Brain cells depend on the extracellular matrix (ECM) for structural and functional support as well as sequestration and transport of key ions and neurotransmitters. Structural and compositional changes to the ECM occur in development and in response to injury and disease. Probing ECM structure and composition in real-time in a dynamic living brain would enhance our understanding of the ECM changes that drive disease. In our work, we use organotypic whole-hemisphere (OWH) brain slices to study the interaction between brain parenchymal cells and the ECM. We have applied multiple-particle tracking (MPT), an imaging technique that tracks movement of nanoscale probes with sub-micron resolution, to OWH slices exposed to different stimuli, including oxygen-glucose deprivation (OGD) and mitochondrial dysfunction by rotenone (ROT) exposure.  Our MPT data confirmed that ECM microstructure changes in a time and stimuli-dependent manner and this was associated with changes in cellular composition and morphology. In this study, we measured changes in expression of ECM transcripts using Reverse Transcription quantitative Polymerase Chain Reaction (RT-qPCR) of RNA isolated from OWH brain slices exposed to OGD and ROT. After exposure to 30 minutes of OGD or treated with 50 nM ROT, OWH slices were preserved at 2h and 24h in RNALater buffer for RNA isolation. The 2h time point aligns with the end of the MPT experimental window. Healthy unexposed OWH slices were controls. We measured expression of genes associated with ECM composition and remodeling including tenascin-R, aggrecan, neurocan, MMP9, and TIMP1; markers associated with cellular activation including Ki67, Cd45, Cd11B, and GFAP; inflammation markers including IL-1β, IL-4, IL-6, IL-9, and IL-10; and cell death markers including iNOS, nNOS, TNF-α, and Casp-3. Our results provide a quantitative measure of ECM composition that can be integrated with our MPT and imaging data to better define microstructural dynamics in the stimuli-exposed brain.

Protein kinases have been found to regulate cellular processes such as growth and stress response. Thus, they act as excellent targets for drug treatment. The Mycobacterium tuberculosis (Mtb) genome encodes 11 serine/threonine protein kinases. Our lab has previously found that two of these kinases, PknF and PknL, show a large survival deficit when induced. They phosphorylate similarly throughout central carbon metabolism (CCM), a process known to be involved in cellular survival. To test kinase regulation of different pathways in CCM, I tested the growth of avirulent Mycobacterium tuberculosis (aMtb) strains expressing PknF or PknL using two different carbon sources: propionate and succinate. Propionate is broken down into propionyl-CoA, a toxic co-intermediate, which passes through the methylmalonyl or methylcitrate pathway to enter the citric acid cycle at succinate. The methylmalonyl pathway requires vitamin B12 to proceed and prevent toxic propionyl-CoA build up. Thus, propionate + B12 was tested to further elucidate regulation of these pathways. I measured colony-forming units (CFU) to quantify aMtb survival in these growth conditions. I compared survival measurements of the PknF and PknL induced strains relative to an empty vector control strain. I found that PknF induced grown with propionate showed a greater survival deficit by day 7 compared to the strain grown in succinate. Interestingly, the addition of B12 did not rescue growth as it did in the empty vector control. PknL induced grown with propionate shows a greater survival deficit compared to succinate; however, the addition of B12 decreased the survival deficit experienced in propionate. Due to this difference between B12 phenotypes, we hypothesize that PknF induction is regulating the methylmalonyl pathway, resulting in no rescue of the survival deficit. These findings can be used to inform future studies on PknF and PknL as potential targets for tuberculosis treatment during infection.

Previous research has determined that nanoparticle systems require a wide parameter space to effectively conduct synthesis and characterization. As a result, the development of high-throughput techniques is essential for efficiently analyzing the large datasets produced in colloidal particle experiments. These methods enable the rapid assessment of particle properties, such as size, shape, and charge, which are critical for modifying nanoparticles for specific applications. In order to do this, advancements in automated synthesis platforms, such as the Jubilee automated multi-tool system, offer the potential to streamline the fabrication of magic sized clusters. This approach has the potential to accelerate the discovery of novel nanoparticles but also allows for real-time adjustment of synthesis parameters to achieve desired properties with high precision, throughput, and reproducibility. As a result of the optimized synthesis process, characterization using techniques such as small angle X-ray scattering (SAXS) and UV-vis spectroscopy can be done at an accelerated rate. Efforts to enhance the durability and performance of the Jubilee automated multi-tool platform are focused on integrating advanced materials to improve system lifespan. This work will incorporate glass syringes and resin-printed components which offer improved chemical resistance and precision compared to traditional plastic components, extending the utility of the platform to be able to work with solvents and chemicals that are corrosive, volatile, or strong solvating agents for typical plastics. These improvements aim to reduce wear and tear, extend the lifespan of critical components, and ultimately ensure the platform's reliability for long-term use in high-throughput nanoparticle synthesis.

Silica nanoparticles have diverse applications in catalysis, imaging, and drug delivery. Tailoring these nanoparticles for specific applications requires precise control over their size, surface chemistry, porosity, and polydispersity. These properties are controlled by a wide range of factors such as reactant type and concentration, pH, reaction temperature, and other synthesis parameters. Due to the large parameter space, determining the optimal reaction conditions for synthesizing silica nanoparticles with the desired size and morphology is time-consuming and challenging. An accelerated experimentation platform integrating automation and artificial intelligence can streamline the selection of reaction parameters for synthesizing silica nanoparticles with targeted size and morphology using machine learning-based iterative design of experiments to optimize material properties. This system uses the Science Jubilee flexible laboratory automation platform to carry out sol-gel synthesis. Small-angle X-ray scattering is used to characterize the sample. The data collected is used to optimize the reaction condition for synthesizing the targeted nanoparticle. We have successfully carried out sol-gel processes and synthesized silica nanoparticles with various sizes and polydispersity using the platform. Currently, we are working on optimizing the selection of sample synthesis conditions.

The ability to manipulate and ligate proteins has been a driving force in advancing our understanding of the complex regulation of biological processes in space and time. Protein ligation, in which two or more polypeptides are covalently linked, is a powerful strategy in biomacromolecular engineering, enabling precise control over protein modifications, stability, and functionality. This is particularly useful in understanding protein function and interactions, as well as modulating protein activity, including immobilization of protein-based signals within materials triggered by cytocompatible light. One proven system known for its specificity and ease of use is SpyTag/SpyCatcher, a peptide-protein pair capable of irreversible ligation via isopeptide bond formation. Recent work has demonstrated the ability to control SpyTag/Catcher ligation using cytocompatible light due to its non-invasive nature and spatiotemporal (i.e., 4D control) manipulation of protein signals on a biologically relevant timescale. However, the application of this reported photoligation strategy is hindered by the use of genetic code expansion which limits protein yield, entails additional orthogonal protein machinery, and involves translational incorporation of a non-canonical amino acid. To address these challenges, we aim to develop a photocontrolled protein ligation strategy using native protein activity while maintaining spatial and temporal control. We predict this strategy will enable dose-dependent reconstitution of ligation by varying light exposure duration and intensity in native protein systems while sidestepping challenges associated with genetic code expansion. We intend to use this strategy to further assess our capability to control split protein reconstitution and for future applications in directing complex cell fate, which has significant utility in stem cell biology and regenerative medicine.

