UCSF

Over the past few decades, regulatory T cell (Treg) therapy has gained traction for the treatment of autoimmunity, transplant rejection, and other inflammatory diseases. A necessary part of this process is cellular manufacturing, where Tregs are activated with immunomodulatory agents, expanded, and quality assessed for CD4+ and Foxp3+ lineage purity before infusion. A challenge that has emerged in certain patient populations is the failure of Tregs to expand. This clinical manufacturing failure has been associated with the pre-expansion phenotype of the Treg, with a higher proportion of terminally differentiated effector Tregs correlating with poorer expansion. While the heterogeneity of Tregs phenotypic subsets and their expansive capacities has been reported, it is unclear what the required signaling strengths are to induce proliferation within each subset. To investigate this question, we used a biomaterial strategy for precisely attaching Treg stimulatory molecules onto polymeric particles using DNA to act as artificial antigen presenters. First, we optimized the fabrication technique, adapting numerous technologies commonly available to immunology labs, and exemplified its use in both human CD4+ and CD8+ T cell activation. Second, we then applied these materials towards activating human Tregs, identifying unique activation thresholds for inducing expansion and further subset-specific expansion behaviors. Third, unable to drive the expansion of effector Tregs using this approach, we turned to paired single cell RNA and TCR sequencing (scRNA/TCRseq) to identify the distinguishing genes associated with high expansion. Using TCR tracking of highly expanding clones, we were able to subset non-activated cells by their TCR clonotype expansion, enabling the comparison between transcriptomes of high expanders and low expanders. This led to the finding that highly expanding clones were associated with numerous markers associated with Treg stemness, proliferative capacity, and degradation of inhibitory intracellular signals, whereas poorly expanding clones express numerous differentiation and activation genes. Further, we identified a group of proliferative Tregs which displayed an inflammatory phenotype, which prompts further investigation into maintaining Treg purity in strategies that improve expansion. This work sets a foundation in studying Treg subset activation biology using precisely controlled signals and has uncovered numerous gene targets for improving Treg expansion. We believe that investigating these targets and optimizing their activation signals may provide a means of rescuing Treg proliferation in cases of manufacturing failures, reenabling Treg therapy as a possible treatment option for affected patient populations.

This dissertation presents a multifaceted study exploring critical aspects of drug development, focusing on the transporters on Blood-Brain Barrier (BBB) and a novel approach for Non-Alcoholic Fatty Liver Disease/Steatohepatitis (NAFLD/NASH).

The first segment of this research offers an in-depth analysis of the BBB transporters, focusing on age-related changes in protein expression and the functional impact of polymorphisms. The Blood-Brain Barrier (BBB) serves as a selective barrier for a variety of small molecules, including chemical carcinogens, environmental toxins, and therapeutic drugs. This barrier is constructed from brain capillary endothelial cells, pericytes, and astrocytic end-feet, working in unison to protect neurons and maintain brain homeostasis throughout life. The selective nature of the BBB is attributed to the tight junctions within the capillaries, which prevent the free passage of small molecules, and also to an array of transporters that regulate the influx of essential nutrients and the exclusion of many xenobiotics. The goal of this part of the dissertation is to elucidate the effect of aging from neonates to elderly on the human BBB, with an emphasis on transporters. A secondary goal is to examine in detail the determinants of function of a key transporter in the human BBB, ATP-binding cassette transporter, ABCG2 (BCRP).

This part of the dissertation begins with an overview of the current understanding of elements that influence the operation of the BBB, particularly transporters. It starts with a review of the BBB's structural components and their joint function in sustaining barrier integrity. It then summarizes how the BBB controls the movement of nutrients and medicines into the brain through various transport methods. The discussion includes the development of the BBB, how it changes as we age, and how diseases may affect it, emphasizing the dynamics of BBB and the challenges it creates for creating brain-targeted drugs. The chapter then shifts to how genetic variants in transporters influence BBB function and drug disposition, focusing particularly on the prominently expressed ATP-binding cassette transporters ABCB1 and ABCG2. After the overview, this dissertation provides a rich set of comprehensive analyses how age and genetic variations in transporters impact the functionality of BBB to advance our understanding of the multifaceted regulatory framework that controls BBB in physiological and pathological contexts. Major gaps in our understanding of the BBB, which are addressed by this dissertation research, are highlighted.

