A Semi-Automatic Method for the Quantification of Astrocyte Number and Branching in Bulk Immunohistochemistry Images

Autores da FMUP

• Susana Isabel Ferreira Da Silva De Sá

  Autor

Participantes de fora da FMUP

• Marques, SI
• Carmo, H
• Carvalho, F
• Silva, JP

Unidades de investigação

• Biomedicina

• Laboratório Associado RISE- Rede de Investigação em Saúde

  

  Investigador

  Fernando Carlos De Landér Schmitt

• RISE-Health

  

  Investigador

  Fernando Carlos De Landér Schmitt

Abstract

Immunohistochemical staining of cell and molecular targets in brain samples is a powerful tool that can provide valuable information on neurological mechanisms. However, post-processing of photomicrographs acquired after 3,3 '-Diaminobenzidine (DAB) staining is particularly challenging due to the complexity associated with the size, samples number, analyzed targets, image quality, and even the subjectivity inherent to the analysis by different users. Conventionally, this analysis relies on the manual quantification of distinct parameters (e.g., the number and size of cells and the number and length of cell branching) in a large set of images. These represent extremely time-consuming and complex tasks, defaulting the processing of high amounts of information. Here we describe an improved semi-automatic method to quantify glial fibrillary acidic protein (GFAP)-labelled astrocytes in immunohistochemistry images of rat brains, at magnifications as low as 20x. This method is a straightforward adaptation of the Young & Morrison method, using ImageJ's plugin Skeletonize, coupled with intuitive data processing in datasheet-based software. It allows swifter and more efficient post-processing of brain tissue samples, regarding astrocyte size and number quantification, the total area occupied, as well as astrocyte branching and branch length (indicators of astrocyte activation), thus contributing to better understand the possible inflammatory response developed by astrocytes.

Keywords

• glial fibrillary acidic protein (GFAP); skeletonize; image analysis; image post-processing