| Popis: |
Introduction: It is recognized that screening of drugs on 2D models is unable to precisely select clinically active medicaments; therefore, 3D culture systems are emerging and show potential for better simulating the in vivo tumour microenvironment and for eliminating the species differences to allow drug testing directly in human systems before drugs move into clinical trials. The purpose of this study was to automate the production and cultivation of human primary osteogenic sarcoma cells line, SaOS-2, in scaffold-based, multiple spheroids in GrowDex® (GDS) and scaffold-free systems, single spheroids (SS) for high-throughput screening. Methods: For scaffold-based models, SAOS-2 were embedded in nanofibrillar cellulose hydrogel, GrowDex®, in flat ultra-low attachment (ULA) 96-well plates while for the scaffold-free system cells were let to form spheroids in U-bottom ULA 96-well plates. Experiments were conducted on the Fluent® 780 automation workstation. The scripts for automated model production and media exchange were established in the FluentControlTM. Technical parameters such as aspiration and dispensing speed as well as XYZ dispensing positions were empirically defined and optimized i) to allow production of stable models, ii) to avoid disturbing during the medium exchange neither the GDS nor the GrowDex structure/shape, iii) to avoid SS displacement and aspiration. GDS and SS models were cultivated up to 11 days and were characterized for their viability (ATP assay), morphology and size (MTT assay). Dose-response tests with taxol and doxorubicin were carried out for both system types. Drugs were automated dispensed on the 3D models at day 4 for a treatment time of 72h. Cells were analysed for cell viability, measuring ATP levels, morphology, and size. Results: The established scripts allowed the formation of multiple spheroids in GrowDex as well as the aggregation of SAOS-2 in single spheroids in the U-bottom ULA plates. 70% of medium could be successfully automated exchanged maintaining unaltered the shape and the position of the models in the wells. GDS and SS remained stable for up to 11 days and increased in size over time, showing a similar growth rate. Viable GDS populated the entire model at different Z values with a compact morphology, parameter that characterized the SS too. GDS showed, as expected, a wide size distribution while SS were bigger and more homogenous in size in comparison to GDS (SS area 10-fold bigger than GDS area). SAOS-2 responded in both systems to taxol and doxorubicin, showing higher IC50 values for GDS compared to SS. Taxol (fig.1) and doxorubicin were 3.5- and 4.5-fold more potent in SS than in GDS. Collapsed morphology with a viable and compact core and a loose cell layer around border at high drugs concentrations characterized exclusively SS. Discussion & Conclusions: The automation protocols were successfully established allowing the reproducible production and maintenance of GrowDex multiple spheroids and scaffold-free models. Although its viscosity, GrowDex is automation compatible, and the results obtained in this project shows its high potential for high-throughput drug screening. Acknowledgements: The ZHAW authors would like to thank the Tecan for the technical support and the consumables, and the UPM Biomedicals for its contribution. |
