In Vivo Imaging and Quantification of Retinal Gliosis in Transgenic Mice

We propose an in vivo retinal imaging platform for quantifying gliosis in response to neurotoxicity of certain compounds. Gliosis, in the form of retinal GFAP-GFP expression, was non-invasively and longitudinally imaged using a confocal scanning laser ophthalmoscope (SLO) and its progression was monitored over a period of two weeks. However, the retinal images have very poor signal-to-noise ratio (SNR) and cellular resolution such that the analysis of GFAP-GFP expressing retinal cells from these images can be a very challenging task. We report an image averaging method based on a pixel rank matching criterion which significantly enhances both these image attributes. We then compute retinal gliosis by quantifying the fluorescence intensity (FI) around the optic disc at different time points. We demonstrated that the kainic acid (KA)-induced gliosis can be non-invasively detected starting on day 3, and peaked at day 7 (p les 0.05). The proposed method provides great potential for longitudinal studies on various retinopathies and for therapeutic development in a pre-clinical setting.