There are a vast number of pathogens that impact global public health, necessitating an accessible assay capable of detecting multiple targets simultaneously. Lateral flow assays (LFAs) have the potential to fill this role as a cost-effective, rapid, and simple technology instrumental in the detection of many analytes. However, multiplexed detection using nucleic acid LFAs is difficult due to the increased chance of non-specific binding as more targets are added to the assay. In this work, we aim to increase specificity and multiplexing potential in LFAs. We first showcase the process of developing a nucleic acid LFA by evaluating both fluorophores and gold nanoparticles to generate a visible signal. As fluorophores require a fluorescent light source, we moved forward with gold nanoparticles, which have a readout visible to the naked eye. Additionally, we automated the LFA fabrication process using an Echo Liquid Handler. Finally, we assessed methods to convert double-stranded to single-stranded DNA, required for compatibility with LFAs. In the future, we look to optimize signal visibility while increasing multiplexability. This work highlights the potential of multiplexed LFAs as a robust technology capable of significantly improving public health responses and outcomes.

Plastic is a prevalent and useful material, experiencing a 230-fold increase in production from 1950 to 2019. However, problems revolving around plastic production and disposal are becoming increasingly clear. In 2015 virgin plastic production was responsible for 4.5% of global greenhouse emissions, with the United states collecting only 9% of plastic waste for recycling. For polystyrene, a common plastic in food storage containers, less than 1% was recycled. The decomposition of polystyrene can also release dangerous chemicals, such as additives Bisphenol A (BPA) and polystyrene oligomers, directly linked to human health hazards such as diabetes, breast cancer, reproductive harm, thyroid regulation issues, heart diseases, and liver problems. To prevent both the need for new plastic and decrease plastic decomposition products from harming both humans and the environment, it is critical to develop a circular plastic economy, where plastic waste is broken down back into the chemical building blocks, or monomers, which can be re-manufactured into new high-quality plastics. The use of basic heterogeneous catalysts, such as zinc oxide (ZnO) can improve the yields of polystyrene deconstruction back into monomers compared to other chemical recycling techniques such as pyrolysis and acid-catalysis. To further improve yields, it is critical to understand the impact of relative basicity of ZnO on styrene yield. To examine base strength effects, the relative basicity of ZnO will be systematically varied through alterations in pretreatment methods. Reactions were tested in a mini-batch reactor at 623K over three hours using a 10:1 substrate to catalyst ratio, with product distribution examined using Gas chromatography-mass spectrometry (GC-MS) and a flame ionization detector (FID). 

Plastics have transformed the modern materials landscape, as they have a large range of applications at low costs. With this widespread and consistent use it is seen that production will continue to increase, reaching over 1.1 billion tons per year. Of this, only around 16% of plastics are recycled, with complex polymer recycling rates even lower. Remaining plastics are primarily landfilled or incinerated, leading to associated environmental impacts. In response to this issue, chemical recycling methods such as hydrogenolysis have been developed. However, these methods are limited by their ability to separate or otherwise handle mixed plastic waste feedstocks. This work investigates ruthenium-catalyzed hydrogenolysis of mixed polyolefin waste to determine the impact of mixed feedstocks on catalytic activity.  The system of pure polyethylene (PE) and polypropylene (PP) mixtures is studied as PE and PP are two of the most commonly produced plastics, together comprising nearly 50% of all plastics produced. Model compounds tetracosane and squalane are used as analogs of PE and PP, respectively. These compounds are simpler than their polymer counterparts while possessing similar structural elements, therefore making for an easier determination of mechanistic pathways and kinetic parameters. Reactions are performed in 10mL stainless steel reactor systems using a ruthenium on carbon (Ru/C) catalyst. Squalane and tetracosane are mixed at various ratios, maintaining 1g of substrate and 100 mg of Ru/C per reaction. Relatively mild conditions of 225°C, 20 bar of hydrogen are held for 1 hour. Preliminary results find the C-C bonds in tetracosane cleave to a greater extent than in squalane, with much of the squalane preserved. Likewise, reactions with a higher percent of tetracosane result in more bond cleavage compared to squalane-rich reactions. This indicates that branched polymer substrates may decrease overall catalytic activity compared to linear polymer substrates. 

Most of our chemicals come from petroleum, a nonrenewable resource and a significant source of pollution. Purple non-sulfur bacteria (PNSB) also produce some of these chemicals from one-carbon (C1) feedstocks, however, genetic engineering toolkits are underdeveloped for these organisms. The ability to integrate heterologous genes is a crucial component of genetic engineering toolkits, enabling stable and precise gene expression. Despite their metabolic versatility, PNSB lack well-characterized genomic integration sites, limiting advanced strain engineering efforts. Here, we identify and characterize genomic integration sites in Rhodobacter sphaeroides 2.4.1 that can serve as stable integration loci for heterologous gene expression. Using RNA-Seq transcriptomic data, we identified intergenic regions with minimal transcriptional activity, ensuring that insertions into these regions would not disrupt native gene function. Seven candidate integration sites were selected across the genome, spanning both chromosomes and plasmids. Two-step allelic exchange was used to integrate “landing pads” for Serine Recombinase-Assisted Genome Engineering (SAGE), a site-specific recombination system, into candidate sites. Our next step is to use the SAGE system to integrate fluorescent reporters into these sites to assess positional effects on gene expression. These seven integration sites serve as a testbed, allowing us to validate the workflow for integration into a broader range of genomic locations. Our findings will provide a resource for engineering R. sphaeroides and expand the genetic toolkit for PNSB, facilitating their use in synthetic biology and bioproduction applications.

