AI-based Label-Free Imaging of Live Cells
Regenerative medicine. Immunotherapy. Tissue repair. Mesenchymal stromal cells (MSCs) hold high promise in numerous novel clinical treatments. But! Unregulated inconsistencies in the MSC phenotype negatively impact treatment outcomes. Here, we develop a machine learning model to identify MSC extracellular marker expression and localization in phase contrast images of real time, live cell cultures. Our artificial intelligence (AI)-based method converts transmitted light microscopy images of MSCs into quantitative measurements of protein expression levels. This approach allows us to perform quantitative, non-invasive, single-cell, and multi-marker characterizations of heterogeneous live MSC culture which is critical for high-throughput drug screening and quality control in cell therapies. We hope that this technique can be developed into the next gold standard as an accurate, efficient, lower-cost, real-time, and non-invasive method of evaluating cell cultures.
Publications: Investigating heterogeneities of live mesenchymal stromal cells using AI-based label-free imaging. Sci Rep 11, 6728 (2021).
Effects of Microenvironment on Epithelial Phenotype
Epithelial cells line organ surfaces throughout the body and perform vital physiological functions including adsorption of solutes and ions, secretion of biochemical factors, barrier protection against foreign bodies, and sensory reception. They are also the source of a wide variety of diseases from skin disorders to polycystic disease to 90% of cancer types. Due to their wide influence, it is important to fully understand epithelial cell behavior. However, previous works have analyzed cell phenotype of single cells or colonies, which is not reflected in in vivo biological systems. Our lab is interested the individual and combined effects of cell-substrate and cell-cell adhesions on fully confluent epithelium. We aim to reveal a more physiologically relevant model of study. We engineer substrates of varying stiffness to map the phenotypical contribution from substrate stiffness. Additionally, we use a variety of mechanisms to inhibit or degrade intercellular adhesion proteins to modulate cell-cell interactions.
Mechanical Heterogeneties in Epithelial Monolayers
Current approaches for measuring the mechanical properties of tissues are time intensive, invasive, or can only achieve a tissue level resolution, overlooking mechanical heterogeneities which can serve as disease biomarkers. In this project, cell-stretching instrumentation is developed and used with an inference algorithm to resolve heterogeneities in modulus at a cellular resolution. A piezo motor-driven cell stretcher device stretches a live epithelium cultured on an ultra-thin and soft freestanding PDMS membrane while DIC images are captured in real time. Images are processed and used as an input into a neural network that can directly output its corresponding heterogeneous modulus field.
Laminar Flow Culture of 3D Tissues
3D spheroids and organoids are valuable for modeling tumor and tissue microenvironments. 3D tissues incorporate both cell-cell and cell-matrix interactions, which give rise to in vivo-like heterogeneous cell structures and gene expression profiles. In addition, 3D tissues may be derived from human-cell line embryonic stem cells or induced pluripotent stem cells, allowing us to compliment or even replace the need for animal models.
While the benefits of 3D tissues are numerous, they still do not recapitulate the maturity and complexity we see in vivo, which led us to investigate novel cell culture methods using flow culture. We use a novel Microwell Flow Device (MFD) to rotate 3D tissues in 96 well plates to generate superficial laminar shear stress. This stress, in turn, increases diffusion to deeper cell layers and alters cell phenotype and behavior. We apply the MFD to study different growth mechanisms and response to therapeutic drugs in multiple human tissue and tumor types.