, resulting in human exposure to glyphosate and its degradant, aminomethylphosphonic acid. These compounds may disrupt endocrine function and affect maternal breast and fetal development, yet few epidemiologic studies have examined these effects or their interaction with neurotoxic metals/metalloids. This study evaluates associations between prenatal glyphosate concentrations and both early neurodevelopment and lactation duration, including mediation by prenatal hormones and co-exposure with metals/metalloids. Participants from the PROTECT and CRECE/ECHO cohorts in Puerto Rico provided prenatal urine and blood samples for exposure measurement; neurodevelopment was assessed using the Battelle Developmental Inventory, second edition, from six weeks to 48 months. Linear mixed-effects, mediation, and mixture models (Bayesian Kernel Machine Regression, quantile g-computation) will assess relationships. Higher prenatal glyphosate levels are expected to predict lower neurodevelopmental scores and shorter lactation duration via hormonal disruption.

Roles of Phosphoinositide Signaling in Cell Migration

Huseyin Karaburk, Molecular, Cellular, and Developmental Biology

Cell migration is critical for several biological processes, while dysregulation of cell migration is common in cancer. However, there are gaps in knowledge of how cell migration is controlled. Here I demonstrate that cell migration and actin dynamics are regulated by PIKfyve, a phosphoinositide lipid kinase. I show that PIKfyve regulates RAC1, a regulator of actin dynamics via the negative regulation of a RAC1 inhibitor. Using proximity labelling, I discovered that the RAC1 inhibitor, ARHGAP17 resides near PIKfyve. Genetic depletion of ARHGAP17 partially rescues PIKfyve inhibition, which provides key evidence that PIKfyve promotes RAC1 activation via inhibition of ARHGAP17. Moreover, using TIRF live-cell imaging, I show that ARHGAP17 and PIKfyve colocalize on the same endosomes. These findings expand knowledge of how PIKfyve regulates cell migration. PIKfyve inhibitors are in clinical trials for multiple types of cancer. Thus, these studies may contribute to knowledge of how PIKfyve inhibition impairs cancer cells.

Characterizing the Correlates and Outcomes of Problematic Substance Use from Early Adolescence to Early Midlife

Lara Khalifeh, Clinical Pharmacy Translational Sciences

Hazardous alcohol and other substance use is a leading contributor to morbidity and impairment. However, important gaps remain in understanding the heterogeneity of substance use across the lifespan, particularly during midlife, a salient but understudied period for substance use research. Using data from a nationally-representative longitudinal cohort, this dissertation will examine the lifespan correlates and midlife outcomes of substance use, across two studies. In Study 1, I will test whether early-adolescent substance use predicts adverse midlife outcomes across mental, physical, social, and socioeconomic domains, employing propensity-score weighting to strengthen causal inference. In Study 2, I will characterize individuals who initiate high-frequency binge drinking at midlife, through an evaluation of early-life and adult correlates. Together, these studies aim to clarify the enduring impact of adolescent substance use and the developmental epidemiology of hazardous midlife alcohol consumption, with implications for identifying at-risk individuals and informing prevention efforts across the lifespan.

Investigating the Brakes, Inhibitors, and Accelerants of Bacterial Amyloid Formation

Divya Kolli , Molecular, Cellular, and Developmental Biology

Amyloid proteins are typically studied in the context of human misfolding diseases such as Parkinson’s disease and Alzheimer’s disease. Recent research has shown that amyloid proteins are also utilized by all three domains of life to serve important biological functions. Amyloid formation is inherently toxic; therefore, functional amyloids are highly regulated and serve as model systems to study and understand how to control amyloid formation. This project utilizes mutagenesis, aggregation assays, predictive modeling, and microscopy to investigate three distinct mechanisms that dictate amyloid formation and kinetics. This work demonstrates that bacterial amyloid CsgA utilizes electrostatic interactions to self-regulate and slow amyloid formation and that a dedicated chaperone protein, CsgC, can display two distinct mechanisms of inhibiting CsgA amyloid formation. These findings enable E. coli to bypass the inherent toxicity of utilizing functional amyloid CsgA. Lastly, we aim to investigate how bacterial amyloids can accelerate human amyloids implicated in disease.

