One of the most common types of vaccines used today are subunit vaccines. Subunit vaccines consist of an antigen that triggers the adaptive immune system to create antibodies but also require a separately added adjuvant, which is a substance that induces longer-term immunity by stimulating the immune system to pay attention to the antigen. Current adjuvants are non-specific - often things like oil-water emulsions that irritate the immune system and cause inflammation in unknown ways to draw attention to the antigen. This project aims to create a more specific adjuvant by directly stimulating B cells. In order for B cells to replicate antibodies, they need a primary signal from the antigen and a secondary signal that certain ligands on T-cells can initiate. We decided to investigate whether CD40Ligand (CD40L), an immune protein present on T cells that works to signal B cells to either replicate or create antibodies, could be used to achieve this goal. The idea behind the project is to co-display CD40L with antigen on the nanoparticle in hopes of creating a more specific adjuvant. We designed 10 different versions of this nanoparticle, where we tested two versions of CD40L, the placement of CD40L, and the linker length between CD40L and the nanoparticle surface. Our preliminary results also show that our cages retain their ability to bind both antibodies and CD40 as well as activate NFkB transcription - a proxy for B cell activation. We expect CD40L-displaying nanoparticles will promote B-cell proliferation to a greater extent than the nanoparticle vaccine displaying only hemagglutinin (flu) antigen accompanied with an adjuvant like Addavax. Ultimately, we hope to examine how co-display of CD40L with antigen will change the quality of immune response and memory in-vivo in comparison to currently used vaccine adjuvants, and begin testing in-vivo in the coming quarters.

Traditional vaccines use inactivated or live attenuated pathogens to elicit an effective adaptive immune response, but these vaccines can lack safety for immunocompromised individuals. Subunit vaccines–which display characteristic components of pathogens–are safe, stable, and readily engineered, but struggle to elicit a strong immune response. These next-generation vaccines require adjuvants (substances that stimulate the immune system) to increase efficacy. However, many currently used adjuvants lack well-understood mechanisms or wide applicability across vaccines. There is a need for new adjuvant platforms, and protein-based adjuvants are appealing because they are stable, readily engineered, and can be co-delivered with antigens on subunit vaccines. Toll-like Receptor (TLR) proteins are promising adjuvant targets that bind pathogen-associated molecules to activate the innate immune system. Of this family, TLR5 binds the bacterial protein flagellin, but flagellin is not a suitable adjuvant candidate because it is degradation- and aggregation-prone. Therefore, this project aims to design, test, and characterize novel protein-based adjuvants that can bind TLR5 and activate the immune system. Here, I test and characterize de novo mini-proteins that I have computationally designed to bind mouse TLR5 (mTLR5). I use cell-based binding and signaling assays to identify and characterize successful binders. Preliminary results show de novo mini-proteins bind and activate mTLR5 and key knockout mutations at the binder interface decrease signaling. Ultimately, this work hopes to provide a mouse model for these novel protein-based vaccine adjuvants with clinical aims. This project has wide-reaching public health implications, as vaccines offer the potential to improve the health and lives of countless individuals.

