Relative Course of Retinal Nerve Fiber Layer Birefringence and Thickness and Retinal Function Changes after Optic Nerve Transection

Abstract

purpose. To test the hypothesis that alterations of RNFL birefringence precede changes in RNFL thickness in an experimental model of RGC injury and, secondarily, to determine the time course of RGC functional abnormalities relative to RNFL birefringence and thickness changes.

methods. RNFL birefringence was measured by scanning laser polarimetry (GDx VCC; Carl Zeiss Meditec, Inc., Dublin, CA). RNFL thickness was measured by spectral domain optical coherence tomography (SD-OCT, Spectralis HRA+OCT; Heidelberg Engineering, GmbH, Heidelberg, Germany). Retinal function was assessed by three forms of electroretinography (ERG): slow-sequence multifocal (mf)ERG (VERIS; EDI, San Mateo, CA); pattern-reversal (P)ERG (Utas-E3000; LKC Technologies, Inc. Gaithersburg, MD); and photopic full-field flash (ff)ERG (Utas-E3000; LKC Technologies). All measurements were obtained in both eyes of four adult rhesus macaque monkeys (Macaca mulatta) during two baseline sessions, and again 1 week and 2 weeks after unilateral optic nerve transection (ONT).

results. ONT was successfully completed in three subjects. RNFL birefringence declined by 15% 1 week after ONT (P = 0.043), whereas there was no significant change in RNFL thickness (+1%, P = 0.42). Two weeks after ONT, RNFL retardance had declined by 39% (P = 0.018), whereas RNFL thickness had declined by only 15% (P = 0.025). RGC functional abnormalities were present 1 week after ONT, including decreased amplitudes relative to baseline of the mfERG high-frequency components (−65%, P = 0.018), the PERG N95 component (−70%, P = 0.007), and the photopic negative response of the ffERG (−44%, P = 0.005).

conclusions. RNFL birefringence declined before and faster than RNFL thickness after ONT. RGC functional abnormalities were present 1 week after ONT, when RNFL thickness had not yet begun to change. RNFL birefringence changes after acute RGC injury are associated with RGC dysfunction. Together, they reflect RGC abnormalities that precede axonal caliber changes and loss.

Previous studies have shown that cytoskeletal components and their tertiary structure within retinal ganglion cell (RGC) axons cause the retinal nerve fiber layer (RNFL) to exhibit the optical property of form birefringence. ^{1 2 3} This idea is supported both by theoretical analyses ^{1 2} and by evidence demonstrating that RNFL birefringence declines rapidly after chemical disruption of cytoskeletal components, microtubules (MT) in particular, in situ ³ or in vivo. ⁴ This finding, in turn, has clinical relevance because cytoskeletal abnormalities may develop in diseases such as glaucoma before the death of RGCs. For example, in experimental models of RGC injury such as optic nerve transection (ONT) or crush, there is a delay among most surviving RGCs before axonal caliber begins to decline, ⁵ which is preceded by changes in cytoskeletal protein content and mRNA. ^{6 7} Abnormalities of cytoskeletal proteins such as neurofilament (NF) have also been demonstrated in experimental models of glaucoma ^{8 9 10} and may represent a mechanism of susceptibility. ¹¹ Thus, it is possible that measurement of RNFL birefringence can be used to detect early-stage cellular dysfunction and/or to predict subsequent risk of progression and permanent loss. ^{3 12}  

RNFL birefringence can be measured clinically using scanning laser polarimetry (SLP) ^{12 13 14} or polarization sensitive optical coherence tomography. ^{15 16 17 18} Using SLP, Mohammadi et al. ¹⁹ found that measures of RNFL birefringence are an independent predictor of future vision loss in patients with suspected glaucoma who began the study with normal SAP visual fields, regardless of their age, IOP, or optic disc appearance. Although this is consistent with the hypothesis that cytoskeletal abnormalities were present before subsequent progressive changes in optic nerve structure or function, the technique of SLP measures retardance, which is a function of both RNFL birefringence and RNFL thickness. ^{3 12 13 17 18 20 21} Thus, it is possible that, in Mohammadi et al., ¹⁹ the RNFL thickness changes were present at baseline and predictive of future progression of glaucoma. ²²  

The primary purpose of the present study was to test the hypothesis that alterations of RNFL birefringence precede changes in RNFL thickness in an experimental model of RGC injury. The secondary purpose was to determine the time course of RGC functional abnormalities relative to RNFL birefringence and RNFL thickness changes. 