Gold nanoparticles (AuNPs) have unique optical and physical properties that have a range of applications in photovoltaics and medicine. The properties of AuNPs can be adjusted depending on their intended use, which is accomplished by synthesizing AuNPs of a specific size, shape, and surface chemistry. Optimizing AuNP structure is currently performed through a time-consuming approach. In experimental synthesis a multitude of parameters can affect the AuNP structure, including temperature, reagent concentrations, time delays of component addition, and the use of selective passivation molecules during synthesis. In order to achieve robotic control over the large design space, a computational method called phase-mapping can be utilized. These algorithms correlate the different synthesis design variables to the AuNP structure measured using characterization, and from that information the algorithm can provide synthesis parameters to create a desired AuNP structure. In this poster, an experimental case study of creating phasemaps of peptide-based AuNP synthesis by varying temperatures and the ratio of peptides in the growth solution will be presented. To produce enough experimental data to create an accurate phase-mapping algorithm, the synthesis process will be automated using an Opentrons OT-2 liquid handling robot, with an attached thermal module to control the synthesis temperature. After synthesizing the AuNPs, their structure will be characterized using UV-Vis spectroscopy. The structure, alongside the design parameters, will be used to update the phase-mapping algorithm, from which new design parameters will be obtained and synthesized in order to validate if the produced structure matches the algorithm’s prediction. The phasemaps generated will be used to understand the design rules for controlling the colloidal AuNP growth and further guide the bio-inspired synthesis of colloidal nanoparticles.

Industrial chemical separation processes, such as distillation, drying, and evaporation, consume 10-15% of US annual energy production. Membranes, which act as a selective barrier to separate compounds, are substantially more energy efficient than traditional chemical separation methods that require heat and could help reduce this consumption. Inorganic membranes are inherently suitable for many separation processes because they are chemically and thermally stable; however, ceramic membranes are mechanically fragile and costly to produce. Commercial polymeric membranes are comparably more economical but degrade in harsh organic solvents and high-temperature environments. One approach to achieve the necessary membrane properties at low cost is vapor phase infiltration (VPI), a gas-phase synthesis technique consisting of sorption, diffusion, and entrapment of vapor-phase reactants within organic polymers. The infiltration of inorganic oxides through VPI has been shown to enhance the properties of polymeric membranes by producing cost-effective, chemically stable, and temperature-tolerant organic-inorganic hybrid materials. However, the mechanical properties of these hybrid membranes, which are crucial for maximizing lifetime and durability, are generally less well understood. In this study, polyethersulfone (PES) membranes are subjected to trimethylaluminum and water under various VPI process conditions in a custom-built reactor. Thermogravimetric analysis is utilized to quantify the extent of inorganic infiltration by measuring the aluminum oxide loading within PES membranes. Mechanical properties of these membranes are characterized by tensile stress, modulus, and maximum pressure through dynamic mechanical analysis and burst pressure testing. Enhancement in chemical stability is determined by measuring the degradation of VPI-treated PES samples after exposure to organic solvents. These results provide insight into the relationship between infiltration structure, membrane stability, and mechanical properties, which may allow for improved membrane design and more sustainable industrial chemical operations.

Per- and polyfluoroalkyl substances (PFAS) are highly toxic contaminants shed from man-made chemicals which are still being used in consumer and industrial applications. Unfortunately, strong carbon-fluorine bonds present within PFAS prevents their natural degradation in the environment, leading to PFAS accumulation. Membranes, particularly those used for desalination, have been shown to be effective at removing many types of PFAS from water and are less expensive and energy intensive when compared to other removal approaches. However, new membrane materials are needed that can remove even the smallest PFAS molecules. In this project, we are developing new membrane materials aimed at being more effective than commercial nanofiltration and reverse osmosis membranes using molecular layer deposition (MLD), a technique that can deposit and precisely control membrane chemistry. First, commercial membranes from DuPont (NF245, NF270, and Seamaxx) were tested for their pure water permeability as well as rejection of salts and PFAS of varying carbon chain lengths, the results of which were used as an experimental control. Next, polymer membranes were made using MLD. These MLD-based membranes were synthesized and tested, and their results were compared to the commercial membranes for efficacy. This work hopes to develop new membrane chemistries that are more effective at removing PFAS than existing commercial materials.

Future technological developments in fields including alternative energy and medicine require next-generation materials. Synthesizing each new material requires exploring a multi-dimensional parameter space. Developing laboratory automation tools for automating lab procedures and data analysis will be key to efficient discovery of optimal, novel materials. Some automation tools utilized in this work include automated sample loading and analysis for both Small Angle X-ray Scattering (SAXS) and Dynamic Light Scattering (DLS), and a custom sonication robot. The goals of this project are to apply these lab automation tools to construct and characterize crystalline structures of nanoparticles encapsulated in lipid membranes and connected with DNA linkers. With high throughput methods, the impact of design parameters on the crystal structure can also be determined. Parameters of interest in the self-assembly of particles include the molar ratio of lipid membrane components and the nanoparticle surface area to membrane surface area ratio. The first assembly step is embedding the nanoparticles in a lipid membrane of optimal composition. Next, the cholesterol end of synthesized DNA-cholesterol fragments embeds in the membrane and complementary DNA fragments are added to connect the nanoparticles when combined with a complementary DNA bridge. The aggregates formed are analyzed with Zeta potential, SAXS, and DLS to determine if crystals are formed. Preliminary results from this project are presented here.

Multiple Particle Tracking (MPT) is a powerful technique for studying microscopic particles, such as viruses and nanoparticles, by tracking individual displacement and movement. One application of MPT is to measure microstructural changes in the brain extracellular environment (ECM) in development, aging, and disease progression. MPT of nanoparticle probes generates thousands of trajectories, from which geometric features, diffusion coefficients, and viscosities can be extracted. The vast array of trajectories presents an opportunity for deep learning models to uncover meaningful insights. However, to enable MPT data to be trainable and predictable by deep learning models, we need to curate the data to be useable by these models. To enable this, I have created a database and developed a data architecture that would allow MPT data to be useable within deep learning models. Building upon this foundation, I am currently working on creating a Self-supervised deep learning model utilizing equivariant graph neural network, equivariant transformer, and Explainable AI methods. The current iteration of this model can predict a masked point of a trajectory with a 34% error rate. The goal is to reduce this error to 10% and, more importantly, to differentiate between healthy and pathological trajectories. To achieve this, we will use Saliency Maps, an Explainable AI method, to understand how the model distinguishes between these two datasets. This approach will provide insights into which part of the trajectory the model finds most relevant. My hypothesis is that the model can effectively learn to distinguish between healthy and pathological trajectories based on the trajectory properties with an error rate of 10%. I will verify my model by modifying the trained model’s output layer to explicitly classify trajectories as healthy or pathological. By fine-tuning this model, we will evaluate performance using error metric, which I will further validate using Saliency Map visualizations.

