Back

A multi-tiered microDicer with hierarchical blades achieves protein-preserving microdissection down to 10 microns

Arif, A.; Kumar, R.; Kwon, Y.; Rodriguez, R.; Cordts, S. C.; Koppaka, S.; Zhu, Y.; Pasa-Tolic, L.; Tang, S. K. Y.

2025-12-16 bioengineering
10.64898/2025.12.12.694048 bioRxiv
Show abstract

To study tissue heterogeneity at the sub-millimeter scale, laser capture microdissection (LCM) has been the leading technology for isolating regions of interest (ROI) for downstream molecular profiling. As the ROI approaches cellular dimensions ([~]10 m), laser-induced photothermal damage and challenges in capturing microtissues in conventional LCM can compromise protein preservation and quantitative fidelity. This work introduces multi-tiered Dicers, fabricated by two-photon polymerization, to mechanically dissect tissue slices into uniform microtissues down to 10 m. The hierarchical blade architecture limits instantaneous blade-tissue engagement and lowers the cutting force relative to single-tier designs. For benchmarking, proteomic analysis is performed on ethanol-fixed human squamous cell carcinoma microtissues generated by Dicers and by LCM. Under identical Nanodroplet Processing in One pot for Trace Samples (nanoPOTS) and liquid chromatography-mass spectrometry (LC-MS) conditions, Dicers yield more peptides and proteins than LCM, with the largest gains at 10-20 m spatial resolution. Confocal imaging shows catapult-associated cavities in LCM-generated microtissues. This material loss, along with membrane-limited protein extraction, likely reduces protein coverage. In contrast, multi-tiered Dicers enables reproducible microdissection down to 10 m while maintaining high protein coverage. With spatial registration of microtissues under development, Dicers have potential to complement LCM for next-generation spatial proteomic workflows.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.