Plasma membrane (PM) cholesterol abundance plays a critical role in functional T cell receptor (TCR) clustering and activation. The Aster family of proteins dynamically regulates PM cholesterol organization by facilitating intracellular cholesterol transport from the PM to the endoplasmic reticulum. We previously found that knocking out Aster in T helper cells led to cholesterol accumulation in the PM, clustering of TCRs, and increased TCR signaling and immune cell activation. We hypothesized this pathway could be harnessed in the context of chimeric antigen receptor (CAR)-T cells. Using a CRISPR-engineered CD19-CAR-expressing T cell line, I found that Aster-deficient CAR-T cells had higher levels of PM cholesterol and produced more Interleukin-2 in response to the CD19 target compared to Aster-sufficient CAR-T cells. These observations suggested a correlation between PM cholesterol and the activity of the Aster-deficient CAR-T cells. Additionally, compared to Aster-sufficient CAR-T cells, activation of Aster-deficient CAR-T cells by CD19 induced greater phosphorylation of Tyrosine 505 of LCK, suggesting that altering PM cholesterol via Aster can tune proximal CAR signaling. I will harness the genetics of Aster and our established cell line system to test the hypothesis that PM cholesterol impacts CAR clustering. I will further examine whether the deletion of Aster enhances the killing efficiency of CAR-T cells against target tumor cells. This research highlights a potential route to increase CAR-T cell effectiveness.

Hypoplastic Left Heart Syndrome (HLHS) is a form of congenital heart disease in which left-sided heart structures are underdeveloped. While current fetal echocardiography and diagnostic tools are reliable for prenatal identification, predicting which fetuses and subsequently neonates will develop critical illness after birth remains a major clinical challenge. Our team has developed a novel noninvasive computational pipeline that reconstructs the patient-specific fetal aorta and integrates computational fluid dynamics based on 2D fetal echocardiography imaging and data to simulate hemodynamics and blood flow characteristics. We aim to improve prenatal risk stratification through patient-specific hemodynamic modeling using computational fluid dynamics in controls and HLHS fetuses during pregnancy. Our data demonstrates that prenatal modeling utilizing the computational pipeline can enhance early risk assessment, guide clinical decision-making, and improve postnatal outcomes. We anticipate this will reassess the current standard of care, ultimately improving long-term postnatal outcomes for this vulnerable population with critical congenital heart disease.

Facial feminization surgery (FFS) is a cornerstone of gender-affirming care with significant psychosocial benefits, yet postoperative recovery varies and clinicians lack tools to provide individualized predictions for pain, recovery, and quality-of-life outcomes. We evaluated whether baseline clinical and psychosocial variables could be used in artificial intelligence (AI) models to predict postoperative outcomes following FFS and translated these models into an interactive dashboard for point-of-care visualization. We retrospectively analyzed 232 patients (mean age 34.1±10.1 years) who underwent FFS. Preoperative predictors included age, operating room time, baseline PROMIS scores, and mental health diagnoses. Outcomes included PROMIS scores at 3–6 months, inpatient pain, length of stay, and perioperative medication use normalized by body weight. Linear regression, LASSO, Random Forest, Support Vector Regression (SVR), and XGBoost were implemented with grid search cross-validation. Performance was evaluated using MAE, RMSE, and accuracy defined as predictions within 15% of observed values, and feature importance was assessed using SHAP. SVR performed best overall, achieving the highest accuracies across psychosocial outcomes, followed by XGBoost, while linear models performed comparably for select outcomes. Psychosocial outcomes were more predictable than perioperative outcomes. Baseline depression, anxiety, OR time, and mental health history were the most influential predictors. The Streamlit dashboard demonstrates how AI

Pulmonary arterial hypertension (PAH) is characterized by elevated pulmonary vascular resistance, driving right ventricular (RV) remodelling, dysfunction, and failure. Despite its clinical impact, the mechanisms driving RV dysfunction remain unclear, with no approved therapies. We hypothesize that lysyl oxidase-like 2 (LOXL2) upregulation promotes RV fibrosis via trans-differentiation of fibroblasts to myofibroblasts, and we evaluated the therapeutic potential of LOXL2 inhibitor, epigallocatechin gallate (EGCG). Adult male rats received Monocrotaline (MCT) or SU5416 under hypoxic conditions (SuHx) to induce PAH, confirmed by echocardiography and RV catheterization. LOXL2 expression was validated by qPCR, immunofluorescence, and Western blot. EGCG or vehicle (methylcellulose) was administered by oral gavage (days 21-30). Fibrosis was assessed by trichrome and immunohistochemistry staining. In vitro, human cardiac fibroblasts were cultured in TGF-β1 and EGCG to assess myofibroblast formation. RV transcriptomic analysis from MCT and SuHx models revealed dysregulation of pathways involving extracellular matrix remodelling and showed upregulation of LOXL2. EGCG treatment improved RV hemodynamics and attenuated fibrosis; RNA sequencing showed a reversal of genes involved in mesenchymal transition and inflammation. In vitro, EGCG dose-dependently inhibited TGF-β1-induced myofibroblast formation. These findings reveal LOXL2 as a key driver of RV failure and demonstrate EGCG’s therapeutic potential for PAH-induced RV dysfunction.

Kidney stone disease is a urologic condition that affects 1 in 10 people globally. Although kidney stone disease is widespread, research on cellular interactions between calcium oxalate (CaOx) crystals and renal epithelial cells is limited. Our study aims to assess the interactions between CaOx stones, biofilm-forming bacteria, and renal epithelial cells (RPTEC/TERT1 cell line) using an in vitro microfluidic model that replicates renal physiological conditions. RNA sequencing analysis was used to characterize the genetic pathways that facilitate CaOx crystal nucleation and deposition in our model. Renal epithelial cells underwent treatment with (1) growth media, artificial urine, and CaOx, (2) growth media, artificial urine, CaOx, and biofilm-forming bacteria, and were compared to (1) cells with growth media alone, and (2) cells with growth media and artificial urine. RNA sequencing analysis found an upregulation in gene pathways related to DNA repair and cellular replication. More specifically, the FEN1 gene, associated with DNA repair, was downregulated with an average fold change of -0.7 and a P-value of 9.56 x10^-11. Downregulation of the FEN1 gene results in an upregulation of inflammatory pathways. Ongoing evaluation may provide potential avenues for both the prevention and treatment of kidney stone disease.