advertisement

Editors Selection IGR 18-3

Experimental Glaucoma: RNFL reflectance in glaucomatous rats

Comment by Brad Fortune on:

51030 Wavelength-dependent change of retinal nerve fiber layer reflectance in glaucomatous retinas, Huang XR; Zhou Y; Knighton RW et al., Investigative Ophthalmology and Visual Science, 2012; 53: 5869-5876

Find related abstracts

This recent study by Xiang-Run Huang and her colleagues extends a series of excellent papers on the alterations observed within the retinal nerve fiber layer (RNFL) of rats with experimental glaucoma. After varied exposure to increased intraocular pressure, retinal explants from Wistar rats were examined using a custom multispectral microreflectometer. This enabled the investigators to measure reflectance of RNFL bundles in relatively narrow spectral bands (or bins) from 400 to 830 nm. The investigators demonstrate that in control retinas, RNFL reflectance is wavelength-dependent, being strongest for the shortest wavelengths (nearest the 'blue' end of the spectrum) and weakest in the near-infrared portion of the spectrum. This is consistent with theoretical predictions based on the geometry of thin cylinder reflectivity, work done previously by this group and the well-known clinical practice of using 'red-free' filters to optimize visualization of the RNFL (during clinical exam and photography). Retinas were then stained histologically with markers for three major constituents of the RNFL axonal cytoskeleton: F-actin, beta-tubulin and neurofilament and grouped according to the degree of cytoskeletal disruption.

The results for glaucomatous retinas demonstrated clearly that RNFL reflectance becomes abnormal at an early stage judged by the histological degree of cytoskeletal disruption. The reflectance abnormality was characterized by a 'flattening' of the spectraldependence such that the relatively greater reflectance normally observed for shorter wavelengths in control retinas was reduced in glaucomatous retinas (as compared with the middle or 'green' portion of the spectrum from 500-560 nm). This abnormality was present even in RNFL bundles that were judged to have a 'normal-looking' axonal cytockeleton by histology. However, this 'early-stage' finding was limited to the measurement locations nearest the optic disc. The pattern of abnormal RNFL bundle reflectance became more pronounced (i.e., the RNFL reflectance spectrum became progressively flatter and extended to greater distances from the optic disc) as histological signs of cytoskeletal disruption became more severe.

Axonal cytoskeletal disruption precedes axon thinning. Moreover, this early phase of axonal reflectance changes may represent a reversibly stage of injury

Application of theoretical modeling provided additional compelling evidence to suggest that loss of the thin cylinder reflectance source (having dimensions of cytoskeletal elements) occurs prior to loss of a thicker cylinder mechanism (having dimensions closer to typical axon diameters). This finding is consistent with other cited evidence suggesting axonal cytoskeletal disruption precedes axon thinning. Moreover, this early phase of axonal reflectance changes may represent a reversibly stage of injury. The authors also point out that their findings might explain some of the discrepancies within the clinical literature. Other studies have found that RNFL defects apparent on red-free photography are occasionally missed by OCT imaging, which is done most commonly using an infrared source. If early-stage cytoskeletal disruption is causing decreased scatter (reflection) that is most apparent under red-free (short wavelength) illumination, it stands to reason (given the evidence presented by Huang and colleagues) that imaging with an infrared source may not be the most sensitive mode for detecting abnormal RNFL optical properties which derive from its unique ultrastructural characteristics. To this end, these investigators have also recently developed¹ the capability to image the rat retina in vivo using dual-wavelength OCT (415 and 808 nm) in order to compare depth resolved reflectance within the RNFL at different wavelengths (Zhang X, Hu J, Knighton RW, Huang X-R, Puliafito CA, Jiao S. Dual-band spectral-domain optical coherence tomography for in vivo imaging the spectral contrasts of the retinal nerve fiber layer. Opt Exp 2011;19:19653-19659).

One disappointing aspect of this paper was that the authors never presented any data for absolute reflectance (only reflectance normalized to the 'middle' wavelengths examined). It might have been interesting to know whether absolute RNFL reflectance was reduced in addition to the spectral flattening that was well documented. Our own (unpublished) findings in a non-human primate model of experimental glaucoma suggest that absolute reflectance (per unit thickness of RNFL) does not change much (or at all) for the longer end of the spectrum using SDOCT (870 ± 50 nm). It would also have been interesting to know whether retinal ganglion cell and/or orbital optic nerve axon counts were normal at the earliest stage of RNFL reflectance abnormalities documented in this study.

• ← Previous
• Next →