Abstract
Purpose :
The neonatal mouse retina has become a prominent model for studying angiogenesis. However most quantitative assessments of anti-angiogenic therapeutics inhibiting vascular growth, sprouting and maturation are performed on randomly selected fields of microscopic view that introduce a high risk of biased measurement. We aimed to develop a semi-automated image analysis solution for annotation and segmentation of flat-mounted retinal scans, enabling computation of large specimen numbers and differentiation between applied anti-angiogenic modalities.
Methods :
Neonatal mice were treated i.p. at day of birth, P2 and P4 with anti-angiogenic compounds influencing vascularization and pericyte (PC) attachment. Pups were sacrificed on P5. Retinal flatmounts, stained for endothelial cells (EC) and PCs, were imaged with an automated scanner (Olympus) and subjected to quantitative image analysis employing an algorithm for vessel growth and pericyte attachment (Definiens XD). Quantification of the superficial vascular plexus and EC coverage by PCs was analyzed by segmentation and analysis of the digitized data per gray-scale analysis followed by channel co-localization. Parameters to differentiate anti-angiogenic compounds were radial expansion of vessel growth, % vascular plexus area, branching, endothelial tip cells, vascular front tortuosity, and % PC-covered vessels.
Results :
Analyses were performed on wing-based computation and data visualized using Tibco Spotfire. Results significantly differentiated the effects of single receptor antagonists as well as their monomeric combinations and revealed comparable and statistically relevant regression on vascular growth and maturation. Quantification of PCs attached to annotated retinal arterioles vs venules vs capillaries revealed clear differences between the vascular beds. However a statistical compound effect was only detectable on the venules.
Conclusions :
We report generation of an image analysis solution that allows semi-automatic quantification and differentiation of anti-angiogenic therapeutics in a mouse model of angiogenesis. This unbiased approach delivers significant and reproducible results over a series of studies and provides a promising alternative to manual procedures.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.