Biological and Health Sciences

Structural and Functional Analyses of Oligomeric Protein Complexes that Interact with Membranes
Sarah Connolly, Cellular and Molecular Biology

Cellular membranes are essential for life and mediate cellular entry and egress. Membrane proteins regulate traffic across the lipid barrier and transmit signals across membranes. My research delineates how Vacuolating toxin A (VacA) and human Caveolin-1 form oligomers that insert into and modify membranes. VacA is secreted by the pathogenic bacteria Helicobacter pylori and is associated with gastric cancer. I determined the structure of VacA’s intermediate pore state and found evidence for a multi-step process of pore formation using single-particle cryo-electron microscopy (cryo-EM), and cryo-electron tomography (cryo-ET) analyses. Human Caveolin-1 oligomers are essential building blocks of caveolae, which are important for regulating membrane tension and whose dysregulation leads to cardiopathies, lipodystrophies, and cancers. Leveraging phylogenetic analysis, cryo-EM, and cryo-ET, I am defining human Caveolin-1 and evolutionarily distinct Caveolin-like proteins’ oligomerization, and their interaction and alteration of membranes. This work provides molecular models of two membrane proteins interacting with cellular membranes.

Testing Hidden Roles and Physiological Tradeoffs in the Evolution of Squamate Coloration
Hayley Crowell, Ecology and Evolutionary Biology

Most organismal traits serve more than one function. These functions are often in conflict, resulting in “tradeoffs” that underlie the evolution of many phenotypes in nature. I study tradeoffs in animal coloration, in which complex colors and patterns simultaneously function for predator avoidance and thermoregulation. Using snakes and lizards, I combined phylogenetic comparative analyses, field observations, and experimental manipulations to assess ecological and physiological tradeoffs structuring the evolution of adaptive color. First, I tested among hypotheses driving the evolution of ultraviolet coloration across the snake tree of life – a group of animals where it has not been previously documented. Within one species, I then tested how color morphs affect responses to physiologically-stressful environments. Finally, I used experimental gradients to test the role of color variation in thermoregulatory behavior in lizard color morphs. Together, these results support tradeoffs as foundational to the maintenance of color diversity across the natural world.

Unveiling the Complexity of Pancreatic Ductal Adenocarcinoma: A Comprehensive Study of Tumor Progression and Microenvironment
Ahmed Elhossiny, Bioinformatics

Pancreatic Ductal Adenocarcinoma (PDAC), a deadly disease, includes epithelial (tumor) and stromal (non-tumor) compartments. Through a partnership with Gift of Life – Michigan, we obtained donor pancreata and characterized them using single-cell RNA sequencing, spatial RNA sequencing, and immunostaining. We discovered that most healthy individuals, at any age, present with precursor lesions. As the incidence of pancreatic cancer is relatively low, the question arises as to what factor(s) prevent or promote progression to malignancy. To address this, we evaluate the mutation burden by whole exome sequencing and perform spatial transcriptomics to compare the microenvironment surrounding precursor lesions or tumors. Differentially expressed genes are studied in an experimental model using organoids and co-cultures fibroblasts, a key stromal component, to identify key tumor-promoting factors. The long-term goal of this research is to identify druggable targets, paving the way for the development of innovative therapies for pancreatic cancer patients.

Regulation of Eating Behavior by the Melanocortin-3 Receptor (MC3R) in AgRP Neurons
Yijun Gui, Molecular, Cellular and Developmental Biology

Eating provides nutrients that are essential for energy homeostasis maintenance. Adaptive modifications of eating behavior in response to energy deficiency are critical for survival. The circuitry of hypothalamic AgRP neurons is the neuronal basis for the sensing of hunger, the activation of which is both necessary and sufficient to drive food intake. However, molecular mechanisms underlying the regulation of AgRP neuron activation are incompletely understood. In my dissertation, I investigate the role of the melanocortin-3 receptor (MC3R), a conserved membrane protein in the central nervous system from mouse to human, in regulating AgRP neuron activation. Utilizing animal genetics, pharmacology, immunohistochemistry, and calcium imaging, I aim to determine if MC3R expression is required for the normal activation of AgRP neurons in response to energy deficiency, and to identify the relevant molecular mechanisms. The completion of my dissertation will greatly enhance our understanding of the neuro-molecular basis for eating behavior, spotlighting MC3R as a prospective therapeutic target for anorexia and weight-loss treatments.