Methods

Subjects

The subjects of this study were four adult rhesus macaque monkeys (Macaca mulatta). Table 1lists the age, weight, and sex of each animal. All experimental methods and animal care procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the local Institutional Animal Care and Use Committee (IACUC). 

Anesthesia

All experimental procedures began with induction of general anesthesia using ketamine (15 mg/kg IM), along with a single SC injection of atropine sulfate (0.05 mg/kg). Animals were then intubated and breathed 100% oxygen for retinal function testing by electroretinography (ERG), during which anesthesia was maintained with a combination of ketamine (5 mg/kg/h IV) and xylazine (0.8 mg/kg/h IM). On completion of retinal function testing, ketamine-xylazine administration was discontinued, and isoflurane gas (1%–3%) was mixed with oxygen to provide anesthesia during structural imaging of the retina and optic nerve head (ONH). Isoflurane (1%–1.5%) was also used to provide anesthesia during optic nerve transection surgery. 

During all procedures, heart rate, and arterial oxyhemoglobin saturation were monitored continuously (Propaq Encore model 206EL; Protocol Systems, Inc., Beaverton, OR) and maintained above 75 minutes⁻¹ and 95%, respectively. Body temperature was maintained with a warm-water heating pad set at 37°C. 

Retinal Function Testing

Retinal function was evaluated by three different modes of ERG, as previously described. ^{23 24} Custom-designed Burian-Allen contact lens electrodes (10 mm diameter, +3.0 D; Hansen Ophthalmics, Iowa City, IA) were used for all ERG testing. The corneal ring on the stimulated eye served as the active electrode, and the corneal ring of the unstimulated (patched) contralateral eye served as the reference electrode. An SC ground electrode was placed on a rear limb. Electrode impedance was accepted if <5 kΩ. Before insertion of ERG contact lens electrodes, one drop of topical anesthetic (0.5% proparacaine) and an ocular lubricating agent (Celluvisc; Allergan, Irvine, CA) were applied to each eye. Head position was stabilized with a bite bar apparatus capable of rotation in three axes. 

Multifocal ERGs.

mfERGs were recorded with a commercial system (VERIS; ver. 4; EDI, San Mateo, CA). Residual refractive error was measured by retinoscopy for the test distance (25 cm) and corrected to the nearest half diopter. The mfERG stimulus was presented on a 21-in. monochrome monitor with a 75-Hz refresh rate. An initial set of brief recordings (2 minutes each) was used to center the stimulus on the visual axis such that the foveal and “blind spot” responses were positioned appropriately within the response array. 

The mfERG stimulus consisted of 103 unscaled hexagonal elements subtending a total field size of ∼55°. The luminance of each hexagon was independently modulated between dark (1 cd/m²) and light (200 cd/m²), according to a pseudorandom, binary m-sequence. Stimulus luminance was measured with a calibrated spot photometer (SpectraScan PR-650; Photo Research, Chatsworth, CA). The temporal stimulation rate was slowed by insertion of seven dark frames into each m-sequence step (7F). The m-sequence exponent was set to 12; thus, the total duration of each recording was 7 minutes, 17 seconds. The signals were amplified (gain = 100,000), band-pass filtered (10–300 Hz; with an additional 60-Hz line filter), sampled at 1.2 kHz (i.e., sampling interval, 0.83 ms), and digitally stored for subsequent off-line analyses. Two such recordings were obtained for each eye at each time point and averaged. 

From the average of the two recordings at each time point, a subset of local responses was exported for further analyses. Figure 1Ashows the stimulus locations of this subset, consisting of the central part of the array where RGC contributions are largest. ^{24 25} The response from the central stimulus element (marked with a C) and those from the two surrounding concentric rings were evaluated: locations are numbered 1 to 6 around the first ring and 1 to 12 around the second ring. 

Each local mfERG response was band-passed filtered (−3 dB at 65 and 250 Hz) to extract the high-frequency components (HFCs; Fig. 1B ). The low-frequency component (LFC) of each response was represented as the raw response minus the HFC. The amplitude of the HFC was calculated as the root mean square (RMS) for the epoch between 0 to 80 ms of each filtered record. Peak amplitudes for LFC features (Fig. 1B)were quantified as follows: the first negative feature (N1) was calculated as the maximum negative excursion from baseline in the epoch up to 30 ms; the amplitude of the first positivity (P1) was calculated as the voltage difference between the maximum peak and the N1 trough; and the second negativity (N2) was calculated as the difference between baseline and the minima from 30 to 80 ms. 

Pattern-Reversal ERGs.

