Purchase this article with an account.
Leandro B. Teixeira, Owen Bowie, T M. Nork, Richard R. Dubielzig, Gillian J. McLellan; Optic nerve axon loss in a Cat Model of Inherited Glaucoma : validation of a semi-automated targeted sampling method. Invest. Ophthalmol. Vis. Sci. 2012;53(14):2492.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To validate a semi-automated method for quantifying axons in the optic nerves of cats with inherited congenital glaucoma
Previous non-invasive IOP and structural and functional data (including SD-OCT and electrophysiology respectively) were available for all cats included in the study. Based on these data, single optic nerves were selected from cats that demonstrated mild to severe glaucoma (n= 9) or were normal (n=6). Following transcardial perfusion fixation with 4% paraformaldehyde, eyes were enucleated and 2mm-long samples of optic nerve were dissected 2mm posterior to the globe. Optic nerve samples were fixed in glutaraldehyde, osmicated and resin embedded. Semi-thin (1µm) sections stained with 1% p-phenylenediamine were evaluated by light microscopy. Axons were quantified in one section per nerve using both a full-count method and a semi-automated targeted sampling (SATS) method. To evaluate inter-observer reproducibility, 3 masked observers obtained axon counts on the same samples by the SATS method. Full counts were obtained from up to 40 non-overlapping photomicrographs encompassing the entire nerve cross-section. Targeted counts were obtained using commercially available image-analysis software (cellSens Dimension®, Olympus). Up to 3 distinct, relatively homogeneous regions were identified in low power photomicrographs. An automated counting function was used to calculate the average axon density, from up to 5 randomly selected high power fields for each region; adjusting the adaptive threshold as needed to recognize and highlight all axons, with subsequent manual exclusion of non-neural elements. An estimate of the total number of axons in each cross-section was thus obtained.
Median total axon count (full count method) was significantly lower in the glaucomatous cats (39020) than in the normal cats (62713; p=0.0016, Mann-Whitney). Correlation between the semi-automated counts and total axon counts was strong (linear regression slope = 1.10, y-intercept =564 and r2=0.96) with acceptable inter-observer reproducibility.
Our study confirmed mild to severe optic neuropathy in this feline model of congenital glaucoma. Use of a SATS method provides a practical, rapid and relatively reliable means of estimating axon loss in cats with glaucoma but may over- or under-estimate axon numbers in normal cats and those with only very mild glaucomatous damage. Evaluation of the relationships between optic nerve damage and measurements of optic nerve structure and function obtained in vivo in this spontaneous animal model of glaucoma is ongoing.
This PDF is available to Subscribers Only