The Implementation of Combination Therapies for the Treatment of Triple Negative Breast Cancer
Kassidy Jungles, Pharmacology

Triple negative breast cancer (TNBC) is an aggressive breast cancer with few treatments. Radiotherapy is a mainstay therapy for treating breast cancer; however, the return of one’s cancer following radiotherapy is common. Consequently, developing novel therapeutic approaches to improve the effectiveness of radiotherapy is clinically warranted. My thesis work examines the implementation of novel combination therapies, employing currently available therapeutic tools (i.e., radiation therapy, small molecule inhibitors) to improve outcomes in TNBC. Specifically, in my research, I have examined the spindle assembly checkpoint (SAC) complex, a crucial regulator of cancer cell growth and survival. Proteins of the SAC complex (monopolar spindle kinase I and Aurora kinase B) have been found to be upregulated in breast cancer patients and correlate with poor prognosis. My thesis examines the effects of combined SAC complex inhibition and radiotherapy in cellular and animal models of TNBC to improve treatment options and outcomes for TNBC patients.

Determining the Effect of Psilocybin on Chronic Pain and Neural Dynamics in Rats
Nicholas Kolbman, Pharmacology

Psilocybin is being increasingly explored for its therapeutic potential to treat psychiatric disorders ranging from depression to addiction, but systematic studies determining the effect of psilocybin on chronic pain and neural dynamics are lacking. To address these gaps in knowledge, I investigated the effect of intravenous psilocybin administration on mechanical hypersensitivity and thermal hyperalgesia in a rat model of chronic pain (aim 1), and neural dynamics in healthy rats and rats with chronic pain (aim 2). Currently, I am investigating (aim 3) the role of the serotonin 2A receptor – known to mediate the psychedelic effects of psilocybin – in attenuation of indices of chronic pain and changes in neural dynamics, as shown in aim 1-2. These studies are expected to provide mechanistic evidence to motivate clinical exploration of psilocybin to treat chronic pain, and insights into psilocybin’s effect on neural dynamics, which is posited to underlie its therapeutic effects.

Traf6 Regulates β-cell Stress Responses through Mitophagy
Elena Levi-D’Ancona, Immunology

Impaired insulin release from β-cells drives diabetes and is linked to dysfunctional mitochondrial function. In immune cells, mitochondria act as junctions for innate immune proteins, including toll-like receptor (TLR) signaling proteins, whose activation enhances mitochondrial damage. Mitochondrial recycling, known as mitophagy, serves as a quality control mechanism, selectively degrading damaged mitochondria. However, the relationship between innate immunity and mitophagy is unknown. This study reveals that Traf6, a regulator of TLR-dependent signaling, mediates blood glucose levels following high-fat diet (HFD). β-cell specific Traf6 knockout mice (Traf6Δβ) exhibited elevated blood glucose levels and developed reduced insulin secretion following HFD. Co-deletion of Parkin, a protein that initiate mitophagy, rescued these defects, highlighting a novel link between Traf6 and β-cell mitochondrial quality control. Islets from HFD-fed Traf6Δβ mice also exhibited increased mitochondrial mass, abnormal mitochondrial structure, and decreased mitophagy. Thus, Traf6 links β-cell responses to metabolic stress with mitophagy regulators, potentially unveiling connections between innate immune signals and mitochondrial health.

Statistical Methods for High-Dimensional Genetics and Genomics Data
Zheng Li, Biostatistics

Advancements in genetics research, and more importantly, in statistical methods that power the analysis of large-scale genetics and genomics data have made it possible to provide personalized treatments. Despite a huge success in identifying genetic markers associated with various diseases and complex traits, understanding the molecular mechanisms underlying these associations remain challenging. Through methodological developments that leverage gene expression studies, we have the potential to fill this critical gap. In my dissertation, Chapter I focuses on a data integrative framework to identify genes associated with a disease/trait of interest and proposes methods that extend the current framework to include analysis of data from multiple ancestries and to reveal causal relationships. Chapter II and III propose methods to characterize spatial transcriptomic landscapes of complex tissues by inferring cell types, tissue anatomic structures, and gene-gene interactions, catalyzing new discoveries in many areas of biology.