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease that damages motor neurons, leading to severe disability within 1-3 years of diagnosis. Though its precise mechanism is unknown, chronic microglial activation has emerged as a hallmark of ALS pathophysiology. This results in persistent neuroinflammation and a positive feedback loop of cell death. Anti-inflammatory drugs could help restore microglia to a neuroprotective state. However, delivering these therapeutics across the blood-brain barrier and into disease-mediating cells presents a major challenge. Our prior work demonstrated that poly(lactic-co-glycolic acid)-poly(ethylene glycol) (PLGA-PEG) nanoparticles can overcome barriers to the brain in models of neurodegeneration such as Huntington’s disease. PLGA-PEG nanoparticles further exhibit localization and uptake in microglial cell populations. In this study, we aimed to develop PLGA-PEG nanoparticles for targeted delivery of danirixin (DNX), an anti-inflammatory agent, in ALS. We formulated DNX-loaded PLGA-PEG nanoparticles (PLGA-PEG/DNX) with different mixed organic solvents via sequential nanoprecipitation. Nanoparticle characterizations included dynamic light scattering for size, dispersity, and surface charge determination. We quantified drug loading and release using liquid chromatography-mass spectroscopy. PLGA-PEG/DNX achieved physical properties for effective brain delivery, including a small hydrodynamic diameter (<100 nm) with narrow dispersity (<0.20) and near-neutral surface charge (-10-0 mV). We identified an optimal mixed organic solvent system for synthesizing PLGA-PEG/DNX with high drug loading (>30%) and encapsulation efficiency (>70%). We further show that DNX retains activity following PLGA-PEG encapsulation with suitable lyophilization stability for in vivo administration. Future work will evaluate dose response, therapeutic efficacy, and pharmacokinetic properties for PLGA-PEG/DNX in pre-clinical ALS models. Successful completion of this study could help advance nanoparticle-based therapies into ALS clinical trials.

Architectural and spatiotemporal aspects of epigenetic regulation and cell behavior are critical for maintaining overall health. Unintentional genetic mutations can create dynamic dysregulation in the epigenome and transcriptomes at the cellular level which is implicated in diseases ranging from fibrosis to cancer. However, our tools to probe and understand these behaviors are limited by a lack of spatiotemporal control. To address this, we propose installing four-dimensional control over the potent CRISPR inhibition transcriptional effectors to establish epigenetic control at cellular scale resolutions.  CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) is a genetic modification system that relies on Cas9 proteins to splice and deactivate genes as controlled by a programmable guide RNA sequence. CRISPR inhibition relies on a deactivated Cas9 protein that does not directly alter the genetic material in order to sterically hinder transcription. Our work aims to formulate a CRISPR inhibitor system which can unbind from the target DNA with two photon activation via a photo-cageable noncanonical amino acid insertion. This would allow for four-dimensional spatiotemporal control over the system, thus increasing the level of control in epigenetic regulation. Currently, work is being done to test the CRISPR inhibition system in HEK 293 cells that have been lentivirally transduced with both a test sequence and the deactivated Cas9 protein. After testing is completed for this simpler system, we will move toward creating a system that incorporates the photocaged noncanonical lysine variant, giving us control over the CRISPR inhibition system with regards to both space and time.

Biological processes rely on the intricate functions of proteins, which drive essential biochemical reactions. Given their critical role, various methods have been developed to regulate protein functions in biomaterials and in vitro. Enhancing the precision of gene editing is crucial for advancing applications in gene therapy and minimizing off-target effects. My project focuses on integrating photoactivatable proteins with prime editors, a modified version of the widely known gene editor CRISPR/Cas9, to improve spatial and temporal control over gene modifications. By utilizing genetic code expansion, non-canonical amino acids are incorporated into human cells to express photocaged prime editor proteins and altering host genomes. This system enables optical stimulation to precisely regulate protein activity. Through the deployment of well-characterized photolabile groups, we expect to be able to render protein activity controllable in a dose dependent way. A key application of this approach is the development of a photoactivatable prime editor system to induce precise gene edits. Traditional CRISPR/Cas9 methods lack spatiotemporal control over activation. To address this, the system is adapted for use in hydrogels, where two-photon patterning allows visualization of prime editor protein activation in three dimensions. Our study aims to demonstrate the feasibility of optically controlling gene editing with high specificity, offering a novel strategy for advancing cell lineage tracing and gene therapy applications.

Nanoparticles are drug delivery carriers on the nanometer-length scale, and are promising targeted drug delivery solutions due to their small size and tailorability. However, current materials used to produce nanoparticles are synthetic and typically lead to large amounts of chemical waste and high costs. To explore more sustainable technologies, the Nance and Roumeli labs established a novel bacterial cellulose nanoparticle (BCNP) platform. BCNPs are formulated with a bacteria that produces cellulose and no byproducts when cultured, allowing for less reagents required and non-toxic biodegradable wastes. To be comparable to synthetic nanoparticles as a drug delivery platform, BCNPs must load and release drugs and be biocompatible with mammalian cells. In this project, I explored the tunability of BCNPs through size modification, performed cytotoxicity studies on a microglial cell line, and carried out drug loading studies. I found that higher mixing speeds during BC culturing led to a smaller BCNP size and variable particle concentration. Through cytotoxicity analysis in cell culture, I showed BCNPs were not toxic. Ongoing studies are assessing BCNP cytotoxicity as a function of BCNP dose. To demonstrate drug loading, I am incorporating catalase, an enzyme with the ability to mitigate oxidative stress markers, into BCNPs to analyze their efficacy in an in vitro model of oxidative injury. These results show BCNPs have the potential to become a sustainable nanomedicine platform and provide an important step towards reducing the environmental impact of synthetic nanoparticles.

Therapeutic delivery to the brain is challenging due to restrictive barriers such as the blood-brain barrier and the brain-parenchymal barrier. Although nanoparticles help overcome these barriers and improve therapeutic uptake, many nanoparticles are developed from synthetic materials and generate significant harmful waste. Bacterial cellulose nanoparticles (BCNPs) offer a sustainable alternative to current synthetic carriers. As a new platform, evaluating cytotoxicity and localization is essential to determine BCNP biocompatibility and potential for targeted drug delivery. To produce BCNPs, a BC pellicle was grown with gram-negative bacteria in the presence of yeast and washed with sodium hydroxide and deionized water. The BC was chemically and mechanically dissolved via sonication with dimethylacetamide and lithium chloride. Then, the BC dissolution media was added dropwise into a Pluronic F127 surfactant solution at room temperature and incubated for 2 h under stirring conditions to produce BCNPs. After washing and filtration, BCNPs were ~100 nm in size, had a slight negative zeta-potential, and demonstrated a polydispersity index <0.3, all parameters necessary for brain-targeting drug delivery. BCNPs were labeled with varying concentrations of carbotrace 680, a fluorescent dye used to specifically label cellulose materials. Cytotoxicity of BCNPs was assessed using healthy 10-day-old postnatal rat brain slices cultured for 4 days in vitro. BCNPs were topically applied to the brain slices (n=3 per experimental condition) at doses of 97 µg/mL – 290 µg/mL and incubated for 24 h. Slices were stained with propidium iodide (PI) before fixation and 4’,6-diamidino-2-phenylindole after fixation and imaged on a confocal microscope to quantify PI+ cells and determine BCNP localization. BCNPs resulted in <20% cytotoxicity at the applied doses confirming BCNPs do not cause cell death. These results demonstrate BCNPs are biocompatible and a promising alternative to synthetic carriers for drug delivery to the brain.