Chapter 2 examines changes in the human BBB proteome throughout a human lifespan, noting the significant shifts in protein expression that influence barrier permeability and the transport of nutrients and drugs. It is acknowledged that the BBB matures after birth, adjusting its transport mechanisms to align with each stage of development, and later alters due to aging and neurodegenerative disorders. Yet, fully grasping these modifications in the human BBB is an ongoing challenge. This chapter introduces a comprehensive proteomic analysis of the evolution and senescence of proteins in brain microvessels (BMVs). Samples from healthy individuals across a wide age spectrum and Alzheimer’s disease patients were analyzed using LC-MS/MS. A plethora of proteins, including numerous SLC and ABC transporters, were identified. Network analysis of the BMV proteome suggested potential alterations in BBB permeability over time and pinpointed transporters crucial for nutrient supply and drug penetration that exhibit age-dependent expression patterns. This investigation sheds light on the dynamic regulation of BBB proteins, emphasizing how transporter variations with age can affect drug permeability. These findings are crucial for refining pharmacokinetic modeling and therapeutic approaches across different stages of life.

The dissertation (Part A) then pivots to explore how genetic factors may alter the functionality of transporters, potentially causing variances in drug distribution within the brain. It focuses particularly on ABCG2, a transporter highly expressed at the BBB, noting that genetic variations leading to functional changes can result in differing drug responses. Utilizing deep mutational scanning (DMS), an innovative technique that combines next-generation sequencing (NGS) with functional outcomes of numerous variants, this study evaluated 12,724 variants of the ABCG2 gene. Our experimental setup was crafted to assess over ten thousand of missense, synonymous, and deletion variants of ABCG2 in a high-throughput manner. The abundance of ABCG2 was quantified, its surface expression was measured, and the functional effects of each variant were examined using the anti-cancer drug, mitoxantrone. The resulting detailed functional map, visualized through heatmaps and integrated with the structural data of ABCG2, helped identify crucial residues essential for ABCG2's function and poly-specificity. This study enhances our understanding of ABCG2 and lays the groundwork for future investigations into other ABC transporters. It underscores the value of DMS in dissecting the intricacies of pharmacogenetics and the mechanisms underlying drug resistance.

In summary, this part of the dissertation presents a comprehensive examination of the factors important for the functionality of BBB, shedding new light on transporters. Importantly, we unveil the BBB's dynamic protein regulation across the human lifespan, demonstrating how age-related changes affect drug permeability through a detailed proteomic analysis. Furthermore, the dissertation explores how mutations influence transporter functionality, ABCG2 as an example for developing the platform. The innovative use of Deep Mutational Scanning (DMS) to assess thousands of ABCG2 variants provides a rich functional map, revealing key insights into the transporter's operation and offering a valuable resource for future pharmacogenetic and drug resistance research. Overall, these studies highlight the necessity of understanding the BBB's complex mechanisms to enhance drug delivery strategies and overcome barriers in treating neurological disorders including neurodegenerative diseases.

The second part of the dissertation (Part B) shifts focus to the global health issue of NAFLD and its more severe form, NASH. It investigates Cis-Regulation Therapy (CRT) as a novel treatment approach, utilizing nuclease-deficient gene-editing technologies to modify gene regulatory elements for therapeutic ends. Approximately 30% of people worldwide are affected by NAFLD, and about 25% of these cases may advance to NASH. NASH represents a more serious stage of NAFLD, marked by liver inflammation and damage due to fat accumulation in the liver. About 25% of those with NAFLD progress to NASH, characterized by significant liver inflammation and damage due to fat accumulation. Currently, the pharmacological treatment options for NAFLD/NASH are severely limited. Our study investigates the potential of CRT as an innovative treatment strategy. In our research, we explored the effectiveness of CRT as a promising new treatment strategy. CRT employs nuclease-deficient gene-editing technologies, such as dead Cas9 (dCas9) combined with transcriptional modulators, to alter the activity of gene regulatory elements for therapeutic purposes. The goal of this part of the dissertation research specifically focuses on the nuclear receptor-like protein 1 (NURR1, NR4A2), a transcription factor critical in regulating inflammation which is a hallmark of NASH.

