TY - JOUR
T1 - Deep Learning-Aided Time-Discrete Bioprinting for Single-Cell Encapsulation in Microgel With Composable Stiffness Gradients
AU - Zhao, Yanfeng
AU - Hou, Yaozhen
AU - Zheng, Zhiqiang
AU - Dong, Xinyi
AU - Shi, Qing
AU - Huang, Qiang
AU - Fukuda, Toshio
AU - Wang, Huaping
N1 - Publisher Copyright:
© 1996-2012 IEEE.
PY - 2026
Y1 - 2026
N2 - Single-cell microgels, engineered to replicate complex 3-D in vivo niches, have shown tremendous potential for advancing biomedical research. Studies have demonstrated that microgels recapitulating tissues’ stiffness facilitate the observation of cellular behavior in vitro. However, encapsulating single cells in microgels with composable stiffness gradients to understand cellular responses at microscale interfaces remains challenging. Here, we propose a versatile fabrication method that integrates an enhanced vision model with time-discrete bioprinting to encapsulate single cells in microgels featuring composable stiffness. An enhanced deep learning algorithm was developed to localize single cells in microscopic images, integrating traditional image processing algorithms for improved precision. By combining single-cell positioning data with a time-discrete array based on a digital micromirror device, dynamic digital masks can be generated to control each micromirror in real time. Using these masks, MDA-MB-231 cells were encapsulated in Gelatin methacrylate microgels with tunable stiffness ranging from 1.42 to 11.42 kPa. Experimental results indicated that when the microgel stiffness gradient exceeded 2 kPa, MDA-MB-231 cells exhibited a significant response, extending toward the softer regions of the microgels. These findings illustrate that our approach provides a valuable biofabrication strategy for investigating single-cell behaviors in microgels, showing great potential for applications in 3-D single-cell research.
AB - Single-cell microgels, engineered to replicate complex 3-D in vivo niches, have shown tremendous potential for advancing biomedical research. Studies have demonstrated that microgels recapitulating tissues’ stiffness facilitate the observation of cellular behavior in vitro. However, encapsulating single cells in microgels with composable stiffness gradients to understand cellular responses at microscale interfaces remains challenging. Here, we propose a versatile fabrication method that integrates an enhanced vision model with time-discrete bioprinting to encapsulate single cells in microgels featuring composable stiffness. An enhanced deep learning algorithm was developed to localize single cells in microscopic images, integrating traditional image processing algorithms for improved precision. By combining single-cell positioning data with a time-discrete array based on a digital micromirror device, dynamic digital masks can be generated to control each micromirror in real time. Using these masks, MDA-MB-231 cells were encapsulated in Gelatin methacrylate microgels with tunable stiffness ranging from 1.42 to 11.42 kPa. Experimental results indicated that when the microgel stiffness gradient exceeded 2 kPa, MDA-MB-231 cells exhibited a significant response, extending toward the softer regions of the microgels. These findings illustrate that our approach provides a valuable biofabrication strategy for investigating single-cell behaviors in microgels, showing great potential for applications in 3-D single-cell research.
KW - Biofabrication
KW - single-cell heterogeneous encapsulation
KW - time-discrete control
KW - vision-based cell localization
UR - https://www.scopus.com/pages/publications/105030693136
U2 - 10.1109/TMECH.2026.3656633
DO - 10.1109/TMECH.2026.3656633
M3 - Article
AN - SCOPUS:105030693136
SN - 1083-4435
JO - IEEE/ASME Transactions on Mechatronics
JF - IEEE/ASME Transactions on Mechatronics
ER -