Statistical Inference for Large-scale and Complex Structured Data
Bo Meng, Statistics

This dissertation addresses three important research challenges in high-dimensional statistical inference and complex structured data analysis. Specifically, the first part addresses problems in hypothesis testing for high-dimensional datasets. The high-dimensional settings often render traditional hypothesis testing methods invalid, potentially resulting in biased conclusions. Our results offer a comprehensive guide to determining the validity of testing methods in high-dimensional contexts. The second part focuses on developing a general framework for recurrent event data. Conventional recurrent event models frequently encounter computational inefficiency when estimating infinite-dimensional functional parameters. To address this issue, we introduce an innovative Ordinary Differential Equation (ODE) framework that unifies many existing recurrent event models as special cases and leverages established numerical tools to improve computational efficiency. The third part develops efficient statistical tools for analyzing complex-structured data. Specifically, we introduce a novel statistical inference framework for network data with signed edges, which signifies relationships of liking and disliking between different entities.

Dissecting the Mechanisms Shaping Liver Macrophages in Metabolic Liver Disease
Ziyi Meng, Molecular and Integrative Physiology

Nonalcoholic steatohepatitis (NASH) represents a severe stage of metabolic liver diseases characterized by liver injury, inflammation, and fibrosis. Macrophages are an integral part of the innate immune system and play a critical role in host defense and disease pathogenesis. However, the nature of macrophage heterogeneity, disease-associated reprogramming, and contribution to NASH pathogenesis remain obscure. My thesis is designed to dissect the transcriptomic signatures, disease regulation, and pathophysiological role of macrophages in NASH. In Aim 1, I established intrahepatic transforming growth factor β (TGF-β) as a driver of macrophage polarization in NASH liver. Specific ablation of TGF-β receptor in macrophages disrupted liver microenvironment and exacerbated diet-induced NASH pathogenesis. In Aim 2, I identified Basp1 as a novel mediator of proinflammatory signaling in macrophages through NLRP3 inflammasome activation. Together, these findings provide important new insights into macrophage biology in metabolic liver diseases.

Exploring Repetitive Elements in Neurodegenerative Disease Using Targeted Long-Read Sequencing
Camille Mumm, Genetics and Genomics

The repetitive nature of the human genome, with repeat sequences accounting for around 50% of its content, poses significant challenges to genomic analysis. With advances in long-read sequencing techniques, we are now better equipped to investigate variation in repeats and how they contribute to disease risk. My research utilizes nanopore targeted sequencing strategies to investigate two repetitive elements, short tandem repeats and mobile element insertions. Taking advantage of these new long-read technologies, we apply targeted sequencing to investigate neurodegenerative diseases. Our work uses post-mortem brain tissue samples and cell lines to explore repeat variation in neurodegenerative diseases and control samples across diseases including late-onset ataxias and Alzheimer’s disease (AD). In addition, we apply these techniques to investigate variation between cells. Thus, this work informs the role of repeat sequences in the human genome while simultaneously promoting the potential of long-read sequencing techniques in unmasking the mechanisms driving neurodegenerative diseases.

Synaptic Mechanisms of Top-Down Control by the Auditory Cortico-Collicular Pathway
Hannah Oberle, Neuroscience

Descending auditory cortical projections target the dorsomedial and lateral shell inferior colliculus (IC) to provide contextual information to ascending sound signals. However, the biophysical properties supporting ascending and descending pathway integration are unknown. Further, it is unclear the extent that corticofugal activity differentially modulates the dorsomedial and lateral shell. We addressed these questions using in vivo and slice electrophysiology, along with optogenetics and pharmacology. We determined ascending and descending signals arrive to the dorsomedial shell in quick succession and their integration can evoke an NMDA receptor-dependent supra-linear response. While descending activity enhances responses at most dorsomedial neurons, corticofugal-driven activity at lateral shell neurons is sparse and weak. Thus, corticofugal activity can powerfully modulate individual dorsomedial IC neurons but has a surprisingly limited effect on lateral shell neurons. These results suggest that the top-down control from the auditory cortex differentially shapes IC sub-circuits, with substantial modulation occurring at the dorsomedial shell.