Transient pattern-reversal ERGs (PERGs) were recorded (Utas-E3000 system; LKC Technologies, Inc., Gaithersburg, MD). The PERG stimulus was a checkerboard pattern (1° check size), reversing at 2.5 Hz (five reversals/second). The stimulus subtended 32° × 24° at the 50-cm test distance. Stimulus luminance was 75 cd/m² and contrast was >90%. The position of the foveal projection determined during mfERG testing was used to align the center of the PERG stimulus on the visual axis. Residual refractive error was measured for the test distance and corrected to the nearest half diopter. Signals were band-pass filtered 1 to 500 Hz and sampled at 2 kHz. Two records were obtained for each eye and then averaged. Each single record was an average of 200 sweeps. Eye position was monitored continuously and remained stable with sufficient depth of anesthesia. Amplitudes were measured for the primary features commonly known as P50 and N95 (see Fig. 2A ); the P50 was calculated as the difference between the peak and baseline; and the N95 was calculated as the difference between the peak around 50 ms and the trough (minimum) around 95 ms. 

Full-Field Flash ERGs.

Using the same UTAS-E3000 system, photopic full-field flash ERGs (ffERGs) were obtained after 5 minutes of light adaptation to a rod-saturating blue background (30 scotopic cd/m²; Wratten 78). Red stimulus flashes (Wratten 29; Eastman Kodak Co., Rochester, NY) with an intensity of 0.42 log photopic cd-s/m² were presented monocularly at 0.5 Hz via a Ganzfeld integrating sphere. Stimulus and background intensities were measured with a calibrated photometer (Spectra Pritchard PR-1980A; Photograph Research). Signals were band-pass filtered 0.3 to 500 Hz and sampled at 2 kHz. Two records were obtained and then averaged. Each single record was an average of 10 sweeps. 

Amplitudes were measured for four features, the a-wave, b-wave, oscillatory potentials (OPs), and photopic negative response (PhNR; Fig. 2B ). The a-wave amplitude was measured at the criterion time of 10 ms after the stimulus flash; the b-wave as the difference between the peak and the a-wave trough values (i.e., a peak-to-trough amplitude); and the PhNR as the difference between the value at the criterion time of 85 ms after the stimulus flash and the value at the b-wave peak. OPs were isolated with a Blackman filter (−3 dB at 65 and 240 Hz) and their summed amplitude was quantified as the RMS of the filtered waveform between 0 and 100 ms. 

Clinical Imaging of Retinal and ONH Structure

RNFL Birefringence.

Retardance measurements were obtained by scanning laser polarimetry (SLP; GDxVCC; Carl Zeiss Meditec, Inc., Dublin, CA). The instrument compensates for the effects of anterior segment (primarily corneal) retardance, to more accurately determine RNFL retardance. ^{14 26} Thus, anterior segment retardance measurements are obtained before initial baseline RNFL scans and then are used as compensation in all subsequent RNFL scans. A bite bar, which rotates in three axes, was used to align the head and eye properly, and autorefraction was used for each scan. Three RNFL scans were averaged for each eye at each time point. 

Figure 3provides an individual example (ONT3) of RNFL retardance data obtained by SLP. The pseudocolor maps in Figures 3A 3B 3C 3Drepresent retardance as RNFL thickness in micrometers. The SLP instrument detects the retardance of a cross-polarized source after a double pass through the tissue sample, assumes that RNFL thickness is linearly related to retardance, and then calculates and reports an estimate of RNFL “thickness” by using a linear conversion factor of 0.67 nm/μm ¹³ (as stated in the instrument manual ²⁷ ). 

Values of RNFL birefringence were exported for the small peripapillary locus (the SLP instrument’s default). The exported data consist of 64 samples along a peripapillary locus beginning on the temporal side of the ONH; proceeding around the superior, nasal, and inferior aspects of the ONH; and completing a circle at the temporal location, thus representing a profile commonly referred to as a TSNIT curve (e.g., Fig. 3E ). Each of the values in the TSNIT curve is an average from an 8-pixel-wide band centered on the optic disc. ²⁷ This band is indicated by the pair of gray circles concentric around the ONH on each of the pseudocolor birefringence maps in Figures 3A 3B 3C 3D . The inner and outer limits of the band are 27 and 35 pixels from the center of the optic disc, so the center of the band has a radius of 31 pixels. ²⁷ This corresponds to a scan angle with a radius of 6.1°, ²⁷ which translates to approximately 1.12 mm on the macaque retina (assuming an emmetropic eye with average axial length of 19 mm). ^{28 29}  

RNFL Thickness.