This part of the dissertation initiates with an overview of existing treatment options for NAFLD/NASH, pinpointing their limitations and the urgent need for more effective interventions. It further explores contemporary strategies in drug and therapeutic development targeting NAFLD/NASH, with a particular emphasis on animal models. After the overview, we present our findings that activating Nurr1 through CRISPR activation (CRISPRa) offers a promising therapeutic strategy for NAFLD/NASH within FATZO mouse models. This technique has shown efficacy in improving glucose metabolism abnormalities and reducing the CCL2-CCR2 axis, a critical inflammatory pathway, both before and after the onset of the disease. Our findings introduce a promising new therapeutic avenue for NAFLD/NASH, highlighting the capability of Nurr1 activation to control and possibly reverse the disease's progression.

In summary, this dissertation delivers a comprehensive analysis of the variables impacting the functionality of BBB transporters and presents a promising therapeutic approach for NAFLD/NASH. Through this research, we aim to pave new pathways for the advancement of treatments for neurological and hepatic disorders.

Bacteria use conserved signal transduction pathways, called sensory systems, to sense environmental stimuli. Most of our understanding of sensory systems in bacteria, however, comes from the chemotaxis system of Escherichia coli, which senses chemical gradients to control the direction of flagellar-based motility (chemosensing). Importantly, bacteria can also sense mechanical stimuli to actively shape their physiology. An in-depth mechanistic understanding of mechanosensory systems, when compared to their chemosensory counterparts, is however lacking. This dissertation presents work towards understanding mechanosensing in the important opportunistic human pathogen Pseudomonas aeruginosa. This Gram-negative bacterium uses Type IV pili (TFP), retractile polarly localized appendages, to sense mechanical forces generated during surface contact at one cell pole. We and others have demonstrated that spatially resolved mechanical stimuli transmitted by TFP activates the Pil-Chp mechanosensory system. Upon surface contact, TFP transmits mechanical stimuli to the Pil-Chp receptor, PilJ, thereby altering the autophosphorylation state of ChpA and thus the phosphorylation of PilG and PilH, the antagonistic Pil-Chp response regulators.

PilG and PilH inversely control two outputs of the Pil-Chp system in P. aeruginosa: cAMP production and twitching motility. Sensing of surface contact by the Pil-Chp system activates the membrane bound CyaB adenylate cyclase, which catalyzes the production of the second messenger, cyclic adenosine monophosphate (cAMP). cAMP binds to the Vfr transcription factor, leading to altered transcription of >200 genes involved in acute virulence as well as selected TFP regulatory proteins. Signal processing through PilG and PilH is critical for surface-dependent cAMP production. PilG promotes cAMP production and upregulation of the surface dependent transcriptional program while PilH has the opposite effect.

The Pil-Chp mechanosensory system is required for twitching motility, partially independently of cAMP levels. In Chapter 2, we demonstrate that P. aeruginosa actively directs twitching in the direction of mechanical input from TFP, in a process called mechanotaxis. The Pil-Chp system controls the balance of forward and reverse twitching motility of single cells in response to the mechanical inputs. We show that the Pil-Chp response regulators PilG and PilH control the polarization of the TFP extension motor PilB. PilG localizes to both poles, but shows greater accumulation at the leading pole, where it stimulates polarization favoring forward migration. In contrast, PilH, is primarily cytoplasmic, thereby globally antagonizing PilG. Subcellular segregation of PilG and PilH efficiently orchestrates their antagonistic functions, ultimately enabling rapid reversals upon perturbations. The distinct localization of response regulators establishes a signaling landscape known as local-excitation, global-inhibition in higher order organisms.

In Chapter 3, we demonstrate that PilG and PilH enable dynamic cell polarization by coupling their antagonistic functions on TFP extension. By precisely quantifying the localization of fluorescent protein fusions, we show that phosphorylation of PilG by the histidine kinase ChpA controls PilG polarization. Although PilH is not inherently required for twitching reversals, upon phosphorylation, PilH becomes activated and breaks the local positive feedback established by PilG so that forward-twitching cells can reverse. To spatially resolve mechanical signals, the Pil-Chp system thus locally transduces signals with a main output response regulator, PilG. To respond to signal changes, Chp uses its second regulator, PilH, to break the local feedback.