Anxiety and depression are becoming more prevalent and are an increasing public health concern. People with certain personalities are particularly vulnerable to developing anxiety and depression. Personality is commonly measured using the Big Five Personality test, which groups personality traits into five scales that a person can score high or low on. One of these traits is Neuroticism (N), which measures people’s propensity towards anxiety, depression, and self-consciousness, as well as the degree to which people find the world distressing and threatening. Emotion reactivity (ER) is a facet of N and falls under the former half of this definition. Past research has shown that those with high N and ER are at risk for developing anxiety and depression. One possible way to reduce this risk is through studying how people regulate their emotions. While past research has already shown that certain regulating strategies reduce the risk of anxiety and depression, my project is examining their efficacy for people with high ER. Specifically, my research is predicting anxiety and depression symptoms from participants’ levels of neuroticism with emotion regulation use as a moderator. I am measuring ER with the Emotion Reactivity Scale (ERS), emotion regulation with the Cognitive Emotion Regulation Questionnaire (CERQ), depression with the Patient-Reported Outcomes Measurement Information System (PROMIS) depression scale, and anxiety with the PROMIS anxiety scale. Data was collected from a baseline study on students at the University of Washington (n=268). Due to past research on different emotion regulation strategies, I predict that participants with high emotion reactivity who use maladaptive strategies will score higher on the anxiety and depression scales. In contrast, people who score high on the emotion reactivity scale and use adaptive strategies will score lower on the anxiety and depression scales. Broadly, this study hopes to further our understanding of potential risk and protective factors related to internalizing disorders in at risk populations. If results are consistent with my predictions, the analysis could suggest possible means of providing relief for people with anxiety and depressive symptoms.

Protein nanoparticles are useful for the design of novel vaccines. We can use these nanomaterials for display of antigens; however, antigens tested thus far have been homooligomers—consisting of a single unique component, and existing protein nanoparticle assemblies are not well-suited for the display of heterooligomeric antigens (such as HCV E1E2). We attempt to solve this problem through designing uniformly symmetric icosahedral nanoparticles that contain two distinct protein chains within their fundamental geometric component—referred to as the asymmetric unit; this allows us to retain the particle symmetry and double the number of accessible linkage points, or chain termini. This doubling of termini could allow us to fuse heterodimeric antigens to our cages. To accomplish this design goal, I utilized RFdiffusion—a generative machine learning model—to generate two-component icosahedral protein backbones, which were then filtered by evaluating subunit packing through a set of contact distance calculations. ProteinMPNN, a DL-based sequence design method, was used to assign candidate sequences to each of the filtered backbones. Finally, complete designs were filtered by using AlphaFold2 to evaluate fidelity to the original design model. I expressed the top 96 designs in E. coli, but saw minimal protein with no indication of assembly. In an attempt to maximize my chances of forming successful cages, I have elected to conditionally generate backbones that favor alpha-helical secondary structure. In this new design round, I hope to see favorable improvements in inter-chain packing; this will lead to an increase in passing design candidates and hopefully allow my computationally generated structures to have a higher chance of assembly in lab. This work serves to streamline the development of a therapeutic platform that can display multi-component antigens, which could enable the creation of new vaccines.

Computational protein design has successfully designed nanoparticle cages that self-assemble and effectively deliver encapsulated therapeutics to cells. These nanoparticle cages are readily taken up by the cell via receptor-mediated endocytosis. Despite the promise of these cages, one of the greatest challenges that remain is the successful endosomal escape of the encapsulated biologics and their precise delivery to the cytosol. To address this, we have engineered a high throughput complementation assay, based on split green fluorescent protein (GFP) construct, that helps screen and quantify cytoplasmic delivery of therapeutics through fluorescence intensity. Split-GFP is a protein complementation assay in which the normally monomeric GFP is made of two fragments: the larger non-fluorescent beta barrel and a 15 amino acid (a.a) peptide. When these two components unite, the GFP fluoresces. In this project, I created a stable HeLa cell line expressing the beta barrel of split GFP using lentiviral transduction under antibiotic selection. The cell line has been further validated, through transient transfection of the complementary 15 a.a peptide to test the assay performance. I propose to test endosomal escape, through introduction of endolytic peptides (EEPs) into model proteins, which force early endosomal membrane fusion and destabilization. Future research will explore adapted designs of nanoparticle cages, incorporating the EEPs and the split-GFP complementary strand in the HeLa cell line, to quantify the endosomal escape of our designs. The outlook of this project has transformative implications for targeted therapeutic delivery. By creating a screening assay that can quantify targeted delivery into cytosol, we can expedite refinement of protein designs for therapeutic delivery, thus accelerating the timeline for developing novel protein-based therapeutics.