Thickness measurements were obtained by spectral-domain optical coherence tomography (SD-OCT; Spectralis HRA+OCT instrument (Heidelberg Engineering, GmbH, Heidelberg, Germany). The optical resolution of the instrument is ∼7 μm axially (depth) and ∼14 μm transversely. The depth of each A-scan is 1.8 mm and consists of 512 pixels, providing a digital depth sampling of 3.5 μm per pixel. Each B-scan spans 15° and consists of 768 A-scans providing a digital transverse sampling of 5 μm per pixel (in an emmetropic human eye with average axial length). For this experiment, volume scans consisting of 145 horizontal B-scan sections were centered on the ONH. Each B-scan in the volume spanned 15° horizontally and the block of 145 B-scans spanned 15° vertically (thus, B-scans were separated by 0.1034°, vertically). Radially oriented B-scans were also acquired with 48 sections arranged in a star pattern centered on the ONH. 

A real-time eye-tracking system measures eye movements and provides feedback to the SD-OCT scanning system to stabilize the retinal position of the B-scan. This system thus enables sweep averaging at each B-scan location to reduce speckle noise. For this experiment, nine sweeps were averaged for each B-scan. 

RNFL thickness measurements were derived from manual delineation of anterior (internal limiting membrane) and posterior borders along a single A-scan at the appropriate eccentricity within each radial B-scan. This eccentricity was chosen to correspond with the location of the RNFL birefringence measurement acquired by SLP. This eccentricity was determined to be equivalent to “1400 μm” from the center of the ONH, as indicated by a ruler within the OCT visualization software. In converting angular span to linear distance, the SD-OCT instrument assumes an emmetropic human eye with average axial length. This dimension translates to ∼1120 μm on the macaque retina (assuming 19 mm in axial length) ^{28 29} and thus corresponds to the locus of SLP birefringence measurements. Figure 4provides an example of the RNFL thickness measurement method. The radial B-scans were used for all RNFL thickness measurements reported herein. Cross-validation was performed by using the horizontal volume scans in which measurements at the same eccentricity and polar angle differed by ≤5 μm (∼1 pixel) from those made within radial B-scans. 

Confocal Scanning Laser Tomography.

CSLT was performed with a retina tomograph (HRT II; Heidelberg Engineering). During each imaging session, seven 15° × 15° scans were acquired, from which at least three were used to create a mean (Engelman CJ, et al. IOVS 2004;45:ARVO E-Abstract 5524; Cull G, et al. IOVS 2006;47:ARVO E-Abstract 745). ^{30 31 32 33} The effect of ONT on ONH surface topography was evaluated by two methods of analysis: topographic change analysis (TCA) ^{34 35} and the parameter mean position of the disc (MPD). ^{30 31}  

Stereoscopic Photographs.

Simultaneous stereoscopic photographs (3-Dx; Nidek Co., Ltd., Aichi, Japan) of the ONH and peripapillary retina were obtained at baseline and at the final time point (2 weeks after ONT). 

Digital Video Fluorescein Angiography.

Fluorescein angiography was performed 1 week after ONT the HRA+OCT instrument (Spectralis; Heidelberg Engineering) to evaluate retinal and choroidal circulation. After background fluorescence was briefly recorded, 0.8 mL of 10% sodium fluorescein (100 mg/mL, 80 mg total dose, Fluorescite; Alcon Laboratories, Inc., Fort Worth, TX) was injected intravenously as the clock timer was set to 0. 

Intraocular Pressure.

IOP measurements were made at the start of each session (Tonopen; Oculab, Inc., Glendale, CA). The value for each eye was taken as the average of three successive measurements. 

Optic Nerve Transection.

Nerve transsection was performed with the animal under general anesthesia (isoflurane). Complete transection of the intraorbital, retrobulbar optic nerve was achieved under direct visualization via lateral orbitotomy. Tissues were closed on multiple planes with 3-0 nylon sutures, and intraorbital and SC antibiotics were administered on closure. Pain medications were administered for the first five postoperative days in conjunction with veterinary staff. 

Experimental Design and Protocol

Two baseline sessions of retinal and ONH structural imaging (SLP, OCT, CSLT) and two baseline measurements of retinal function testing by ERG were completed in each eye before ONT. ONT was then performed on the right eye of each of the four animals. Structural imaging and ERG testing was performed 1 week and 2 weeks after ONT. The animals were killed 14 days after ONT by barbiturate overdose, and ocular tissues were harvested for a separate, ongoing study on proteomics. Thus, retinal and optic nerve tissues were not available for histopathologic evaluation for this study. 

