3D reconstruction of granulomas from transmitted light images implemented for long-time microscope applications.

Image analysis tools are essential to describe and quantify dynamic biological phenomena, such as early stages of granuloma formation. Granulomas are constituted of a collection of immune cells that contain pathogens, leading to their elimination. We presented here a new method to obtain granuloma 3D reconstruction from transmitted light images. Granulomas were generated by incubating peripheral blood mononuclear cells with beads coated with sonicated Coxiella burnetii, a bacterial pathogen. Biological samples were observed under a confocal microscope, and recorded during several hours, providing a large set of data of several gigabytes. Our image processing, called Focus Detection Plugin (FDP), allowed to extract relevant images from large datasets and to perform a deblurring of image stacks. This FDP method, that was implemented as an ImageJ plugin, did not require powerful computer resources and was simple to use. To validate our FDP method, we compared our results with 3D reconstruction of fluorescent images. Both methods yielded comparable results. We concluded that our FDP method was able to generate processed images yielding robust 3D reconstruction of whole cell bodies, and presented major advantages for long-time recordings since no cell labeling was needed. This method was convenient to study the early stages of granuloma formation and may be applied to other complex biological systems. 2010 Elsevier B.V. All rights reserved.