Understanding Mechanisms of Respiratory Pathogen Virulence and Transmission

Katy Krupinsky, Microbiology and Immunology

Respiratory pathogens impose significant burdens on society. Available methods provide inadequate data to mitigate their impacts. One alternative approach is mathematical modeling. Specifically, I focus on lymph node (LN) Mycobacterium tuberculosis (Mtb) infection, impact of viral and host factors on influenza A virus (IAV) transmission bottleneck size, and drivers of SARS-CoV-2 transmission bottlenecks. I developed a mathematical model of LN Mtb and performed sensitivity analysis on this model. I determined transmission bottleneck size for IAV using two established maximum likelihood estimation models and samples from a multi-season household study. I will explore drivers of SARS-CoV-2 transmission bottlenecks through creation of an agent based model. My Mtb model captures experimental data and predicts potential drivers of LN fate. I also found IAV transmission bottlenecks are tight and not altered by viral or host factors. The SARS-CoV-2 project is ongoing. These models provide new tools for understanding infections of public health importance.

Evolvability on the Fitness Landscape

Bhaskar Kumawat, Ecology and Evolutionary Biology

My research examines the role of the fitness landscape in shaping evolvability—the capacity of organisms to undergo adaptive evolution. In Chapter 1 of my dissertation, I show that evolution in a changing environment can steer populations to regions of the fitness landscape where mutations that are beneficial in future environments arise frequently, increasing evolvability. In Chapter 2, I derive and solve a new partial differential equation that describes evolution on time-varying fitness landscapes; paired with a metric based on landscape structure, these results enable us to predict rare, evolvability-enhancing mutations that emerge in changing environments. Finally, in Chapter 3, I perform directed evolution experiments with bacteriophages (viruses that infect bacteria) to demonstrate a practical route to generate viral populations with heightened evolvability. Together, this work demonstrates that evolvability is evolvable, develops a framework to predict its emergence, and applies this theory to laboratory experimental evolution of bacteriophages.

Proteolysis-Targeting mRNA Vaccines Enhance Antitumor Immunity

Xiang Liu, Pharmaceutical Sciences

mRNA cancer vaccines offer a versatile platform for inducing tumor-specific T cell responses, but their clinical efficacy has been limited by inefficient antigen processing. Inspired by proteolysis-targeting chimera (PROTAC) technology, I developed a proteolysis-targeting mRNA vaccine (PROTRAX) that enhances antigen processing by recruiting endogenous E3 ligases to mRNA-encoded antigens. Through systematic engineering and in vivo immunological analysis, I demonstrated that PROTRAX drives efficient antigen processing, increases MHC class I antigen presentation, and amplifies antigen-specific CD8⁺ T cell responses while selectively enhancing responses to subdominant epitopes. The balanced CD8⁺ T cell repertoire improves antitumor efficacy in murine tumor models. PROTRAX also induces a stem-like CD8⁺ T cell population with elevated PD-1 and stem cell antigen-1 expression, a phenotype associated with sustained T cell proliferation. Together, PROTRAX provides a generalizable strategy that shifts mRNA vaccine design from antigen expression toward active control of antigen processing to improve immune quality and durability.

Transforming Climate Change Education: A Theoretical and Practical Framework for U.S. Schools

Timothy Muhich, Environment and Sustainability

Climate change is among the most daunting scientific and societal challenges of our time. Young Americans are increasingly worried about climate change, meanwhile, U.S. schools largely ignore the issue. Despite recent development of guidelines and materials for teachers, climate change education (CCE) needs substantial improvement to meet student and United Nations recognised needs. Therefore, my dissertation builds a framework for translating our knowledge of effective CCE practice into classroom use. To do this, I use phenomenological methods and the conceptual framework of expert teaching to study exemplary CCE and create theoretical models of instruction for a societally challenging topic across disciplines. I then apply cultural-historical activity theory to guide collaborative professional development with these models across three urban, suburban, and rural Michigan schools. Ultimately, I aim to shift scholarship around CCE implementation to leverage collective teacher agency and expansive learning, to provide actionable strategies for scaling transformative CCE nationwide.

Development of Inhalable Lipid Nanoparticles for the Treatment of Lung Cystic Fibrosis: Overcoming Barriers to CFTR Gene Delivery

Youngseo Na, Medicinal Chemistry

Cystic fibrosis (CF) is a life-threatening autosomal recessive genetic disorder caused by CFTR gene mutations, resulting in dysfunctional CFTR protein. This dysfunction alters the structure and function of submucosal glands and impairs mucociliary clearance mechanism, leading to progressive pulmonary disease. Although advances in small-molecule modulators have improved symptom management, approximately six to 10 percent of CF patients still lack treatment options.

Gene addition strategy enabling stable CFTR expression represents a promising approach to address this unmet need. However, efficient delivery to target airway cell types, including lung epithelial cells and basal stem cells, remains a bottleneck due to viscous mucus, immune responses, and epithelial tight junctions. Our research focuses on developing inhalable lipid nanoparticles (LNPs) that promote localized protein production in the lungs while minimizing systemic toxicity. We have recently identified an LNP formulation with pronounced lung-tropism, which has the potential to overcome current limitations and improve clinical outcomes for all CF patients.

