Biological tissues have complex, three-dimensional (3D) organizations of cells and matrix factors that provide the architecture necessary to meet morphogenic and functional demands. Disordered cell alignment is associated with congenital heart disease, cardiomyopathy, and neurodegenerative diseases and repairing or replacing these tissues using engineered constructs may improve regenerative capacity. However, optimizing cell alignment within engineered tissues requires quantitative 3D data on cell orientations and both efficient and validated processing algorithms. We developed an automated method to measure local 3D orientations based on structure tensor analysis and incorporated an adaptive subregion size to account for multiple scales. Our method calculates the statistical concentration parameter, κ, to quantify alignment, as well as the traditional orientational order parameter. We validated our method using synthetic images and accurately measured principal axis and concentration. We then applied our method to confocal stacks of cleared, whole-mount engineered cardiac tissues generated from human-induced pluripotent stem cells or embryonic chick cardiac cells and quantified cardiomyocyte alignment. We found significant differences in alignment based on cellular composition and tissue geometry. These results from our synthetic images and confocal data demonstrate the efficiency and accuracy of our method to measure alignment in 3D tissues.

A single right retina from a black rhinoceros was whole mounted, stained and analyzed to determine the visual resolution of the rhinoceros, an animal with reputedly poor eyesight. A range of small (15-μm diameter) to large (100-μm diameter) ganglion cell types was seen across the retina. We observed two regions of high density of retinal ganglion cells at either end of a long, but thin, horizontal streak. The temporal specialization, which receives light from the anterior visual field, exhibited a ganglion cell density of approximately 2000/mm2, while the nasal specialization exhibited a density of approximately 1500/mm2. The retina exhibited a ganglion cell density bias toward the upper half, especially so, the upper temporal quadrant, indicating that the rhinoceros would be processing visual information from the visual field below the anterior horizon for the most part. Our calculations indicate that the rhinoceros has a visual resolution of 6 cycles/degree. While this resolution is one-tenth that of humans (60 cycles/deg) and less than that of the domestic cat (9 cycles/deg), it is comparable to that of the rabbit (6 cycles/deg), and exceeds that seen in a variety of other mammals including seals, dolphins, microbats, and rats. Thus, the reputation of the rhinoceros as a myopic, weakly visual animal is not supported by our observations of the retina. We calculate that the black rhinoceros could readily distinguish a 30 cm wide human at a distance of around 200 m given the appropriate visual background.

This study presents a morphological and quantitative analysis of displaced ganglion cells performed in the chick retina at different stages of development (E11, E17, & P2). The lipophilic dye DiI inserted into the optic nerve provided a complete staining of displaced ganglion cell population and of their dendritic trees. From 11 days of incubation (E11) all neurons present a uniform morphology and are unmistakably recognizable by the dimensions of their cell body and by the presence of the axon. Their dendritic arborizations shape a complex network through the retina. Cell bodies are randomly distributed but dendritic trees of neighboring cells constantly present two levels of dendritic overlap: about 50% of the cells present a high degree of overlap, more than the 30% up to 95% of their total area, while the other half overlaps less than 20%. The high degree of overlap identifies a cluster, which is regularly formed by three and more rarely by four neurons starting from 17 days of incubation (E17) onward and after birth. From E17 onward, 94% of clusters is constantly formed by three cells and 6% of the clusters by four neurons. The distribution and the spatial arrangement of clusters are very regular, forming a mosaic that gives rise to a precise retinal coverage. The nearest-neighbor analysis unequivocally demonstrated that the mosaic of clusters is the most regular array so far described. This exact arrangement starts from 11 days of incubation and is maintained and refined through the following stages of development and after birth.