Tumor angiogenesis is characterized by unregulated blood vessel formation, impairing vascular networks and biological transport. It represents a critical stage in cancer progression, where malignant tumors metastasize and exploit the human body’s resources, which lie in vascular networks. However, the complex tumor microenvironment presents significant challenges in studying tumor angiogenesis and identifying its biomarkers. Towards addressing this concern, hydrogels—water-swollen, polymeric networks—can be used to recapitulate the tumor microenvironment, whose physiochemical properties can be precisely tuned to match that found in vivo. The DeForest Lab has developed methods and techniques in bioorthogonal chemistry and light-based subtractive manufacturing to tune such hydrogel materials with precise and unique 4D control, all at subcellular resolutions. In this project, I will exploit image-guided multiphoton lithography to engineer natively complex tumor vasculature patterns within photodegradable hydrogels. We will further embed tumor vascular spheroids within these hydrogels, providing a platform to model and assay tumor progression in vitro. This study has exciting implications for translational research and preclinical studies, particularly for disease modeling and therapeutic screening, as well as reducing ethical concerns regarding tissue and animal models in preclinical studies.

Hydrogel biomaterials have many applications in tissue engineering and drug delivery. Stimuli-responsive hydrogels allow for controlled drug release, dependent on a user-defined trigger. However, current stimuli-responsive hydrogels are case-specific and cannot be used for broader applications, such as targeted disease treatment. Most hydrogels can only respond to one input, making them difficult to use in treating diseases with multiple markers. We developed a fully recombinant protein-based material with protease degradable cross links that follow Boolean logic (YES/AND/OR) in response to multiple inputs to allow for user controlled material degradation and drug release. The protease degradable sequences can be easily switched out before expression depending on the application, making our hydrogel generalizable. The hydrogel will be crosslinked with Boolean logic constructs, each of which are flanked by a click-like chemistry protein system. This allows the crosslinks to be covalently ligated to a linker made from elastin-like polypeptides (ELP), which holds the hydrogel network together. The crosslinks and ELP were expressed recombinantly in E. coli and purified on an ӒKTA Pure (Cytiva). A degradation study was conducted by adding different combinations of proteases to prove that material degradation is dependent on the combination of proteases added. We then conducted rheometry to determine the mechanical properties of the hydrogels, and verified that material stiffness followed the expected logical operation, where correct inputs resulted in material degradation. Finally, we tested the hydrogel’s ability to release drugs by incorporating human epidermal growth factor (hEGF) into the gel and measuring activation of the ERK signaling pathway through a Western Blot. The Western Blot showed activation of the ERK pathway only when the correct combination of proteases was added, indicating release of a bioactive protein drug. If successful, this hydrogel could be used for therapeutic delivery of drugs and broader tissue engineering applications.

User-controlled cell behavior is useful for studying wound healing because the isolated therapeutic effects of individual signals can be observed at the wound site. Aptamers are single-stranded oligonucleotides that fold into three-dimensional structures that can capture and inhibit proteins. The biological capacity of cells to deploy traction forces as a release mechanism for extracellular proteins can be engineered through clever deployment of aptamer-bound proteins with peptide handles. Scientists at the Imperial College London recently synthesized TrAPs: Traction Force-Activated Payloads that enable precise control of cell behavior using such a strategy. We bound photocaged TrAPs lacking adhesion handles to functionalized collagen hydrogels. Peptide immobilization was then selectively patterned using 365 nm light to spatially confine cell access to captured vascular endothelial growth factor (VEGF). After surface seeding endothelial cells, observations were made regarding changes in cells’ physical characteristics as a result of  protein release. Through SELEX (Systematic Evolution of Ligands by Exponential Enrichment), TrAPs can be designed for any target protein in the extracellular matrix. The wide scope and biorthogonality of this project allows for many applications in medical technology and user-controlled cell fate. 

Linear electrochemical impedance spectroscopy (EIS) is widely used in the characterization of electrochemical systems, such as batteries, although the results of EIS are only as good as the scientist's model of their data, as it’s possible to fit multiple models to the same data. Nonlinear EIS (NLEIS) can also be helpful when characterizing batteries - as they are nonlinear devices - and reveal additional information, such as the asymmetry of the charge transfer between charge and discharge. Combining EIS and NLEIS results in multiple, interrelated data sets, which when fit together drastically reduces the set of models that fit the same data, providing a better understanding of battery physics. However, NLEIS is not as widely developed or used as traditional EIS methods. The goal of this research project is to further develop the use of NLEIS for battery characterization in order to combine EIS and NLEIS to ultimately provide a more accurate picture of battery health. To reach this goal, I plan to test fresh and aged lithium nickel manganese cobalt (NMC) pouch cell batteries with my group’s EIS/NLEIS model. Using materials and equipment from the Washington Clean Energy Testbeds, I will then deconstruct these batteries and fabricate coin cell batteries from the harvested electrode materials and run EIS/NLEIS experiments on these coin cells, comparing the results of the coin cells to the results of their parent pouch cells to assess the accuracy and usefulness of the NLEIS model. Advancing battery health testing is critical for the future development and use of batteries, as understanding battery health allows consumers and scientists to make sustainable decisions regarding battery use, recycling, and disposal.

Scarcity of usable water has quickly become one of the world’s greatest problems. Most of Earth’s water is saltwater, and much of the limited available freshwater contains harmful contaminants. One type of contaminant, per- and polyfluoroalkyl substances (PFAS), are particularly hazardous as they are toxic to humans and do not naturally decompose due to their strong carbon-fluorine bonds. Of the available methods of removing PFAS from water, including adsorption, ion exchange, and membrane filtration, membrane filtration is an appealing separation technology since it does not require expensive, energy intensive regeneration steps used in adsorption and ion exchange. Our project aims to use molecular layer deposition (MLD) to create polymeric thin films selective to  PFAS for water filtration. MLD involves cycles of dosing and purging reactant vapors to create a thin film layer by layer, allowing for better control over the surface uniformity, composition, and thickness. These thin films, synthesized on polyethersulfone (PES) membranes, will ideally be rejective of PFAS while preserving membrane permeability. We synthesize thin films of various chemistries and measure their water contact angle to determine the impact of hydrophilicity on long- and short-chain PFAS rejection. Here, we provide our measurements of the pure water permeability, long- and short-chain PFAS rejection, and water contact angle of MLD-treated PES membranes.

