Mouse reconstituted ovaries (rOvaries) are biomedical research tools used to model ovarian development in vitro. While many of the ovary’s cellular processes are reproducible in the rOvary, key differences between the in vitro and the in vivo system still exist. Two classes of follicles exist: one that activates growth immediately after assembly, and another which remains dormant in the ovary until puberty. The rOvary is thought to only produce the first class of immediately growing follicles. However, this was only shown from the perspective of the oocyte, so we aimed to gather a more structural view of follicle assembly and activation in the rOvary. Through counting the number of granulosa cells per follicle, we established that count is a reliable indicator of follicle type, in addition to granulosa cell morphology and layering. Using these indicators, we then demonstrated that rOvaries undergo a similar process of follicle formation and 1st wave activation, without establishing a pool of dormant primordial follicles. Furthermore, disorganized granulosa cell morphology during rOvary follicle formation suggests that rOvary follicles are like medulla-located, early activated first wave follicles. Visualization of FOXO3, however, revealed that while FOXO3 in early activated first wave oocytes is still imported into the nucleus, FOXO3 in rOvary oocytes is never imported into the nucleus. With the information gleaned here, we will be able to better the rOvary model and its relevance as a research and therapeutic tool.

Human pluripotent stem cells (hPSCs) offer a potentially unlimited supply of cells due to their pluripotency and self-renewal capacity, revealing a new avenue for Type 1 Diabetes therapy, as opposed to existing transplantation of increasingly scarce cadaveric human islets. Scalable production of functional human β-cells is difficult due to intensive labor and high production cost, but we can lessen these burdens by controlling proliferation. However, the molecular network regulating human β-cell proliferation remains under-explored. Here, we identify 8 novel drugs from a 4,280 small-molecule library with a high-throughput drug screening system for promoting hPSC-derived β-like cell proliferation. In EndoC-βH1 cells, a human β-cell line, we demonstrate that transcriptomes associated with cell cycle regulating genes are upregulated by each drug. To further investigate specific pathways impacting human β cell proliferation, we established EndoC-βH1 overexpression lines of drug target genes. With FACS and Incucyte live-imaging, we confirmed that our target genes increase human β-cell proliferation in vitro. Future directions include applying our findings to hPSC-derived β-like cells, examining the insulin-secreting function of drug-treated cells, determining mechanisms involved in human β-cell proliferation, and exploring potential synergistic effects of combinations of our target genes. In this project, we move towards our primary goal of establishing methods for scalable expansion of hPSC-derived β-like cells for clinical use.

Obesity is characterized by accumulation of adipose tissue, which disrupts metabolic homeostasis and contributes to the development of insulin resistance and obesity-related diseases. There are three classes of adipose tissue: white, beige, and brown, each with different functions. While white adipose tissue primarily stores energy and contributes to metabolic dysfunction in obesity, brown adipose tissue promotes energy expenditure through thermogenesis and supports metabolic health. Mitochondria play a large role in adipose tissue function by maintaining metabolic health through balancing mitophagy and mitochondrial biogenesis. We previously found that Parkin, an E3 ubiquitin ligase, regulates mitochondrial homeostasis by balancing mitophagy and Pgc1α-mediated mitochondrial biogenesis in white adipocytes. However, the role of Parkin in BAT remains unclear. Our preliminary findings show that Parkin protein levels are significantly reduced in brown adipose tissue (BAT) of high-fat diet (HFD) mice compared to normal chow (NC) mice, suggesting a diet-sensitive regulation of Parkin and a potential contribution to impaired thermogenesis in obesity. Given the central role of mitochondria in BAT thermogenesis and Parkin’s function in mitochondrial quality control, we hypothesize that Parkin has a unique role in maintaining thermogenic function in brown adipocytes. We will investigate how Parkin regulates mitochondrial function and thermogenesis in BAT to uncover potential therapeutic approaches for preventing metabolic disorders.

Caenorhabditis elegans (C. elegans) is a powerful model organism for studying embryogenesis, with its invariant cell lineage and transparency, which enable precise observation of development. To do so, traditional differential interference contrast (DIC) microscopy uses genetically modified worms that express fluorescence track cell division. This induces phototoxicity, photobleaching, and the need for genetic modifications, which hinder analysis. To address these issues, the embGAN pipeline (Waliman et al., 2024) leverages machine learning generative adversarial networks (GANs) for markerless cell-division tracking. While promising, embGAN’s current cell tracking is still inferior to that of fluorescence-based methods, highlighting the need for further optimization. This study focuses on improving the embGAN pipeline by refining the mask-making step used to train the embGAN network. Various image processing techniques, including mask dilation, erosion, and watershed segmentation, were tested to generate higher-quality training labels. Preliminary results suggest that uniform and asymmetrical dilation masks improve tracking accuracy. An optimized embGAN pipeline can leverage this, by using uniform mask dilation when analyzing cell-division in later embryo cell-divisions (past 100 seconds).