Investigating the role of TGFB1-Fibronectin crosstalk in Epithelial-Mesenchymal Transition using in silico and in vitro spheroid models.

Introduction: Epithelial-mesenchymal transition (EMT) is a transdifferentiation process in which epithelial cells lose key markers like strong cell adhesions and apicobasal polarity, and increase expression of mesenchymal characteristics like increased motility. Progression of EMT plays a major role in fibrotic diseases. The cytokine Transforming Growth Factor-β1 (TGF-β1) is a known regulator of EMT. This transition has been previously shown to rely on crosstalk between TGF-β1 and the ECM protein fibronectin (FN), where we see a correlation between upregulated TGF-β1 activity and FN fibril assembly, forming a positive feedback loop that promotes EMT. However, the overall development and maintenance of TGF-β1-FN crosstalk is still not fully understood.
This study seeks to analyze TGFB1-FN crosstalk through a 3D multicellular agent-based model of TGF-β1-induced EMT. We have incorporated the bistable switch model developed by Tian et al. (2013) into a 3D multicellular agent-based model of EMT, incorporating cell-cell interactions and ECM remodeling that affects TGF-β1 diffusion and tethering. Validation of the model has been done using an in vitro assay of 3D epithelial spheroids, mimicking the same parameters defined in the model. We hypothesize that increased FN deposition by epithelial cells in response to TGF-β1 promotes spheroid depolarization and forms a localized reservoir of TGF-β1 and other profibrotic signaling factors that drives EMT. Furthermore, we hypothesize that integration of FN-TGF-β1 tethering into the computational model will improve the ability of the 3D model to accurately predict spatiotemporal EMT signaling.

Materials and

Methods: We developed a 3D multicellular agent-based model that simulates TGF-β1 signaling that incorporates the bistable switch model into each cell. We define EMT states based on the concentration of the mesenchymal marker N-Cadherin. ECM remodeling is captured through changes in the TGF-β1 diffusion coefficient, which affects the ability of TGF-β1 to diffuse throughout the tissue. Decreased diffusion coefficient values lead to decreased TGF-β1 diffusion, which we use to represent growth factor tethering to the ECM. Unknown parameters of TGF-β1 signaling were estimated using Monte Carlo optimization and compared to published TGF-β1-induced EMT longitudinal datasets.
Validation of the model was done using in vitro experiments. Madin-Darby Canine Kidney subclone II (MDCKII) epithelial cells were embedded in a Collagen I hydrogel and cultured into spheroids via single-cell spheroid formation. Fibrogenesis was induced using 4 ng/mL TGF-β1 over 2, 4, and 6 days (known time points where epithelial cells exhibit different EMT states). We characterized changes in spheroid function and morphology via confocal imaging (LSM 980, Zeiss), observing changes in polarity, TGF-β1 signaling, and ECM reorganization. qRT-PCR was used to quantify changes in gene expression of EMT markers defined in the computational model. We compared in vitro experimental results to model simulations to validate outputs.

Results, Conclusions, and Discussions: Results and Discussion. Initial tests of the model were done by comparing outputs to the original bistable switch model. We ran simulations using different TGF-β1 concentrations known to induce epithelial, partial state, or mesenchymal phenotypes. Results show that the model captures the biphasic transition in response to different TGF-β1 concentrations over different time points. Integration of ECM remodeling simulates TGF-β1 tethering to the surrounding matrix, improving simulations of localized TGF-β1 signaling. Comparisons of simulated results to our in vitro spheroid experiments demonstrated that we were able to capture morphological changes in response to TGF-β1 stimuli, where we saw increased spheroid size but decreased cell count in response to increased exogenous TGF-β1.

Conclusions. The development of a combined in silico and in vitro approach provides a novel research tool to study cell-ECM interactions in a 3D environment, with possible future application in therapeutic studies to identify novel targets and individual responses of different cell types to improve current treatments of fibrotic diseases.

Acknowledgements (Optional): This research was supported through the IGNITE KUH Training Program (NIH U2C/ TL1 DK132771-1) and NSF Award CBET-2302580.