The Effects of Limiting Citrate-Derived Acetyl-CoA Synthesis on the Development of Exhaustion in CD8+ T Cells

Overview

Tcell exhaustion is a hypofunctional fate observed in cancer that causes loss of full mitochondrial function. Preventing Tcell exhaustion could sustain the immune system’s anti-tumor response. An accumulation of lipid droplets in exhausted Tcells has been observed, raising the question of how lipid synthesis might be related to the progression of Tcells towards exhaustion. I investigated the impact of limiting intracellular Acetyl-CoA production on the development of Tcell exhaustion by targeting ATP-citrate lyase (ACL Y), which converts citrate to Acetyl-CoA. CD8+ Tcells from C57Bl/6 mice were activated for 24 hours through the TCR in vitro. Cells either received no additional stimulation (acute stimulation - AS) or continuous TCR stimulation for 6 days (chronic stimulation - CS). Within each AS/CS group, cells were treated with ACL Yinhibitor concentrations of 0μM, 10μM, 20μM, or 40μM per subgroup. Cells were analyzed by flow cytometry for markers of exhaustion. In both AS and CS conditions, a greater number of cytokines were produced by cells treated with 40µM ACL Yi than cells treated with lower concentrations. CS cells were most impacted: 5.39% of cells without ACL Yi were IFNγhi TNF-αhi, compared to 51.0% with 40µM ACL Yi. No notable difference in PD-1 or TIM-3 expression was observed in any cell group. These data suggest that ACL Yinhibition preserves Tcell effector function under chronic stimulation conditions. The optimal dosage of ACL Yi may lie around 30µM. Future steps include repeat testing under hypoxia and in vivo experimentation with mice tumors. Optimizing Word2Vec Nishka Kacheria Interlake High School, Bellevue, WA From the video game Semantle to ChatGPT, Natural Language Processing (NLP) systems, such as GloVe (Global Vectors for Word Representation) and Word2Vec (transforming Words to Vectors) are necessary in communicating with humans, as well as answering questions and classifying texts. Previous research has used these systems to predict genomes or stocks. The wide array of applications provided motivation for research into reducing the runtime and memory usage of GloVe by reducing the dataset size through dimensionality reduction techniques. I investigated how using Principal Component Analysis could reduce the dimensions of GloVe, and tested the preservation of data in the vector dataset by using the Birch Clustering algorithm to find semantically similar words. I determined that almost all data was stored in half the number of dimensions, with even smaller reductions. I also identified bias in the GloVe words database stemming from the training dataset from viewing the data on lower dimensions.

Competition history

• JSHS 2023 Category not listed

Resources

• National JSHS abstract book ↗

Related projects

Closest projects by meaning, across every fair and year in the corpus.