In Chapter 4, we report the mechanism of sensory adaptation in the Pil-Chp mechanosensory system. Bacterial sensory adaptation has primarily been studied in flagellar-mediated chemotaxis, where reversible methylation of sensory receptors by a methyltransferase and a methylesterase “tune” their sensitivity of signaling. The Pil-Chp system encodes the PilK methyltransferase, predicted to methylate PilJ, and the ChpB methylesterase, predicted to demethylate PilJ; however, whether sensory adaptation occurs in response to surface contact remained underexplored. Using biochemistry, genetics, and cell biology, we discovered that PilK and ChpB are segregated to opposing cell poles as P. aeruginosa explore surfaces. By coordinating the localization of both enzymes, we found that the Pil-Chp response regulators influence local PilJ methylation in vivo. We propose a model in which spatially resolved mechanical inputs transmitted by TFP not only alter PilG and PilH signaling mechanisms but locally controls PilJ methylation to modulate twitching motility reversal rates and surface-dependent cAMP production. Despite decades of chemosensory adaptation studies, our work has uncovered an unrecognized mechanism that bacteria use to achieve adaptation to mechanical sensory stimuli.

Acinetobacter species are opportunistic pathogens that are ubiquitous throughout the environment and are emerging as a public health threat around the world due to their widespread multidrug resistance. Intriguingly, many Acinetobacter strains encode homologs of the P. aeruginosa Pil-Chp mechanosensory system. In Chapter 5, we demonstrate that A. nosocomialis strain M2, a pathogenic member of the Acinetobacter calcoaceticus-baumannii complex, has a robust surface-dependent transcriptional response. We speculate that the homologous Pil-Chp mechanosensory system is responsible for the surface-dependent transcriptional response that we report in this dissertation.

Overall, this dissertation demonstrates that mechanosensing through the Pil-Chp system takes advantage of the intricate internal organization of bacteria to sense spatially resolved mechanical information. As medically Acinetobacter species exhibit a surface transcriptional response, defining the mechanosensing mechanism of Acinetobacter species represents an exciting area of investigation. Understanding the mechanisms of bacterial mechanosensing may lead to the generation of desperately needed therapeutics to treat multi-drug resistant infections, such as the ones typically caused by P. aeruginosa and medically relevant Acinetobacter species.

Background: The pregnancy desires of adolescents and young adults are not well understood, as researchers and health care providers have traditionally assumed that individuals in this age group want to prevent pregnancy. Efforts to promote reproductive autonomy and promote contraceptive use may have been misguided as they relied on research data based on this assumption. Traditional methods for assessing pregnancy intention use retrospective and binary measures that do not capture the range of feelings around pregnancy. Development of a new validated measure, the Desire to Avoid Pregnancy (DAP) scale, allows for a more holistic perspective, representing a range of pregnancy preferences. Methods: This dissertation study is a secondary analysis of data from the Attitudes and Decision After Pregnancy Testing (ADAPT) study, which recruited participants from March 2019 to October 2022. Using the subset of participants aged 15 to 24 years old at enrollment (N= 1,020), pregnancy preferences were measured with the DAP scale and various demographic, contextual, and economic participant characteristics were analyzed in relation to these preferences and contraceptive use. This dissertation research described the range of youth pregnancy preferences and investigated factors associated with these preferences. The degree to which pregnancy preferences are aligned with contraceptive use was studied and whether these preferences mediated the effect of contextual factors on contraceptive use. Subsequent analysis investigated participants whose pregnancy preferences did not align with their contraceptive use, to identify individuals who may be at risk of compromised reproductive autonomy. Results: Young people had a range of pregnancy preferences, including a high desire to avoid pregnancy, ambivalence, and a low desire to avoid pregnancy. Factors significantly associated with greater desire to avoid pregnancy were identifying as White (compared to Latine), having depressive symptoms, being enrolled in school, having a mother with higher educational attainment, and not having a main partner (compared to being in a high quality relationship). Factors significantly associated with more openness to pregnancy were having one child (compared to none) and being religious. Contraceptive use was more likely among youth who wanted to prevent pregnancy. Interestingly, both youth in high quality relationships (compared to no relationship) and religious youth were more open to pregnancy yet more likely to use contraceptives. As pregnancy preferences acted as a suppressing mediator for these two variables, the degree to which these participants are more likely to use contraception becomes more notable. At some points during the study period, participants’ pregnancy preferences did not align with contraceptive use and a few participants’ characteristics were associated with these discordant relationships. Contraceptive use declined by increasing age for both those who wanted to prevent and those open to pregnancy. Among those with a greater desire to avoid pregnancy, those not in school or not in a relationship were less likely to use a contraceptive. Among youth open to pregnancy, multiparous participants were more likely to use a contraceptive. Conclusion: Having a more nuanced understanding of how youth pregnancy preferences are impacted by contextual factors and how in turn these pregnancy preferences impact contraceptive use, can help to direct policy to help meet the reproductive needs of this population.