Correlates of Protection for IAV, RSV, and SARS-CoV-2
Kalee Rumfelt, Epidemiological Science

This dissertation is focused on identifying adequate antibody correlates of protection for common respiratory viruses, IAV, RSV, and SARS-CoV-2.

Optimizing PDGFRA Inhibition Therapy in Pediatric High-Grade Glioma
Kallen Schwark, Cancer Biology

Pediatric high-grade gliomas (pHGGs) are lethal tumors, with 2-year overall survival of 10% and few effective treatment strategies. Platelet-derived growth factor receptor alpha (PDGFRA) is a receptor tyrosine kinase that is altered (mutated or amplified) in 21% of pHGGs; these tumors have significantly worse prognosis compared to wild type. The tyrosine kinase inhibitor avapritinib is FDA-approved in other solid tumors with PDGFRA exon 18 mutations, and it has been shown to be potent, specific to PDGFRA, and CNS-penetrant. However, avapritinib’s performance against the full spectrum of PDGFRA alterations seen in pHGG is unknown, and early clinical data demonstrates tumor resistance to avapritinib, as is common with targeted monotherapy. Thus, the objective of this project is to characterize avapritinib’s efficacy across PDGFRA variants and identify mechanisms of tumor response that would optimize combinatorial therapy. This knowledge will result in development of treatments that will prolong survival of pHGG patients.

Selections in Molecular and Phenotypic Evolution
Siliang Song, Ecology and Evolutionary Biology

The dissertation explores selections in molecular and phenotypic evolution based on large datasets of genotypes and phenotypes. Chapter 1 challenges the neutral theory of molecular evolution and proposes a quasi-neutral model that reconciles the abundance of beneficial mutations with seemingly neutral, long-term molecular evolution. Chapter 2 delves into the genetics of human same-sex sexual behavior (SSB), revealing a shift in its genetic maintenance due to widespread contraception. Chapter 3 establishes a genetic distinction between bisexual behavior (BSB) and exclusive SSB (eSSB) and discovered that alleles associated with male BSB are reproductively advantageous whereas those associated with eSSB are reproductively disadvantageous. Chapter 4 focuses on the heritability of human sex ratio at birth, using biobanks to identify genetic variants influencing sex ratio and evaluating Fisher’s principle in explaining sex ratio evolution. Collectively, these studies challenge existing paradigms of molecular and phenotypic evolution and provide fresh perspectives on the role of selection.

Integration Methods for Incorporating Published Prediction Models with New Time-to-Event Dataset
Di Wang, Biostatistics

Accurate survival risk discrimination can facilitate detection and prevention of complex diseases. With the abundant availability of data in the public domain, it is becoming an emerging challenge for researchers to consider incorporating external information from large-scale studies to enhance the survival prediction instead of only using a limited-sized dataset collected internally. This dissertation introduces novel data integration methods for effectively and efficiently incorporating different types of external information with internal time-to-event data, accounting for population heterogeneity. In Chapter I, we propose an integration framework that integrates a sequence of external survival probabilities to improve the post-transplant survival risk assessment in kidney transplant recipients. In Chapter II, an integration method that incorporates external risk scores is developed to enhance the cancer risk discrimination in minority populations. In Chapter III, we introduce a flexible cross outcome integration framework, which allows people to incorporate prediction models derived from other types of outcomes.

Novel Small Molecule Inhibitors of the GAS41 YEATS Domain
Alyssa Winkler, Chemical Biology

Epigenetic protein-protein interactions can control gene expression within the cell. Epigenetic reader proteins recognize post-translational modifications, which are chemical markers on histone proteins which are comprised of packaged DNA. Disrupting these interactions with small molecules is an opportunity to control gene expression and treat human diseases, such as cancers where the target protein is expressed in higher levels relative to normal cells. Here, I detail the structure-based development of first-in class, novel small molecule inhibitor of the GAS41 YEATS domain. GAS41 protein is highly expressed in non-small cell lung cancer and contains a YEATS domain which acts as an epigenetic reader to facilitate gene transcription. I demonstrate that blocking the YEATS domain reader activity using these small molecules decreases proliferation and modulates gene expression of non-small cell lung cancer cells. These small molecules represent a class of potential future therapeutics for non-small cell lung cancer.