Statistics

Repeated-measures analysis of variance (RM-ANOVA) was used to test for effects of treatment (ONT) and time (Prism ver. 4; GraphPad Software, Inc. San Diego, CA). Post hoc tests of differences between time points were performed as paired t-tests with the Bonferroni correction for multiple comparisons. 

Results

ONT surgery was successfully accomplished in three of the four animals. In the fourth (ONT4, Table 1 ) an intraoperative hemorrhage was observed within the orbit at the moment of transection, suggesting involvement of the central retinal artery (CRA) despite the transsection’s being ∼7 mm posterior to the globe. Fluorescein angiography 7 days after surgery confirmed that there was no direct perfusion of the CRA, therefore this animal was excluded from the study. Figure 5shows a single frame from the angiograph obtained 1 week after ONT in each of the four experimental eyes. Figures 5A 5B 5Ceach show a frame taken during the venous laminar filling phase, which demonstrate normal circulation for the three subjects included, while Figure 5Dshows the final frame of the angiograph for the excluded subject. Table 1lists the IOP for each eye at each time point. ONT had no effect on IOP (P = 0.72, RM-ANOVA) and there was no IOP difference between right eyes (ONT) and left eyes (control) at any time point (P > 0.05 for all post hoc tests). 

Figure 6shows the TCA change probability maps derived from CSLT data for subject ONT3, which were used to evaluate changes in ONH surface topography after ONT. The first two panels (Figs. 6A 6B)show the reflectance images obtained during each of two baseline sessions. Figures 6C and 6Dshow the change probability maps overlaid onto the reflectance images obtained during the 1- and 2-week post-ONT sessions, respectively. Red and green super pixels indicate that a significant change in height has occurred (with a probability of P < 0.05) at that location relative to baseline height and variability. Green superpixels indicate elevation, and red pixels indicate depression relative to baseline height. Figure 6demonstrates that there was little change in ONH surface topography after ONT. Across the group of three subjects, there was no significant effect of ONT on the MPD parameter (P = 0.84, RM-ANOVA). However, Figure 6Ddoes show that the brightness of the RNFL striations began to decrease, indicating a decrease in RNFL reflectance 2 weeks after ONT. 

Figure 3provides an individual example (ONT3) of RNFL birefringence and RNFL thickness changes after ONT measured by SLP and SD-OCT, respectively. Figure 3A and 3Bshow the birefringence maps obtained during the two baseline time points and Figs. 3C and 3Dshows the maps obtained at 1 week and 2 weeks after ONT, respectively. The TSNIT curves derived from the SLP data for this eye at each time point are shown in Figure 3E . The TSNIT average thickness at each time point was as follows: 53.7, 57.0, 48.7, and 28.0 μm. Thus, there was a 12% decline in RNFL retardance from the average baseline value 1 week after ONT and a 49% decline from baseline 2 weeks after ONT. Figure 3Fshows the TSNIT curves for the fellow control eye obtained during the same four time points. The TSNIT average values were 62.1, 59.9, 53.3, and 61.5, representing a coefficient of variation (COV) in this control eye of 6.8%. 

Figures 3G and 3Hshow the TSNIT curves for RNFL thickness measured by SD-OCT in the same subject (ONT3). The coefficient of variation across the four time points in the control eye was 0.9%, better than that for retardance. The TSNIT average values for RNFL thickness in the experimental eye were: 100.1, 112.8, 101.9, and 85.1 μm, indicating that RNFL thickness declined by 4% 1 week after ONT and by 20% 2 weeks after ONT. Figure 3Galso demonstrates emergence of the retinal blood vessels as RNFL thickness begins to decline and recede around the vessels, note how spikes in the TSNIT curve become prominent but maintain consistent position and thickness compared with baseline. 

Figure 7shows the results in the group of three subjects. The average TNSIT values for each eye at each time point were normalized to the mean of the two baseline values. Figure 7Ashows that RNFL retardance declined by 15% 1 week after ONT (P = 0.043), whereas there was no significant change in RNFL thickness (+1%, P = 0.42). Two weeks after ONT, RNFL retardance had declined by 39% (P = 0.018), whereas RNFL thickness had declined by only 15% (P = 0.025). Figure 7Bshows that there were no significant changes in the group of fellow control eyes for either RNFL retardance (P = 0.16) or thickness (P = 0.97). The data in Figure 7Balso confirm that RNFL thickness measurements had better intersession repeatability (average COV = 2.9%) than RNFL retardance measurements (average COV = 5.9%) among control eyes. 