Investigating the Role and Therapeutic Potential of Gut Microbe-Derived Metabolites in Colorectal Cancer

Jennifer Nguyen, Medicinal Chemistry

Colorectal cancer (CRC) is a major global health crisis and the second leading cause of cancer-related deaths. While it is well-established that the gut microbiome – the trillions of bacteria living in our digestive tract – plays a critical role in this disease, challenges remain in identifying which specific microbes are helpful or harmful, and how they exert their influences in CRC. My research uses a multi-omics approach, combining advanced computing and machine learning to map the chemical signals produced by gut bacteria. By studying patients at different stages of the disease, we can identify the molecules that either promote tumors or protect against them. Our work has already found unique chemical signatures in early-stage patients and discovered how certain bacteria interact to produce bioactive compounds. Ultimately, this work will lead to non-invasive diagnostic tests for early CRC detection and new microbiome-centered treatments to improve patient survival.

Federated Statistical Inference for Heterogenous Time-to-Event Data

Malcolm Risk, Biostatistics

In the past, advances in clinical research have relied on aggregating a sufficient amount of data in a single location for analysis. This requirement limits impactful multi-center research because privacy concerns constrain the ability to freely share data. To address this constraint, this dissertation develops several federated statistical inference methods for fitting statistical models across multiple institutions without sharing individual-level data. In the first project, we develop a novel federated algorithm for estimating inverse propensity weighted KM (IPW-KM) curves, with an application to adverse event monitoring following COVID-19 vaccination. In the second project, we extend the aforementioned algorithm to a competing risks setting. In the third and final project, we develop a novel method for analysing heterogenous data in a federated context by relaxing the assumption of identical data distribution across sites. Taken together, these methods provide researchers with a wider toolkit for conducting multi-center research without compromising patient privacy.

Defining Novel Pathway Dependencies in Cancer Cells Using CRISPR/Cas9 Screening

Nicholas Rossiter, Molecular, Cellular, and Developmental Biology

Understanding the interdependency of cellular processes remains a major goal of cell biology, particularly for the opportunity to expand and refine the catalog of actionable targets in cancer and other diseases. The advent of CRISPR/Cas9 genetic screening has enabled the identification of genes underlying a specific phenotype of interest with unprecedented efficiency. This work leverages CRISPR-based screens to uncover two distinct and novel functional relationships in human cells, motivated by known alterations to zinc homeostasis and glucose metabolism in cancer. Utilizing large-scale screening datasets, we identify and validate a previously unknown role for zinc export in maintaining RNA turnover in the mitochondrion of pancreatic cancer cells. Separately, we employ a genome-wide CRISPR screen to discover that the Fanconi anemia DNA repair pathway protects against the toxic glycolytic metabolite methylglyoxal in human blood cancer cells. Together, these projects define new mechanistic connections between multiple cancer-associated pathways in human cells.

Cytoplasmic Localization of Pseudouridine Synthase 7 Facilitates a Pseudouridine-Dependent Enhancement of Cellular Stress Tolerance

Minli Ruan, Biological Chemistry

Pseudouridine (ψ) is an abundant post-transcriptional modification found across all classes of RNA. It is widely speculated that ψ inclusion in messenger RNAs (mRNAs) might provide an avenue for cells to control gene expression post-transcriptionally. Here we demonstrate that one of the principal mRNA pseudouridylating enzymes, pseudouridine synthase 7 (PUS7), exhibits a stress-induced accumulation in the cytoplasm of yeast and human epithelial lung cells. Stress-induced and cytoplasmic localization of PUS7 promote ψ-incorporation into hundreds of mRNA targets. In contrast, the modification status of tRNA sites targeted by PUS7 (ψ13 and ψ35) is unperturbed. Furthermore, engineered PUS7 cytoplasmic localization increases cellular fitness under reactive oxygen species and divalent metal ion stress. Quantitative proteomics reveal a reshaping of the proteome upon PUS7 relocalization under stress. Collectively, our data demonstrate that PUS7 localization alters mRNA pseudouridylation patterns, reshapes the proteome, and enhances cellular fitness.

Exploration of Endoplasmic Reticulum-Associated Degradation of Membrane Proteins

Jennifer Russ, Biological Chemistry

In eukaryotes, most newly synthesized membrane proteins are folded and integrated into the membrane at the endoplasmic reticulum (ER). Accordingly, the ER has evolved robust protein quality control processes to detect and remove misfolded proteins, which otherwise contribute to proteotoxicity. One conserved process is ER-associated degradation (ERAD). During ERAD, misfolded membrane proteins are recognized, ubiquitinated, and extracted from the ER membrane for proteasomal degradation. How ERAD achieves substrate specificity is unknown. Additionally, during extraction, substrates must be physically dislocated from the phospholipid bilayer. The cellular machinery required to exert sufficient force for substrate extraction remain undefined. In this dissertation, we establish reconstituted assays using purified components in combination with genetic and in vivo approaches to elucidate how the ERAD machinery facilitates the selective recognition and extraction of diverse integral membrane substrates. This work explicates fundamental ERAD principles and provides generalizable platforms for future studies of complex membrane-associated systems.