Trichothecenes such as T2-toxin are toxic natural products produced by several species of fungi that grow on grain crops. This class of sesquiterpenes comprises over 200 family members that feature a tricyclic core with varying degrees of oxidation and further cyclization, and has received attention for their toxicity as well as their anticancer, antifungal, and immunomodulatory effects. Cellular and animal model studies support the hypothesis that their manifold biological effects arise at least in part due to inhibition of protein synthesis. Here we report a modular synthesis of the minimal trichothecene pharmacophore verrucarol, enabling straightforward structural modifications to investigate toxicity and molecular mechanisms of action. We found that while verrucarol inhibits human cytosolic protein synthesis in vitro, functionalization of the C4 and C15 alcohols is required for potent cellular toxicity in cancer cell lines and fibroblasts. We characterized the binding of verrucarol and diacetylverrucarol to the 50S subunit of the human cytosolic ribosome at 2.7 A resolution, revealing binding determinants that could not be resolved from the previous structures in yeast ribosomes. This work will enable new studies of trichothecenes required for food safety and for exploration of their use as therapeutics.

Like any successful intracellular pathogen, Chlamydia trachomatis must overcome several host defenses to complete its developmental cycle. Chlamydia is thought to accomplish this through the employment of effector proteins, particularly a unique class of effectors which are inserted in the Chlamydial inclusion membrane. To test the hypothesis that these inclusion membrane proteins (Incs) target host cell proteins, our lab conducted an affinity purification-mass spectrometry screen (AP-MS) in which cells were transfected with individually tagged Incs. The predicted interaction between the Inc Tri1 and the host ubiquitin ligase TRAF7 is the subject of this work.In this dissertation, we characterize the Tri:TRAF7 complex during infection and reveal that Tri1 can disrupt native TRAF7 protein-protein interactions (PPIs, Chapter 2). Through additional analysis of this TRAF7 interactome dataset, we revealed that Tri1 can also promote novel TRAF7 PPIs (Chapter 3). Finally, we conducted RNAseq analysis to determine the role of Tri1 during infection and determined that it may play a modest role in altering signaling pathways. Overall, this work contributes to the growing literature on C. trachomatis Incs and highlights the important roles they may play during infection.

Protein-protein interactions (PPIs) are vital for biology as they govern numerous essential biological processes, including signal transduction, enzymatic activity regulation, protein localization, and complex formation, collectively orchestrating cellular functions and maintaining homeostasis. The hub protein 14-3-3 interacts with hundreds of client proteins, including key signaling proteins CRAF and SOS1 and transcription factors Estrogen Receptor α (ERα), FOXO1, and TAZ. These 14-3-3/client interactions can be altered in disease, leading to undesirable activity and signaling. The Arkin Lab has pioneered a mass spectrometry-based site-directed disulfide tethering screen to identify PPI stabilizers and inhibitors. We have used this screen to discover stabilizers of various 14-3-3σ/client interactions including, CRAF, FOXO1, ERα, SOS1, and USP8. Disulfide fragment stabilizers lead to a 4 – 250-fold stabilization of 14-3-3σ/client complexes. To further characterize the effects of these stabilizers, we have developed Nanoluciferase bioluminescence resonance energy transfer (NanoBRET) assays to measure 14-3-3σ/CRAF, TAZ, and ERα interactions, where we have measured stabilizer EC50 values between 100 nM – 1 μM. We have optimized 14-3-3σ/CRAF stabilizers for selectivity and potency and measured increased PPI formation in cells, decreased CRAF kinase activity, and decreased signaling through the mitogen-activated protein kinase (MAPK) pathway. Finally, we have expanded the disulfide tethering technology to a new class of hub proteins to discover stabilizers of additional PPIs to modulate their interaction in disease.

