Empirical Analysis and Computational Modeling of Cellular Pattern Formation in the Growing Retina.

Abstract

Abstract: : Purpose: (1) Using a cellular ablation experiment, identify the dominant mechanism of cellular pattern formation in the growing teleost retina: spatial control of cell–fate decisions or lateral migration of differentiated cells. (2) Evaluate a computational model of cellular pattern formation. Methods: Eyes of adult goldfish (Carassius auratus) received injections of BrdU (control) or BrdU + pargyline + 6–hydroxydopamine (6–OHDA; experimental). 6 hr to 71 d after injection, BrdU– and tyrosine–hydroxylase (TH)–positive cells were labeled in retinal whole mounts, and the two–dimensional (2D) patterns of TH cells were analyzed quantitatively (nearest–neighbor, density recovery, and quadrat analyses). The TUNEL method was used to screen for cells undergoing apoptosis. A computational model of cellular pattern formation in the growing retina (Cameron & Carney, J Comp Neurol, in press), based upon a cell–fate decision mechanism, was tested for its ability to predict the growth and characteristics of 2D patterns of TH cells in control and experimental retina. Results: 2D patterns of TH cells in control retinas were non–random and characterized by local anti–clustering. An absence of TH cells 2 d after 6–OHDA exposure confirmed their ablation; selectivity of cellular ablation was suggested by the 2D patterns of apoptotic cells. Post–injection growth was confirmed by the presence of a BrdU–positive, annular band of cells displaced centrally from the retinal margin. The absence of TH cells central to the BrdU–positive band 71 d after 6–OHDA exposure indicated a lack of lateral migration by differentiated TH cells born after the 6–OHDA exposure; the 2D patterns of TH cells immediately peripheral to the BrdU band were non–random, but apparently at elevated density. The computational model, utilizing a homotypic inhibitory signaling mechanism to control cell–fate decisions, successfully predicted quantitative aspects of 2D patterns of TH cells in control and experimental retinas, including an elevated density of TH cells subsequent to ablation of TH cells. Conclusion: Formation of 2D patterns of TH–positive cells in the growing goldfish retina is dominated by mechanisms that control the spatial pattern of cell fate decisions, and apparently not by lateral migration of differentiated cells. A computational model of retinal growth utilizing a ‘one step’ homotypic inhibitory signaling mechanism, arising from differentiated cells, to control cell fate decisions can recapitulate quantitative aspects of 2D cellular patterns in the growing retina.

Keywords: retinal development • computational modeling • cell–cell communication