Poster T8 - Tunable Tissue-Engineered Model for Single-Cell 3D Culture and Proteomics Investigations of Ewing Sarcoma

Introduction: For decades, metastatic Ewing Sarcoma (ES) patients have faced sub-20% survival rates despite attempts to improve therapeutics (Grohar, Pharmacol. Ther., 2013). Heterogeneous ES cell populations complicate treatment, with cells expressing high levels of the oncoprotein EWS-FLI1 being proliferative while migratory, EWS-FLI1-low cells establish secondary tumors (Franzetti, Oncogene, 2017). EWS-FLI1 is known to affect gene transcription and cell phenotype. However, the full impact of EWS-FLI1 on the proteomic landscape of cells is not understood, resulting in knowledge gaps surrounding potential therapeutic targets. On top of complex pathology, use of 2D cell culture and bulk analysis of ES cells in research further hinders the discovery of more effective treatments. 2D culture fails to model in-vivo-like cell behaviors (Jensen, Int. J. Mol. Sci., 2021). This has led to ES therapies showing promise in vitro but being ineffective in vivo. Bulk analysis of cell culture fails to detect cell phenotypes that are differentially impacted by therapeutics.

Here, we developed and optimized the encapsulation of cells in a 3D bone model, which is compatible with a single-cell assay known as SIFTER (Vlassakis, Hensen, Nat. Comm., 2021). SIFTER allows for on-chip protein complex isolation, fractionation, and quantification from thousands of single cells simultaneously. This work shows the accessibility and tunability of our cell encapsulation process and the compatibility of a 3D model with SIFTER for the first time ever. We anticipate this technology, which combines cutting-edge in-vivo mimicry and single-cell analysis, has the potential to provide unique and clinically relevant insight into deadly ES pathophysiology.

Materials and

Methods: The bone model is made of commercially available collagen 1 (Col1) gel supplemented with hydroxyapatite (HA) through in-situ mineralization (Takallu, Biomed. Mat. Res., 2019) before neutralization on ice. Once a pH of 7.4 is reached, 0.5x10^6 GFP-expressing ES TC71 cells/mL Col1 are added into the still sol-state collagen, which is then added to preheated oil containing 0.5-3% (v/v) Span80 in a conical tube. The tube is then agitated with a conical tube spinner at a fixed speed (10-80 RPM) to create an emulsion for 5 minutes, after which the collagen has gelled. The cells in bone model spheres are then separated from the oil with 40 µm cell strainers and imaged on a fluorescent microscope to analyze sphere size and cell encapsulation efficiency. Images were analyzed with a custom image analysis program in MATLAB. The volume ratio of collagen added to oil, Span80 concentration, and spin speed were varied to investigate the impact of these parameters on sphere diameter and single-cell encapsulation efficiency.

Finally, the compatibility of the spheres with the SIFTER assay was investigated. Cell-encapsulating spheres were settled in 100 µm microwells stamped in a thin polyacrylamide gel (PAG) mounted on a microscope slide. A lysis-buffer-soaked PAG lid was then applied on top of the gel slide, and GFP-expressing cell lysis within the spheres was monitored under a fluorescent microscope. When cell lysis had occurred, a 30 V/cm electric field was applied to the gel slide, and migration of GFP out of the cell-encapsulating spheres was monitored and imaged.

Results, Conclusions, and Discussions: To determine which sphere-generation protocol parameters impact sphere size and single-cell frequency most significantly, we employed a design of experiments that systematically altered the parameters. Agitation speed and Span80 concentration significantly impacted the sphere diameter and frequency of single-cell spheres (standard least squares, p < 0.01). For example, 10 RPM agitation speed in 0.5% Span80 oil resulted in ~7% of spheres encapsulating a single cell and an average sphere diameter of 199 ± 55 μm (Fig. 1A). Comparatively, 80 RPM agitation speed in 3% Span80 resulted in ~16% of spheres encapsulating a single cell and an average diameter of 112 ± 45 μm (Fig. 1B). We also observed the inclusion of the 3D bone model did not hinder on-chip cell lysis through the delivery of a buffer with a PAG lid or electrophoretic extraction of proteins from cells in microwells (Fig. 1C), the essential steps of performing the SIFTER assay.

Here, we report on the development of a bone micromodel and the accessible method for single-cell encapsulation within that model. Our method minimizes barriers to 3D cell culture accessibility, including cost and the need for specialized equipment or microfluidic set-ups. On-chip cell lysis within spheres occurred in 90 seconds, which is twice as long as the original SIFTER protocol but still in a short amount of time, allowing for the high-throughput nature of the assay to be maintained.

Future work involves investigating single-cell morphology within spheres when cultured and comparing these profiles with ES mouse tumor samples to investigate how well the spheres model an in-vivo environment for cancer cells. Additionally, the ability to perform SIFTER through the bone model will be rigorously investigated and compared to unencapsulated-cell results to confirm the bone model does not interfere with SIFTER read-outs/ capabilities. The information gained by utilizing this model with single-cell proteomics can guide the development of targeted therapeutic approaches and improve clinical outcomes. We also anticipate this model can be easily adapted to model other disease states, making the fabrication method described here widely applicable to other researchers and allowing for the accessible and rapid 3D modeling of cells.

Acknowledgements (Optional): Data was collected by Jocelyn Baker with assistance from undergraduate researcher Mel Zhou. Cell lines were provided by the Texas Children’s Hospital in collaboration with Dr. Jason Yustein, M.D., Ph.D.. Rice Nanofabrication Cleanroom and Shared Equipment Authority facilities and systems were used heavily for SIFTER device fabrication/validation throughout data collection.