Histologic and morphometric procedures followed standard techniques. Left eyes were collected for sectioning, and right eyes were collected for RPE flatmounts. Left eyes were fixed in fixation solution (97% methanol, VWR, Cat. #BDH20291GLP; 3% acetic acid, Cat. #Fisher BP2401-500) at −80 °C for 4 days, embedded in paraffin, and sectioned through the sagittal plane on a microtome at thickness of 5 µm, with minor variation of the freeze-substitution method of Mary-Sinclair et al.28 (link) Sections were used for hematoxylin and eosin staining, as described previously.29 (link) All sections were imaged with a bright-field microscope with a 20× objective.
ONL nuclei were counted in a semiautomated fashion using QuPath (University of Edinburgh, Division of Pathology, Edinburgh, Scotland; https://qupath.github.io/)30 (link) to outline the regions of the nuclei and then identify and count them within 100-µm-wide segments. The entire retina was divided into parts radially from the ONH, with five parts in the superior direction and five parts in the inferior direction. Nuclei counts were done inside a 100 µm box centered in each of the 10 parts. The first 100 µm box on each side of the ONH was placed 250 µm from the ONH. This resulted in ONL nuclei counts in ten 100 µm regions of the retina for each mouse. Mean ONL counts from three to five mice per group were plotted as “spidergrams.”