Abstract
Purpose :
Clinical management of glaucoma includes assessment of the RNFL by various means including quantitative measurement of RNFL thickness (RNFLT) by OCT. However, previous studies indicate that OCT measurements of RNFLT underestimate the magnitude of axon loss throughout a wide range of glaucomatous damage.1-3 Here we use transmission electron microscopy (TEM) to evaluate the RNFL in a nonhuman primate (NHP) model of experimental glaucoma (EG) in order to refine interpretation of clinical imaging results.
Methods :
Longitudinal measurements of RNFLT were obtained by OCT in both eyes of 3 NHP (ages 15, 13 and 9) during baseline and after induction of EG.1-3 After perfusion fixation and enucleation, a 2-mm trephine was used to biopsy a specimen of retina from the inferotemporal edge of each ONH, which then underwent processing for TEM. Four sets of TEM images, each set spaced by two mesh grid squares, were recorded from an ultra-thin section near the center of each specimen. Axon density for each image was determined by manual counting (11-17 images per specimen); EG vs control eye densities were compared by non-parametric t-test.
Results :
At the final in vivo imaging session, loss of inferotemporal sector RNFLT (EG/Fellow Control) was 26%, 26% and 58% in these 3 NHP. Axon density measured by TEM in the same 3 EG eyes was reduced relative to fellow control eyes by 38% (p=0.02), 23% (p=0.03) and 68% (p<0.0001). Consistent with their reduced density measured by TEM, axons within the RNFL of EG eyes were generally larger (see Fig, note scale difference for NHP1) and commonly exhibited ultrastructural signs of degenerative damage, including loss or distortion of the cytoskeleton and other organelles. Automated image analysis for quantification of axon size and ultrastructural features is ongoing.
Conclusions :
NHP experimental glaucoma led to a reduction of axon density within sectors of the RNFL that had undergone progressive thinning as measured in vivo by OCT. The reduced axon density observed here by TEM is consistent with our previous reports that RNFL thickness measurements underestimate axon loss.1-3 Similarly, the presence of damaged axons with disrupted cytoskeletal ultrastructure is consistent with polarization-sensitive imaging being more sensitive than OCT to detect early-stage damage of the RNFL.2
1IOVS 2012;53:7766-73; 2IOVS 2015;56:3936-44; 3IOVS 2016;57:OCT404-12.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.