Addiction, including alcohol use disorder, is a maladaptive form of learning and memory. Initially, the effects of an addictive substance are rewarding, reinforcing the drug-seeking behavior. Over time, repeated exposure leads to changes in neuroanatomy, altering the brain’s circuitry and resulting in cycles of drug binge, withdrawal, and craving, undermining the individual’s health and functioning. This dissertation aims to examine the role of Rac1 in maladaptive learning and Prosapip1 in normal learning and memory. The first chapter examines a protein that has not previously been studied in mammals in the context of AUD, Rac1. In this chapter, we present evidence that Rac1 is activated in the DMS in response to repeated cycles of alcohol binge and withdrawal. This subregion-specific activation leads to phosphorylation of downstream proteins and promotes F-actin assembly, which then causes increased dendritic arborization and dendritic spine maturation. We also show that Rac1 in the DMS is involved in alcohol-associated goal-directed behavior, and therefore likely contributing to the progression of AUD. The second chapter investigates the physiological role of a protein previously associated with AUD, Prosapip1. We developed a Prosapip1 neuronal knockout mouse line to examine its mechanism of action in vivo. We present data to suggest that Prosapip1 is vital in regulating the PSD scaffold. Disruption of the PSD leads to a loss of LTP. Finally, these biomolecular changes result in a spatial learning and memory deficit, which is localized to the dorsal hippocampus. In summary, this dissertation contributes to the knowledge of molecular mechanisms controlling normal and maladaptive learning, and identifies potential targets for the treatment of AUD.

Liver cancer is diagnosed yearly in over 800,000 people worldwide. While several gene coding mutations have been found to be associated with this cancer, we lack an understanding of gene regulatory driver mutations that could also lead to liver cancer. Using short sequences, 16 base pairs in length, termed neomers, that are generally absent from the human genome but appear in liver cancer, we identify numerous potential causative gene regulatory mutations in promoters, potential enhancers and 3’UTRs. We identify noncoding neomers that are enriched in numerous patients and are thought to regulate cancer-associated genes. We also characterize specific one kilobase regions in the human genome that are enriched for noncoding neomers that reside near many relevant liver cancer genes. Combined, our study identifies novel mutations in gene regulatory elements that could be key regulators of liver cancer, providing a basis for potential downstream diagnosis and therapeutics.

Differential expression analysis of scRNA-seq data is central for characterizing how experimental factors affect the distribution of gene expression. However, it remains challenging to distinguish between biological and technical sources of cell-cell variability and to assess the statistical significance of quantitative comparisons between cell groups. In this thesis, we introduce memento to address these limitations and enable statistically robust and computationally efficient differential expression analysis of the mean, variability, and gene correlation from scRNA-seq. We used memento to analyze 70,000 tracheal epithelial cells to identify interferon response genes with distinct variability and correlation patterns, 160,000 T cells perturbed with CRISPR-Cas9 to reconstruct gene-regulatory networks that control T cell activation, 1.2 million PMBCs to map cell-type-specific cis expression quantitative trait loci (eQTLs), and arbitrary cell groups within the entire 50 million cell CELLxGENE Discover data corpus. In all cases, memento identified more significant and reproducible differences in mean expression but also identified differences in variability and gene correlation that suggest distinct transcriptional regulation mechanisms imparted by cytokines, genetic perturbations, and natural genetic variation. These results demonstrate memento as a first-in-class method for the quantitative analysis of scRNA-seq data, scalable to millions of cells and thousands of samples.