Figure 1shows mfERG responses for the same individual subject ONT3. Figures 3C and 3Dshow the mfERG responses obtained during baseline from the left eye (control) and right eye (experimental), respectively. The HFCs are prominent in responses from both eyes and a strong nasal–temporal asymmetry is evident as the stimulus location changes around the two rings surrounding the central location. Figure 3Eshows the responses from the same locations 1 week after ONT was performed in the right eye, while 3F shows the responses from the fellow control eye that same day. The amplitude of the HFCs had declined substantially (66% on average across the 19 locations) 1 week after ONT. The response morphology had also become more similar around the rings, reflecting a decrease of the nasal-temporal asymmetry. The amplitude of LFC features also became smaller after ONT, though to a lesser extent than the HFC. 

Figure 2shows the PERG (Fig. 2A)and ffERG results (Fig. 2B)for the same individual subject ONT3. There is a marked decline in the amplitude of the PERG N95 component 1 week after ONT (91%, arrows) compared with either the baseline responses from the same eye or the responses from the fellow control eye (dashed curves). The effect of ONT on the PERG P50 component in this eye was to speed the apparent implicit time (by ∼2–4 ms), but the P50 amplitude did not decrease as much as the N95. 

Similarly, Figure 2Bdemonstrates that ONT had a strong and relatively selective effect on the amplitude of the ffERG PhNR (arrows), compared with either the baseline responses from the same eye or the responses from the fellow control eye (dashed curves). ONT had a larger effect on the PhNR compared with the a-wave, b-wave, or OPs. 

Table 2lists the results of retinal function testing for the group of three subjects. The values listed for 1 week and 2 weeks after ONT show the change from the average baseline for each parameter, each value represents the mean change for the three ONT eyes. The results of statistical testing are also listed (two-way RM-ANOVA). The probabilities in the first column represent the chance that the “treatment” effect (ONT) was due to random variation rather than to ONT; the probabilities in the second column represent the chance that the observed interaction between time and treatment (ONT) was due to random variation. The last column lists the COV calculated for the group of three fellow control eyes for each parameter, which provides a basis for comparing the observed effect of ONT against normal intersession variability. The results in Table 2demonstrate that there were significant changes in retinal function 1 week after ONT. The largest effects of ONT were on the mfERG HFC, the PERG N95 component and the PhNR of the photopic ffERG. There was a trend toward recovery of function 2 weeks after ONT for the N1 component of the mfERG N1 and the a-wave of the photopic ffERG. 

Discussion

The results of this study demonstrate that RNFL birefringence decreased 1 week after ONT, while RNFL thickness had not yet changed. By the second week after ONT, RNFL thickness had declined by 15%, while the decrease in retardance measured by SLP was more than twice that amount (39%), suggesting that RNFL birefringence (retardance per unit thickness) had declined further still. ^{3 12 13 17 18 20 21} It is thought that form birefringence of the RNFL is due to the orderly structural array of thin cylindrical cytoskeletal components within RGC axons such as MT and NF. ^{1 2 3} Previous studies have shown that RNFL birefringence declines rapidly after chemical disruption of MT, a component of the RGC cytoskeleton, in situ ³ or in vivo. ⁴ Thus, it has been suggested that measurements of RNFL birefringence could provide a sensitive indicator of compromised cytoskeleton within RGC axons. ^{3 12} The results of this study indicate that there is a stage during RGC degeneration where the axonal cytoskeleton has become abnormal enough to result in altered RNFL birefringence and that this stage precedes thinning of the axon bundles of the peripapillary RNFL. 

This intermediate stage observed in this study in response to a relatively acute experimental injury (ONT) may be common to other forms of RGC injury, such as in glaucoma. For example, it has been suggested that reduced gene NF expression represents a general response to RGC injury. ⁷ Moreover, abnormalities of cytoskeletal proteins such as NF and MT have been demonstrated in experimental models of glaucoma ^{8 11} and after short-term elevation of IOP. ^{9 10 36 37} This finding is important partly because critical functional capabilities depend on intact MT, as they provide the “tracks” on which most active axoplasmic transport takes place. ^{38 39 40} Interruption of axoplasmic transport has been proposed as a fundamental pathophysiological process in glaucoma and has been demonstrated to occur during acutely and chronically elevated IOP states in several mammalian species. ^{9 10 36 37 41 42 43 44} Chemical disruption of MT by colchicine or vinblastine halts axonal transport in RGCs. ^{38 39} Therefore, MT disruption may not only be caused by elevated IOP, but also leads to further abnormalities of axoplasmic transport, perhaps resulting in a vicious cycle. Further studies are under way to determine whether a similar intermediate stage of RGC degeneration, in which RNFL birefringence precedes changes in RNFL thickness, also occurs during experimental glaucoma. Initial results in four animals suggest that it does (Fortune B, et al. IOVS 2008;49:ARVO E-Abstract 3761). 

