Following nanoparticle uptake by cells using high-throughput microscopy and the deep-learning based cell identification algorithm Cellpose

Abstract

Read online

How many nanoparticles are taken up by human cells is a key question for many applications, both within medicine and safety. While many methods have been developed and applied to this question, microscopy-based methods present some unique advantages. However, the laborious nature of microscopy, in particular the consequent image analysis, remains a bottleneck. Automated image analysis has been pursued to remedy this situation, but offers its own challenges. Here we tested the recently developed deep-learning based cell identification algorithm Cellpose on fluorescence microscopy images of HeLa cells. We found that the algorithm performed very well, and hence developed a workflow that allowed us to acquire, and analyse, thousands of cells in a relatively modest amount of time, without sacrificing cell identification accuracy. We subsequently tested the workflow on images of cells exposed to fluorescently-labelled polystyrene nanoparticles. This dataset was then used to study the relationship between cell size and nanoparticle uptake, a subject where high-throughput microscopy is of particular utility.

Keywords

• nanoparticles
• cell uptake
• fluorescence microscopy
• high-throughput
• machine learning
• image segmentation