Optical metasurfaces used in nanophotonic devices are designed and optimized to display remarkable emergent photonic properties beyond what is possible for single-component materials. Traditionally, metasurfaces are designed in response to a particular incident angle of light impinging on its surface. However, in practice these metasurfaces have limited functionality if the incident angle varies. A metamaterial whose function is independent of incident angle would overcome this limitation and be more efficient in practice. For example, angle independent metamaterials that trap light in solar panels can function efficiently for all solar positions. Because a forward approach of screening many candidate materials through trial-and-error is time-consuming and expensive, in this poster we instead employ an inverse computational-based design strategy. We develop a strategy to optimize geometry/dielectric design of nanoparticles (NPs) metamaterials that have an optical response independent of angle of incidence of light. We leverage a computationally efficient and differentiable electromagnetic simulator based on couple dipole methods, the “mutual polarization method”, to perform numerical optimization of these materials. By encoding multiple incident angles and polarization states into an objective function, we ensure that the optimizer reduces the angle-variation of the metamaterials it designs. We use our inverse design tool to create multilayer plasmonic nanoparticle films, whose extinction spectra are insensitive to incident angle and polarization. We also show that we can use our inverse design method to control the spectral line shape of these NP films. Our inverse methodology will greatly accelerate the development time to synthesize new nanophotonic materials.

For decades, nanomaterials formed by protein-nanoparticle interactions have been attractive to researchers for applications that include biosensing and drug delivery. Previous work demonstrated bifunctional superfolder green fluorescent protein (sfGFP) genetically encoded with two silica-binding peptides (Car9) can induce the assembly of 10 nm silica nanoparticles (SiNP) in a pH responsive manner. The pH responsiveness arises from an intricate balance between the attraction of protein-decorated SiNP for other SiNPs, and the electrostatic repulsion of SiNP-SiNP, leading to cluster formation at pH 7.5 and dissociation at pH 8.5. In this study, we expand on the work by introducing steric forces using a monofunctional sfGFP variant chemically conjugated to polyethylene glycol (PEG). We investigate how (i) the molecular weight of the PEG extension, (ii) the molar equivalent of pegylated protein to SiNP, and (iii) the use of a mutant bifunctional protein with lower SiNP affinity, influence cluster size and polydispersity. We find that increasing steric hindrances by adding up to 5-fold molar equivalent of pegylated protein to SiNPs, or by using a longer PEG chain, leads to a progressive decrease in cluster size that is accompanied by 6-fold decrease in polydispersity to 10%. We also demonstrate that while cluster size can be controlled in the 1500-800 nm range with the wild-type bifunctional protein, its mutant version enables access to the 250-50 nm size range. We exploit the facts that sfGFP is inherently fluorescent and that our SiNPs encapsulate rhodamine to investigate how cluster size influences the Förster resonance energy transfer (FRET) efficiency between multiple donors and acceptors. We find that whereas ensemble FRET efficiency doubles as cluster size increases from 50 to 230 nm, it only increases by 15% as the assemblies grow to 1450 nm. We discuss the implication of our results for the design of environmentally responsive opto-electronic nanomaterials.

As the demand for electronics increases, so does the need for efficient recycling methods of electronic waste. The goal of electronic waste recycling is to recover critical metal components that can be used again in future electronics. However, a key challenge is selective separation of metal component mixtures into pure phases. My research in Dr. Zachary Sherman’s lab studies a promising and low-energy solution to this problem involving magnetic separation using external magnets and magnetic fields. Many precious metal ions are magnetizable in the presence of an external magnetic field, and therefore metal ion mixtures can be separated magnetophoretically by taking advantage of differences in their magnetic susceptibility. Using Brownian dynamic simulations to model transport of metal ion mixtures, I have quantified the magnetophoretic separation efficiencies of mixtures of paramagnetic, diamagnetic, and nonmagnetic ions mixtures when exposed to an external magnetic field. I have investigated how separation efficiency is affected by a variety of physical parameters including the strength of the external magnetic field, relative concentrations of ion species, strength of interactions among ions, and the magnetic susceptibilities. I also show that hydrodynamic flows generated by ion motion as well as ion structuring and aggregation have an enormous impact on separation efficiency. These results will guide further research to determine the optimal conditions for selective separation and purification of metal components.

Methane is one of the most attractive targets for controlling near-term climate change due to its short lifespan and high potency (34 times that of CO₂). Methanotrophs are bacteria that can consume methane and convert it into CO₂ and biomass. There is growing interest in using these bacteria to mitigate greenhouse gas emissions from sources such as landfills, agricultural feedlots, and abandoned coal mines. However, a key challenge is that to achieve large scale methane sequestration, as well as economic viability of deploying these in the field, we have to significantly improve the growth of methanotrophs at low concentrations of methane. Regulatory genes play an important role in determining how bacteria allocate energy. By deleting specific regulatory genes and measuring the growth rate of these mutants under low methane conditions, we can assess their importance in helping the bacteria survive and thrive in nutrient-limited environments. Using this approach, we can also replicate mutations that have naturally emerged in strains cultivated for over a year under low methane conditions. This allows us to confirm whether these mutations provide a growth advantage. By identifying and testing key genes involved in low-methane growth, we are guiding efforts to engineer a more efficient and resilient strain for real-world applications.

Bacterial metabolic engineering shows great promise for sustainable chemical production. Non-model microbes such as Pseudomonas putida, Rhodobacter sphaeroides, and Rhodopseudomonas palustris offer unique opportunities for metabolic engineering, given their tolerance to environmental stressors, their ability to grow on waste substrates, and their natural production of industrially relevant compounds. However, tools for engineering these bacteria are underdeveloped. Here we present genome engineering and gene regulation tools that are generalizable to multiple non-model microbes, offering improved versatility for metabolic engineering. Firstly, we employed a high-efficiency genome engineering tool using serine recombinases (SAGE) in R. sphaeroides and R. palustris. We evaluated integration efficiency for 10 different recombinases using a fluorescent reporter screen, revealing variation in recombinase performance across microbial hosts. We used BxbI, the top-performing recombinase, to integrate a heterologous metabolic pathway into the genome of R. palustris for the bioproduction of a biofuel precursor. In addition to genome engineering tools, we developed gene regulation tools using dCas13, a protein which regulates genes at the translational level. Genome-wide functional screens were conducted in P. putida using an inducible guide RNA system to study levels of gene regulation in native aromatic biosynthesis pathways. Overall, this work advances tools for genomic integrations and gene regulation in non-model microbes, offering new strategies for metabolic engineering and expanding the host range for synthetic biology applications. 