The secondary purpose of this study was to determine whether abnormalities of RGC function are associated with the intermediate stage of degeneration (i.e., whether altered function also precedes RNFL thinning after ONT). Three different modes of ERG testing were used to monitor retinal function. The pattern of results for each mode is consistent with loss of RGC function predominantly, although there was some evidence that mild disruption of function of other retinal elements may also have occurred. The photopic ffERG revealed a greater effect on the PhNR than on the a-wave, b-wave, or OPs. The PhNR is thought to be dependent on intact RGC function while the a-wave and b-wave represent responses of more distal retinal generators. ^{45 46 47} The N95 component of the PERG was affected more than the P50 component during the second week post-ONT. This pattern, again, is consistent with loss of predominantly inner retinal function, particularly of RGCs. ⁴⁸ The mfERG HFCs were affected by ONT to a greater extent than the LFC features N1, P1, and N2. This pattern is also indicative of abnormal RGC function. ^{23 24 25 49 50} It is possible that restricting the band-pass filter to isolate only the highest frequency content of these mfERG responses could provide an even more sensitive indicator and greater dynamic range over which to study effects on RGC function. ⁵⁰  

To the extent that the photopic ffERG a-wave represents cone photoreceptor signaling, ^{47 51 52} the results of this study suggest that there was a relatively mild, transient effect of ONT on function of the outer retina (although not statistically significant given the relatively small sample size and variability of the a-wave amplitude). The 10% reduction in a-wave amplitude 1 week after ONT, which seems to have resolved to 96% of baseline values by the second week after ONT, may have been due to the trauma associated with the orbitotomy. For the conditions of this study (background and flash intensities and chromaticities, criterion time of a-wave amplitude measurement), it is likely that the a-wave amplitude measurement reflects a relatively large contribution from hyperpolarizing second-order retinal neurons such as horizontal and off cone bipolar cells in addition to cone photoreceptor responses. ^{51 53} Nonetheless, the results suggest that any mild outer retinal functional changes recovered almost completely by the second week after ONT. 

In summary, the results of this study demonstrate that RNFL birefringence declines before and faster than RNFL thickness in the weeks after experimental injury of RGCs by ONT. This result suggests that disruption of the RGC cytoskeleton precedes RNFL thinning after injury. Abnormalities of RGC function were present along with decreased RNFL birefringence 1 week after ONT, when RNFL thickness and ONH topography were still normal. Collectively, the results are indicative of a stage of RGC dysfunction preceding changes in RNFL thickness. Further studies are under way to determine whether similar intermediate stages of degeneration and abnormal function are detectable in experimental glaucoma before changes in RNFL thickness. Finally, although the results reported herein were robust to rigorous statistical analysis and similar in all three subjects, the conclusions are based on a small sample size of three, which should be considered a limitation of this study. 

 Supported by National Eye Institute R01-EY011610 (CFB); the Glaucoma Research Foundation (BF); the Legacy Good Samaritan Foundation; Heidelberg Engineering, GmbH, Heidelberg, Germany (equipment); and Carl Zeiss Meditech, Inc. (equipment).

 Submitted for publication May 6, 2008; revised June 2, 2008; accepted August 21, 2008.

 Disclosure: B. Fortune, Carl Zeiss Meditec, Inc. (F); G.A. Cull, None; C.F. Burgoyne, Heidelberg Engineering, GmbH (F)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Study Subjects’ Age, Weight, Sex, and Intraocular Pressure

ID	Age (y)	Weight (kg)	Sex	Intraocular Pressure
								Baseline 1			Baseline 2			1 wk Post ONT			2 wk Post ONT
								OD	OS	OD	OS	OD	OS	OD	OS
ONT1	10	5.0	F	15.7	14.3	10.0	9.3	12.0	13.0	9.0	10.0
ONT2	11	9.5	M	17.0	17.0	16.0	15.3	9.0	12.0	12.0	12.0
ONT3	6	6.5	F	16.3	15.7	16.3	14.7	15.0	9.0	15.0	14.7
ONT4	6	4.7	F	10.0	10.0	10.0	10.0	9.0	12.0	NA	NA

Figure 1.

