Introduction

Despite common use, 2D cell culture fails to model in-vivo environments thereby providing inaccurate data on pathophysiology. Two-D cell culture analyses are also commonly performed in bulk, further convoluting information about diverse cell populations such as those found in tumors. Single cell proteomics tools allow for analysis of protein expression in hundreds to thousands of single cells at once but have yet to accessibly incorporate 3D models. Biomimicking single cell proteomics tools are necessary to better study complex diseases and develop improved therapies.

Methods

Here, we present a method for the fabrication of bone model microspheres encapsulating Ewing sarcoma (ES) cells followed by incorporation of spheres with SIFTER (Vlassakis, Nat. Comm., 2021), a cutting-edge single-cell proteomics assay. Collagen 1 (COL1) and hydroxyapatite (HA) were used to model bone and introduce 3D cell culture to SIFTER. COL1 (5 or 10 mg/mL) supplemented with cells (0.5x10^6 cells/mL) and/or HA (70:30 Col1:HA mass ratio) was allowed to gel at 37°C for mechanical analysis (cell free). Alternatively, the gel with cells was processed under agitation in oil supplemented with Span80 to create a microemulsion, thereby forming microspheres for incorporation with SIFTER. Sphere morphology and encapsulation efficiency following Col1 emulsion were analyzed in ImageJ.

Results

Interestingly, we observed no significant difference between gel formulations (p >0.05 for all pairs, Tukey’s HSD). 5 mg/mL Col1 without HA had an average CS = 1661 Pa, (coefficient of variation (CV) = 136%) while 10 mg/mL gels without HA had an average CS = 724 Pa (CV = 45.2%), 5 mg/mL gels with HA had an average CS = 649 Pa (CV = 48.5%) and 10 mg/mL gels with HA had an average CS = 697 Pa (CV = 33.8%). These findings were inconsistent with our hypotheses that higher concentration gels and HA would result in stiffer gels and these findings may be a result of sample preparation or mechanical testing discrepancies. These findings will be validated with larger sample sizes in the future.

With Col1 microemulsion, we can generate more than 3x10^6 spheres in ~1 hour from adding 1 mL of Col1 to Span80-oil for emulsion. These spheres have an overall single cell encapsulation efficiency of 8% (n = 16,000 with 3 replicate batches included). We observe that empty spheres (no cells encapsulated, diameter (d) = 69.6 µm, CV = 34.6%) and spheres with one cell encapsulated (d = 73.1 µm, CV = 33.9%) have significantly smaller (p < 0.05) average diameters than spheres with 2 or more cells encapsulated (d = 109 µm CV = 32.7%). While microfluidic cell encapsulation methods result in more uniformly sized sphere populations, our method results in similar single cell encapsulation efficiency with minimal preparation time and resources. The generated spheres have also been observed to be easily incorporated with SIFTER and do not hinder the single cell proteomics capabilities of the device.

Conclusion

Here, we report on the development of tissue engineered models of the bone microenvironment as well as an accessible method for single cell encapsulation within that model. Our method is capable of encapsulating single cells in micro-scale, in-vivo-representative tissue environments while minimizing barriers to 3D cell culture accessibility including cost, need for specialized equipment or microfluidic technique experience. Future work involves investigating mechanics of spheres made with different concentrations of Col1 and HA, single-cell cytoskeletal morphology and adhesion profiles within spheres and comparing these profiles with ES patient tumor samples to investigate which model formulation is most in-vivo-representative. Information gained by utilizing this model with single cell proteomics tools in the future can be used to guide targeted therapeutic approach developments and ultimately lead to improved clinical outcomes for ES and other complex disease states.