Global plastic production has already surpassed 300 million tons annually, which poses serious environmental challenges due to the limited recycling and effective management of plastic waste. Current mechanical recycling methods are not efficient, since only a small fraction of plastic waste goes through recycling processes, resulting in severe environmental degradation and pollution. In view of this, the use of bimetallic nanoparticles as catalysts is critically evaluated in the case of plastic recycling. In this Literature Review, I look into different metal combinations, such as ruthenium-platinum, ruthenium-nickel, and ruthenium-cobalt bimetallic catalysts. These catalysts are known to have great potential for enhancing the selectivity and efficiency of the hydrogenolysis process, hence increasing the conversion of plastic into more valuable hydrocarbons like fuels and chemicals. It intends to draw attention to various advances in chemical recycling methods that may offer sustainable solutions to the global plastic waste crisis through a critical review of the synthesis methods, catalytic mechanisms, and practical applications of these bimetallic catalysts. Further work on the unique properties of the bimetallic nanoparticles provides insight into their catalytic role in enhancing efficient C-C bond cleavage in plastic, ultimately providing higher yields of desirable products and reduced formation of unwanted byproducts.

Microbial bioproduction supports the manufacturing of sustainable chemicals but requires accurate and easy-to-use tools for monitoring cell growth. A simple and effective tool for estimating cell concentration in aqueous systems is optical density (OD). However, commercially available OD measurement systems are expensive and require manual sampling, which is time-consuming and disrupts culture growth, particularly in anaerobic microbes. To address this, I developed a low-cost OD sensor for continuously monitoring anaerobic bacteria in culture tubes. The sensor design, based on Deutzmann et al. (2022), consists of a 3D-printed sample holder with an LED and a photosensor positioned on opposite sides. The photosensor generates a voltage, which a Python script processes to calculate optical density values for each bacterial species. Plotting these OD values provides researchers with insights into bacterial growth behavior and enables optimization of culture conditions. This device's advantage over commercial spectrophotometers is that it can measure optical density directly from sealed culture tubes, eliminating the need for manual sampling into cuvettes and saving researchers valuable time. It can be configured to run autonomously, further minimizing measurement time and disruptions to bacterial growth. Additionally, the design is fully open-source and customizable while costing less than $100 to reproduce, making it accessible for a wide variety of lab setups. Overall, this low-cost, open-source OD sensor offers a practical, efficient, and customizable solution for continuous monitoring of anaerobic bacterial growth, making it a valuable tool for research laboratories.

The Lidstrom Lab aims to better understand methane-consuming microbes (also called methanotrophs) so that we can develop technologies to remove anthropogenic methane emissions, which will reduce the severity of global warming. Our research explores how the methanotroph Methylotuvimicrobium buryatense 5GB1C can be bioengineered to grow well at the low methane concentrations found in human-made emission sites, while providing value-added products like biomass from dead bacteria that can be used as animal feed. Understanding bacterial methane utilization will allow us to create effective biocatalysts at a far lower monetary and environmental cost. My research project involves deleting cytochrome genes that may be important for the 5GB1C strain to grow in low methane conditions. Manipulating these genes may allow for further improvement of growth at low methane. My targets are three genes that encode cytochromes, which are electron carriers that take electrons from particular reactions and supply them to other reactions that are otherwise energetically unfavorable. My hypothesis is that these cytochromes are involved directly in supplying 5GB1C with electrons needed for the oxidation of methane into methanol. If these cytochromes supply electrons required for methane consumption at low methane, then deleting them would generate a mutant that would grow poorly on methane because it lacks the electron carrier(s). I have generated two possible cytochrome deletion mutants and continue to work on a third cytochrome. Once the mutants that can be generated are sequenced to verify the deletions, cultures will be grown under low methane and methanol conditions to determine how their ability to grow has been affected by the knockout mutations. In this manner, our lab is building a valuable knowledgebase of genes that are suitable for manipulation to improve growth in low methane for the technologies that one day will help curtail the worsening of global warming.

Methane is an extremely potent greenhouse gas, with a warming potential 86 times greater than that of CO2 on a 20-year timescale, and is therefore a top priority for mitigation efforts to combat climate change. Methanotrophic bacteria, such as M. buryatense 5GB1C, metabolize methane as their main source of carbon and chemical energy, a trait that could help slow climate change by reducing emissions. A major obstacle is the rate at which methane consumption occurs at low methane concentrations, which tends to be too low to be appreciable. This project seeks to answer whether currently unknown genes involved in the growth of M. buryatense 5GB1C on low methane could be discovered by comparing its genome with that of a closely related methanotroph, M. alcaliphilum 20Z. While the two have very similar genomes and metabolisms, M. alcaliphilum is not able to grow at low methane concentrations (500 parts per million), while M. buryatense is. I analyzed the two genomes and isolated all genes present in M. buryatense without homologs in M. alcaliphilum. Because they are unique to M. buryatense, they may be involved in the observed growth difference. I systematically performed targeted deletion mutations on many of these candidate genes, and then tested them for growth on low methane compared to the wild type strain, looking for any defect that would suggest a gene directly essential to growth at 500ppm. I confirmed several genes to have no impact on growth at low methane, as well as one that appears to be essential to growth in any conditions, and anticipate reaching conclusions on several more mutants. These findings will help to develop microbial methane mitigation technologies that can be utilized in a great range situations and at a larger scale, essential characteristics for a global impact.

According to the CDC, there are over 42,000 female deaths from breast cancer a year in America. In particular, triple negative breast cancer is a clinical subtype that lacks estrogen, progesterone, and HER2 expression, making it more aggressive and harder to treat compared to other subtypes. There is an increased demand for targeted treatments such as immunotherapy, but little is still known about the disease’s immunological progression. Thus, we aim to integrate multiplexed imaging techniques with computational algorithms to capture immune distributions and uncover unique immune spatial architectures. We will study the immune interactions  between neutrophils and different T cell populations as they play an important role in immune signaling in the tumor microenvironment. This is important as neutrophil interactions are currently not well understood. Using a cohort of multiplexed immunofluorescence (mIF) images, we will characterize helper, cytotoxic, and memory T cells as well as neutrophils using the following biomarker panel: CD3, CD4, CD8, CD45RO, CD66b. Custom-trained CNN-based models using spectrally unmixed data for each marker is used for phenotyping with high accuracy. We annotated cells from our dataset to generate the training dataset for these phenotype classifier models. After phenotyping, we utilize spatial point pattern analyses (e.g., G-Function) to identify spatial interactions such as clustering effects between the immune cell phenotypes. We also compute patient level metrics such as the median nearest neighbor distance between pairs of phenotypes and custom-designed inter-phenotype clustering scores. Finally, we utilize Kaplan Meier analyses and log-rank test to correlate the above spatial metrics with recurrence-free survival.

