Prostate cancer is a leading cause of morbidity, with prostate specific membrane antigen (PSMA) as the primary molecular target for imaging and radioligand therapy. Despite clinical success, a subset of tumors exhibits low PSMA expression, resulting in false negative PET scans and resistance to PSMA targeted interventions. This research profiles the molecular landscape of PSMA low phenotypes to identify alternative cell surface targets within the human surfaceome. Using a clinical cohort of patients who underwent preoperative PSMA PET imaging and radical prostatectomy, we integrated imaging derived SUVmax values with matched bulk RNA sequencing data. Differential gene expression analysis was performed to isolate transcripts upregulated in tumors with low PSMA uptake. These candidates were filtered through curated surfaceome databases to prioritize proteins with high therapeutic accessibility. Our analysis identified promising cell surface candidates, including specific G protein coupled receptors and structural proteins, enriched in the PSMA low population. These results suggest that PSMA loss involves distinct biological reprogramming or lineage plasticity. These alternative markers provide a biological foundation for developing the next generation of theranostic agents, ensuring that patients with PSMA invisible disease can still benefit from precision oncology.

Despite the success of antiretroviral therapy (ART) in controlling HIV, ART fails to clear the infection as it fails to target HIV latent reservoirs. Chimeric Antigen Receptor (CAR) Therapy is a promising approach, engineering immune cells to enhance their HIV reactivity. Specifically, CAR-modified HSPCs, which differentiate into lineages such as T cells, macrophages, and NK cells in humanized mice and nonhuman primates, show potential for broader immune responses against HIV. We aim to optimize anti-HIV multilineage CAR immune cells by selecting intracellular domains that enhance various immune cells. By developing CAR constructs incorporating D1D2-CAR with macrophage intracellular domains, FcRγ and PILR-beta, and NK signaling, DAP12, we aim to enhance multilineage CAR immune cell efficacy. Moreover, we incorporate CD40 and OX40, which mediate NK cytotoxicity with CD40L, engaging a prolonged CAR-T response alongside DAP12.​ In-vitro data shows DAP12 enhances CAR-NK cytokine secretion. FcRγ and PILR-beta in CAR-MQ showed comparable phagocytosis to the original D1D2-41BB-CD3ζ CAR. Notably, combining FcRγ with OX40 and PILR-beta with DAP12 significantly improves CAR-T proliferation and cytotoxicity. These constructs also enhanced CAR-T memory in co-cultures with target cells. We selected three promising CAR constructs to progress to in-vivo studies, in which we will evaluate for anti-HIV properties in HIV-infected humanized mice to boost CAR immune cells and endogenous cytotoxic T-cells while targeting anatomical reservoirs.

"Retinitis Pigmentosa (RP) is a group of genetic degenerative eye disorders characterized by the progressive loss of rod photoreceptors, then cones, resulting in blindness. Normally, rods and cones signal to their rod and cone bipolar cell counterparts. However, it has been shown in Rd10 mice that when rods die, their bipolar cells rewire, connecting to cones to maintain signaling and relay information to the brain. I sought to investigate this rewiring in the whole eye by measuring changes in ex vivo electroretinogram (ERG) and the ratio of the isolated b-wave (from ON bipolar cells) and a-wave (from photoreceptors) as a function of time during degeneration. In my research, I recorded pharmacologically isolated a-waves and b-waves from whole RD10 mouse retinas at various ages (P21, P28, P35, and P42). However, I found that the b and a wave ratio remained unchanged throughout degeneration, similar to the ratio for the photopic ERG in Gnat1-/- mice, suggesting that despite photoreceptor loss and synaptic changes, the retina still maintained information from the outer to inner retina. This preservation of information is due in part to the bipolar cells making contacts with cone photoreceptors. Investigating this process could show how universal this rewiring is, leading to progress for enhancing this rewiring response. "

Magnetic resonance imaging (MRI) is a powerful tool for diagnosing and managing liver conditions due to its high soft tissue contrast and lack of ionizing radiation. In particular, MRI provides unique strengths for guiding minimally invasive procedures, such as microwave ablation of liver cancers. However, it is challenging to achieve accurate and automated low latency tracking of both interventional devices and tissue targets on real-time MRI for procedural guidance. To address this challenge, this project investigates the performance of the SAM 2++ foundation model for tracking the motion of liver tissue structures and interventional needles on 2D real-time MR image series. The foundation model is assessed for motion tracking across multiple granularities, including bounding boxes and key points, and its performance is compared to manually annotated reference data. Using this multi-granularity tracking information, we analyze the nonrigid liver motion by comparing displacement from key points around the liver with the average motion represented by the bounding box. This allows us to identify the liver regions which experience the greatest amount of displacement due to respiratory motion. Ongoing work focuses on expanding the test dataset and fine-tuning the model on MRI data to improve accuracy. Overall, this project demonstrates the potential of using foundation models to enhance MRI-guided interventions through real-time motion tracking.

Pathologic vascular remodeling is characterized by alterations in extracellular matrix composition and smooth muscle cell behavior. Previous literature has implicated the SLIT3 glycoprotein as a regulator of tissue remodeling and collagen biology, yet the role of SLIT3 remains insufficiently explored in vascular mural cell biology. This project investigates SLIT3 regulation of extracellular matrix (ECM)-related gene expression in human coronary artery smooth muscle cells (HCASMC) as a first step toward understanding its connection to arterial remodeling and aneurysm physiology. HCASMC treatments using positive control cytokines such as AngII and TGFB1, along with siRNA knockdown of SLIT3, were performed. HCASMCs were cultured and analyzed by RNA extraction, reverse transcription for cDNA synthesis, and RT-qPCR for a variety of ECM markers, including collagen, fibrillin, fibronectin, and elastin. Results showed a roughly 2x increase in expression of ECM markers after positive-control cytokine treatment and a 50% decrease after siRNA knockdown of SLIT3. This suggested a dose-sensitive response in ECM gene expression, but further optimization of the experimental pipeline is necessary to minimize variability across replicates. These findings provide a basis for future analyses of SLIT3 perturbations in HCASMCs, including dose-response experiments and assays of cell contractility.