Leveraging Metabolomics to Assess Patterns of Phytochemical Defense Across Space, Time, and Trophic Levels

Abrianna Soule, Ecology and Evolutionary Biology

Plants and insects constitute the greatest macroscopic diversity on Earth, making them powerful systems for exploring species interactions. Among the most studied plant-insect interactions is herbivory. To defend against herbivores, plants employ a suite of defenses, including phytochemicals that either directly poison herbivorous insects or recruit beneficial insects that aid in herbivore removal. Several hypotheses have been proposed to understand and explain variation in phytochemical defense investment both within and between species, including coevolution with natural enemies (i.e., insect herbivores), competition and niche theory, resource availability, and increased selection with proximity to the equator. In my dissertation, I combine field sampling, meta-analysis, metabolomics methods, and laboratory experiments to explore several of these hypotheses across multiple spatial, temporal, taxonomic, and trophic scales. My research expands our understanding of phytochemical defense investment and the causes and consequences of variation in this key ecological phenotype.

Evaluating the COVID-19 Vaccine and Infection-Induced Antibody Responses

Jade Yangyupei Yang, Epidemiological Science

My dissertation investigates the influence of vaccination and prior infection on antibody immunological protection against SARS-CoV-2 variants during the Omicron period (2022 until the present). By linking population-level immunologic responses with longitudinal epidemiologic data, this research provides integrated evidence on antibody protection, population immunity dynamics, and vaccine effectiveness. The findings carry broad implications for public health preparedness and immunization policy, particularly at a time of rising vaccine hesitancy, shrinking vaccine research investment, and uneven access to COVID-19 vaccination. Three aims are: 1) to identify antibody correlates of protection against SARS-CoV-2 Omicron infection; 2) to characterize antibody kinetics and antigenic landscape across multiple vaccination and infection exposures; 3) to compare conventional and causal inference approaches to evaluate COVID-19 vaccine effectiveness during the Omicron period.

Fibroblast-Epithelial Crosstalk Drives Cell Death in Inflammatory Bowel Disease via ACSL4-Mediated Lipid Reprogramming

Yuezhong Zhang (Barbour Scholar), Molecular and Integrative Physiology

Inflammatory bowel disease (IBD) is a chronic, relapsing inflammatory disorder affecting millions of Americans, yet its etiology remains poorly understood. A hallmark of IBD is epithelial cell death, but the mechanism driving this injury was unknown. In our research, we discovered that intestinal epithelial cells die through ferroptosis, an iron-dependent, metabolically regulated form of cell death. We found that iron in IBD tissues transcriptionally induces Acyl-CoA synthetase long chain family member 4 (ACSL4), a key enzyme that drives lipid peroxidation, in intestinal fibroblasts. Using pharmacological and genetic approaches, we demonstrate a critical function of fibroblast ACSL4 across multiple murine colitis models. These fibroblasts reprogram lipid metabolism and sensitize neighboring epithelial cells to ferroptosis. This process is further potentiated by hypoxia and impaired antioxidant responses, revealing a novel stromal-epithelial crosstalk mechanism. These findings highlight the critical intersection of metabolism, cell death, and tissue injury in IBD and identifies potential therapeutic targets.

Supporting Personalized Medication Treatment Decisions from Real-World Evidence with Causal Machine Learning

Yifan Zheng, Clinical Pharmacy Translational Science

Medication exposure, defined by the timing and duration of medication use, is a key determinant of treatment effectiveness, yet is inadequately captured in real-world studies. This dissertation develops MedExposure, a longitudinal framework that combines electronic health records with pharmacy fill and medication pick-up data to reflect medication-taking behavior. Using hypertension as a case study, the work shows that common adherence measures can differ significantly depending on data source and calculation method, resulting in meaningful misclassification of patients’ treatment evaluation. By accounting for delays, gaps, switching, dose changes, and add-on therapies, MedExposure captures medication use more realistically over time. The dissertation then evaluates whether improved exposure measurement strengthens the prediction of cardiovascular outcomes. Finally, causal modeling methods are applied to estimate individual treatment responses and inform medication switching decisions. Together, this work establishes medication exposure as a foundational construct for real-world evidence and advances personalized treatment decision-making.