Bacterial cellulose (BC) nanoparticles (BCNPs) are a promising sustainable nanomedicine platform for drug delivery and provides a scalable, eco-friendly alternative to synthetic counterparts. We aim to develop a small library of BCNPs with different chemical moieties to incorporate a broad range of active agents for drug delivery use. To produce BCNPs, a BC pellicle is grown in a kombucha media of tea, sugar, vinegar, and bacterial co-cultures. The pellicle is isolated and chemically and manually broken down using dimethylacetamide, lithium chloride, and an ultrasonicator probe to produce an organic BC dissolution. The BC dissolution is precipitated into an aqueous Pluronic F-127 (F127) surfactant solution under 650 rpm stirring conditions and incubated for 2 h to form nanoparticles ~100 nm, near neutral charge, and low polydispersity index (<0.3). In this study, we optimize the dissolution and nanoprecipitation processes using acetylated and methylated BC pellicles to form acetyl- and methyl-functionalized BCNPs. The functionalized BCNPs were characterized using Fourier transform infrared spectroscopy, nanoparticle tracking analysis, electron microscopy, and light scattering to assess physicochemical properties. Our results demonstrate that functionalized BCNPs can be formulated using similar formulation parameters to unmodified BCNPs. Ongoing work evaluates drug loading and encapsulation efficiencies in the functionalized BCNPs using curcumin as a model drug. Engineering BCNPs with different chemical moieties enables incorporation of a wider array of drugs, which can improve the utility of BCNPs as a sustainable alternative to current synthetic nanomedicines.

Hydrogels with tunable stiffnesses are a versatile method to study the interactions of human cells in vitro. These systems recreate human extracellular matrix (ECM) and capture the stiffness changes associated with a variety of biological processes and diseases, like cancer and cirrhosis. Photoresponsive chemistries allow light to be used to modulate the stiffness in these materials with high resolution. However, when creating more complex patterned gels with photomasks, bulk property analysis cannot capture the variation. To circumvent this and measure the stiffness of these complex gels, I performed rheology and fluorescence recovery after photobleaching (FRAP) to establish a correlation between diffusivity and stiffness in flood-illuminated gels. By finding and using the correlation, I am able to calculate the stiffness of the more complex patterned gels based off of their FRAP-derived diffusivity measurements. This method allows for better fine tuning of gels for use as a platform to study human cell growth through a range of stiffening events in multiple different parts of the body.

The 4-letter genetic alphabet found in Nature is the fundamental basis of biological information storage. As synthetic biologists continue to manipulate the genetic alphabet, they have begun to push against the boundaries of nature itself. Unnatural base-pairing xenonucleic acids (XNAs) are synthetic nucleotides that can pair orthogonally with the standard bases. By increasing chemical and structural diversity, XNAs are poised to enable a plethora of next-generation biotechnologies, including XNA-containing nucleic acid therapeutics (XNAptamers), catalytic nucleic acids (XNAzymes), and an expanded genetic code through a larger codon table. Although the potential of XNAs is near-limitless, the infrastructure required to study XNAs, notably sequencing, is antiquated. Previously, the Marchand Group leveraged commercial nanopore sequencing devices from Oxford Nanopore Technologies to sequence XNAs. This process outputs characteristic current signals that need to be decoded or “basecalled.” The first XNA basecallers used statistical k-mer models to decode XNA containing current signals, yet, their basecalling accuracy is a far cry from commercial basecallers (k-mer: 60-80%, commercial: >95%). Modeling our approach after commercial DNA basecallers, we have built a binary classification training pipeline that leverages long short-term memory (LSTM) neural networks and commercial nanopore sequencing to achieve more precise sequencing of XNAs. Thus far,  we have built models to effectively basecall three XNA base pairs with varying motivations: B≡Sⁿ for studying XNA replication fidelity in PCR due to high error rates, and P≡Z/Ds:Px for their unnatural functional groups (e.g. nitro groups and hydrophobicity) making them useful for applications such as XNAptamers. Currently, our binary classification models have testing accuracies as high as around 95% and we look to further improve our training methods through new model architectures such as transformers. Moving forward, we look to expand our basecaller to perform multi classification, allowing for generalized, de novo basecalling similar to commercial basecallers.

Hypoxic ischemic encephalopathy (HIE) is a neurological condition resulting from reduced blood and oxygen flow to the brain and is a leading cause of morbidity and mortality in neonates. Limited treatment options necessitate accessible and scalable interventions to improve outcomes in newborns impacted by HIE. Extracellular vesicles (EVs) have been previously shown to attenuate oxidative stress and inflammation in the brain. Further research suggests that EVs secreted by astrocytes, a brain cell type involved with the inflammatory and injury response, may elicit neurotrophic or neuroprotective properties. In this study, I isolated, characterized, and evaluated the therapeutic potential of astrocyte-derived EVs (AEVs) in an ex vivo model of hypoxic-ischemic (HI) brain injury. AEV characterization via protein assays and nanoparticle tracking analysis showed that we were able to produce AEV particles about 100 nm in size at concentrations up to 10^11 particles/mL. To assess their therapeutic efficacy, I administered AEVs at varying doses (5, 12.5, 25, and 50 µg) to neonatal rat brain slices exposed to oxygen-glucose deprivation (OGD), an ex vivo model for HI injury. Following 24h of exposure, I evaluated cell viability. Our results indicate that AEVs decrease cytotoxicity in a dose-dependent manner. To further elucidate AEVs’ mechanisms of action, we conjugate AEVs with quantum dots to track AEV localization and cell-type specific uptake in brain tissues. Understanding AEV interactions with neural cells provides insight into both the roles of AEVs and different brain cells in modulating inflammatory responses and promoting neuroprotection. By characterizing AEVs and their therapeutic potential, these findings contribute to the growing body of research on EV-based therapeutics and lay a foundation for developing reliable and scalable therapies with the potential to advance treatments for neurodevelopmental disorders and aid brain injury recovery.