 

(A) mfERG stimulus pattern. Gray oval: position of the blind spot within the stimulus field for left eye recordings. Numbered locations: position of responses that were exported for analysis: C, central element; 1 to 6, first eccentricity ring; 1 to 12, second eccentricity ring. Analysis included amplitude measurements for the mfERG HFC (B) and three features of the LFC (raw response − HFC): the N1, P1, and N2 amplitudes. Representative individual example (ONT3) showing mfERG responses from the 19 locations listed in (A) for the left eye (control, C) and right eye (experimental, D) at baseline. Responses from the same locations 1 week after ONT in the left eye (control, E) and right eye (ONT, F).

Figure 2.

 

Representative individual example (ONT3) of PERG (A) and photopic full-field flash ERG responses (B). Broken traces: left eye (control) responses; solid traces: right (experimental) eye responses. Responses were repeatable in both eyes for two baseline sessions. Both the PERG and the PhNR of the ffERG response were reduced (arrows) 1 week and 2 weeks after ONT (p ONT).

Figure 3.

 

RNFL birefringence maps for one representative eye obtained by SLP during two baseline sessions (A, B), 1 week (C) and 2 weeks (D) after ONT. TSNIT curves for RNFL birefringence in the same eye at the same four time points (E): blue curve, first baseline; green curve, second baseline; orange curve, 1 week after ONT; red curve, 2 weeks after ONT. TSNIT curves for RNFL birefringence in the fellow control eye at the same four time points (F). TSNIT curves for RNFL thickness obtained by SD-OCT at the same four time points in the experimental eye (G) and fellow control eye (H).

Figure 4.

 

Representative example demonstrating method of RNFL thickness measurement by SD-OCT. Measurements were obtained at two locations along each of 80 radial sections centered on the ONH. Vertical arrow: the position and direction of the vertical section overlaid onto the fundus reflectance image in (A). (B) The actual SD-OCT B-scan section at this location. The RNFL thickness measurement was made at locations indicated by the instrument as being 1400 μm from the center of the ONH (this corresponds to ∼1120 μm on the macaque retina). The vertical white line corresponds to the horizontal hash just inferior to the ONH margin shown in (A). Higher magnification of (B) shown in (C); black double-headed arrow: RNFL thickness at the location shown by the hash mark in (A).

Figure 5.

 

Single frame taken from a digital video fundus fluorescein angiograph 1 week after ONT in each of the four subjects (ONT1, A; ONT2, B; ONT3, C; ONT4, D). Frame taken from laminar venous filling phase of angiograph in (A–C) and from the final frame of the angiograph in (D).

Figure 6.

 

Representative example of CSLT TCA map (ONT3). First baseline session (A); second baseline session (B); 1 week after ONT (C); 2 weeks after ONT (D). Red pixels: significant depression relative to baseline height; green pixels: significant elevation relative to baseline height.

Figure 7.

 

Group average (n = 3) RNFL measurements for the 360° peripapillary scan circle (TSNIT curves) obtained during each of two baseline sessions (BL1 and BL2, respectively), 1 and 2 weeks after ONT (wk 1 and 2, respectively) in the experimental eyes (A) and during the same four time points in the fellow control eyes (B). Data are presented for each time point as the difference from the average baseline value for each eye (i.e., normalized to the average baseline value for each eye). Open symbols and broken curves: RNFL retardance measurements; filled symbols and solid curves: RNFL thickness measurements.

Table 2.

 

Results of Retinal Function Testing

	1 wk Post ONT (%)	2 wk Post ONT (%)	P			Control COV (%)
							Treatment (ONT)	Time–Treatment Interaction
mfERG
N1	−33	−19	0.0618	0.0465	16
P1	−43	−26	0.0052	0.0041	12
N2	−56	−31	0.1082	0.1254	23
HFC	−65	−66	0.0176	0.0040	18
PERG
P50	−49	−39	0.1343	0.1468	15
N95	−70	−76	0.0069	0.0037	13
ffERG
a-wave	−10	−4	0.2384	0.0355	28
b-wave	−21	−24	0.0311	0.0359	17
OP’s	−21	−33	0.1000	0.0716	13
PhNR	−44	−48	0.0053	0.0003	21

 Values listed for 1 week and 2 weeks after ONT represent the change from the baseline average (mean of n = 3 subjects). Probabilities listed in each row are derived from a two-way RM-ANOVA; the first column represents the chance that the treatment effect (ONT) was due to random variation rather than to ONT, the second column represents the chance that the observed interaction between time and treatment (ONT) is due to random variation. The last column lists the COV among fellow control eyes for each parameter.

The authors thank Roger A. Dailey, MD, and Leonard A. Levin, MD, PhD, for helpful consultation and assistance toward optimization of lateral orbitotomy for optic nerve transection.