News |  Sitemap |  Contact
PDF Export  | 

Parfait Evouna Mengue: Semi-Automated Cell Segmentation for Quantitative Digital Image Analysis of a Protein Translocation

Abstract

With the combination of a immunofluorescence technique (indirect antibody labeling) and the Laser Scanning Confocal Microscopy, we characterized the translocation of a calcium binding protein complex, anti Migratory-Related Protein (MRP8/MRP14), from the cytosol to the cytoplasm boundaries. It is well known that the neutrophils are among the first cells which participate to the pathogen destruction process in the innate immune response. Since these cells have phagocyte and respiratory oxygen species (ROS) generation activities. The generation of ROS is performed by the enzymatic complex NADPH-OXIDASE. This enzyme is mainly made of 2 components: one which is located to the plasma membrane and the other in the cytosol. This is the case in resting cells. Once the cells are activated by a pathogen agent, the plasma membrane surrounds it to build progressively a phagocytosis vesicle. Then the cytosolic part of the enzyme migrates to join the plasma membrane in order to form the active configuration of the enzyme responsible for the ROS generation. It has been demonstrated that the MRP8/MRP14 mediates this migration or translocation; since its non fixation to the cytosolic part may affect ROS generation. We imaged the channel DAPI for the cell nucleus detection and the DyLight488 (fluorochrome which is coupled to the secondary antibody) for specific recognition of the MRP8/MRP14. The translocation is detected by the fluorescence intensity augmentation corresponding to the MRP8/MRP14 proteins aggregation.

The visual inspection of a large amount of 2D images, even if these images have been obtained by a laser scanning confocal microscope, can be a difficult task. Since microscopy images suffer from inhomogeneous illumination of background and blurring. This semi-automation of the shapes recognition (segmentation) need to be as precise as possible to allow an accurate quantitative image analysis. Thanks to the ImageJ processing tools and the Action Bar plugin (J.Mutterer), we designed an Action Bar which allowed us to automatically achieve quantitative analysis of several confocal 2D images of a protein translocation in fixed neutrophil cells. The procedure show its robustness with images which have been acquired by the same operator either in the same session or in a different sessions. And also when these acquisitions were done by different operators.

To overcome the perturbations associated with the image acquisition, we used macros and Action Bar. We implemented the Action Bar with three levels of usability (Basic, Intermediate and Advanced User). So, the Gaussian filter (rolling ball of 2) was used for de-noising and enhancing the object edges in the converted 8-bit image. Thereafter, the edges detection was facilitated with the function “Find Edges”. This procedure was not adapted for the automatic segmentation of all the nuclei in the data set. Therefore, we use the curve fitting tool with a 4th degree polynomial equation. For the detection of the MRP8/MRP14 cytosolic distribution, we do not apply the “Find Edges” function, since it caused undesired results. We go directly through the curve fitting tool using the same polynomial equation. The polynomial equation fitting allowed us to resolve the inhomogeneous illumination of the background. Once the Cytoplasm or cytosol shape boundaries were approximated, we easily performed in each of them, the aggregations segmentation by the Reny Entropy thresholding method.

Keywords

Neutrophils; anti-Migratory-Related Protein (MRP); Translocation; Immunofluorescence; Laser Scanning Confocal Microscopy; Inhomogeneous Illumination; Curve fitting; Segmentation; Quantitative Image Analysis

Administrative data

Presenting author: Parfait Evouna Mengue
Organisation: CRP Henri Tudor, University of Luxembourg

co-authors: Patrick Pirrotte, Andreas Jahnen, Christian Moll and Sebastian Plançon

Type: Poster (portrait)

© Luxembourg Institute of Science and Technology | Legal Notice