Development of Antigen-specific Therapy for Autoimmune Diabetes Using Nanodiscs
Fang Xie, Pharmaceutical Sciences

Type 1 Diabetes (T1D) is an autoimmune disease characterized by the destruction of insulin-producing β-cells by autoreactive CD4 and CD8 T cells. Antigen-specific Tregs are efficient regulators in suppressing autoimmunity, which makes them attractive targets for tolerogenic therapy. However, it remains unknown how to efficiently induce antigen-specific Tregs in vivo. The goal of this project was to develop a novel synthetic high density-lipoprotein-based nanodiscs (ND) for the delivery of autoantigens and induction of immune tolerance in mouse T1D models. ND was prepared by lyophilization approach. By using BDC2.5 and NY8.3 adoptive transfer models of T1D, we optimized the dosing regimen and improved the efficacy in protecting against disease. We further developed a combinatorial regimen by incorporating multiple CD4 epitopes and IL-2/immunocomplex. In NOD mice at late prediabetic stage, the combination treatment prevented ~75% of mice from diabetes onset. Overall, my work suggests that the combination treatment induces antigen-specific Tregs that dampen ongoing autoimmunity, potentially through recognizing the same tissue-specific self-antigens or through bystander suppression.

Holistic Integration of Deep Learning Models for Mass Spectrometry-Based Peptide Identification
Kevin Yang, Bioinformatics

Identification of proteins and their constituent peptides provides valuable insights into biological systems. Mass spectrometry-based proteomics is high-throughput, allowing researchers and clinicians to quickly see a comprehensive picture of the proteomic landscape. However, mass spectra are noisy and complex. Machine and deep learning methods could improve peptide identification. In chapter I we present MSBooster, a tool that calculates similarity between predicted peptide properties and what is experimentally observed, and the benefits it brings in various experiments are shown. Chapter II extends this framework to different proteomics niches. Nuances of each are exploited for improved rescoring using specialized models. Chapter III presents a framework for transfer learning a user’s own models. MSBooster is open-source and implemented in the FragPipe comprehensive proteomics platform.

Unraveling the Structure and Organization of Circadian Clocks
Ye Yuan, Cellular and Molecular Biology

Circadian clocks control the timing of all vital physiological processes, ensuring synchronization with the daily environmental cycles. While the genetic framework of circadian rhythms is well characterized, the subcellular organization and regulation of cellular clocks in living cells remain unknown. Towards this goal, I employed advanced genetics and high-resolution imaging to investigate the spatial dynamics of clock proteins and discovered that: i) Clock proteins are compartmentalized away from chromatin into ‘inactive condensates’ at the inner nuclear envelope to facilitate gene repression, ii) the BAF chromatin remodeling complex is essential for condensing clock chromatin and silencing expression., iii) the Golgi complex and the centrosome serve as critical signaling hubs for circadian regulation, orchestrating the phosphorylation and subsequent degradation of key clock proteins. Together, my work has the potential to transform our understanding of the inner workings of circadian clocks, offering profound insights into the biological underpinnings of daily life.

Disparities in Flavored Cigarette Use at the Intersection of Multiple Social Identities and the Potential Impact of Banning Flavored Cigarettes in Mexico
Luis Zavala Arciniega, Epidemiological Science

Significance: Smoking is one of the leading causes of preventable deaths in Mexico and about half of the people who smoke used flavored cigarettes. By using a visualization tool, I identified disparities in flavor capsule cigarette use at the intersection of multiple identities. I will conduct an online survey to evaluate the response of people who smoke flavored cigarettes to a hypothetical flavored cigarette ban. I will adapt a mathematical model to estimate the potential public health impact of the flavored cigarette ban on the prevalence of smoking and the number of premature deaths attributable to smoking. Expected results: My dissertation will provide knowledge to tackle the threat of flavored cigarettes, which will help to reduce the incidence of tobacco-related health outcomes attributable to smoking.