3-D Histomorphometry of the Normal and Early Glaucomatous Monkey Optic Nerve Head: Prelaminar Neural Tissues and Cupping

Abstract

purpose. To introduce a three-dimensional (3-D) histomorphometric strategy for characterizing the connective tissue (laminar) and prelaminar neural tissue (prelaminar) components of optic nerve head (ONH) cupping in one bilaterally normal monkey and three monkeys with early experimental glaucoma (EG) in one eye.

methods. Trephined ONH and peripapillary sclera from both eyes of four monkeys were serially sectioned at either 3-μm thickness (three EG monkeys) or 1.5-μm thickness (the bilaterally normal monkey) with the embedded tissue block face stained and imaged after each cut. Digital section images were aligned and stacked to create a 3-D reconstruction of each ONH. Within 40 digital radial sagittal sections of each reconstruction, Bruch’s membrane opening (BMO), the neural canal wall, and the anterior laminar surface were delineated by two delineators. The 80 BMO points were used to establish a BMO-zero reference plane. The parameters prelaminar tissue volume, post-BMO cup (the estimate of the clinical cup), and post-BMO total prelaminar volume (a global measure of ONH connective tissue deformation) were calculated overall and within 15° radial regions. The parameter prelaminar tissue thickness was calculated at each delineated anterior laminar surface point. For each monkey, an intra-animal difference map was generated for each parameter. Overall volume and thickness data were compared between normal and EG eyes by analysis of variance (ANOVA).

results. Regionally variable expansion of post-BMO cup volume and post-BMO total prelaminar volume were present in all three EG eyes and far exceeded the intra-animal, physiologic differences for these parameters in the bilaterally normal monkey. Prelaminar tissue thickness was increased in all three EG monkeys, with the greatest effects present within the peripheral regions of the canal.

conclusions. These data suggest that in young adult monkeys with more compliant connective tissues, clinical cupping in early glaucoma is primarily due to fixed deformation of the ONH connective tissues and occurs in the setting of prelaminar tissues that are thickened rather than thinned.

Cupping is a clinical term that is used to describe enlargement of the optic nerve head (ONH) cup in all forms of optic neuropathy. ^{1 2 3 4 5 6 7 8} However, cupping is also used as a synonym for the pathophysiology of glaucomatous damage to the ONH neural and connective tissues. ^{9 10 11 12} Because the clinical and pathophysiologic contexts for cupping are seldom clarified, there is a large and often confusing body of literature regarding the presence, importance, and meaning of cupping in a variety of optic neuropathies, including those of glaucoma, compressive orbital masses, ischemia, inflammation, and hereditary disorders. ^{1 2 3 4 5 6 7 8 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27}  

We propose that in all optic neuropathies, regardless of the location and etiology of the primary insult to the visual system, the clinical phenomenon of cupping, (herein referred to as clinical cupping) has two principal pathophysiologic components: prelaminar thinning and laminar deformation (Fig. 1) . We define prelaminar thinning to be that portion of cup enlargement that results from thinning of the prelaminar tissues due to physical compression and/or loss of retinal ganglion cell (RGC) axons. We define laminar deformation to be that portion of cup enlargement that results from lamina cribrosa, scleral flange, and peripapillary scleral connective tissue damage followed by permanent, intraocular pressure (IOP)–induced deformation. ^{28 29 30 31 32}  

We further propose that the following clinical and pathophysiologic concepts regarding prelaminar thinning and laminar deformation are true: (1) Prelaminar thinning results in a clinically shallow form of cupping ^{16 33 34} (being limited to the prelaminar tissues) that occurs in all forms of RGC axon loss and is therefore nonspecific; (2) laminar deformation results in a clinically deeper form of cupping that occurs only in those optic neuropathies in which the ONH connective tissues have been damaged and have become susceptible to permanent IOP-induced deformation (whether the IOP is normal or elevated). ^{29 35}  

The purposes of this study were to test the hypothesis that laminar deformation rather than prelaminar thinning underlies the onset of confocal scanning laser tomographic (CSLT)–detected ONH surface change (Fig. 2)in young adult monkey eyes exposed to moderate experimental IOP elevations and to create clinically aligned, three dimensional (3-D) volumetric maps of the magnitude and location of these two cupping components in the same eyes. 

This is the third in a series of five articles devoted to 3-D histomorphometric quantification of the ONH and peripapillary neural and connective tissues. In the first report, ³¹ we introduced our method for 3-D delineation of 13 ONH and peripapillary scleral landmarks, tested the method’s reproducibility, and used it to describe enlargement and elongation of the neural canal at the onset of CSLT-detected ONH surface change in three monkeys with early experimental glaucoma (EG) in one eye. In the second report, ³² we described our method for continuous mapping of the position and thickness of the lamina cribrosa, scleral flange, and peripapillary sclera and used the data collected to report significant posterior deformation and thickening of the lamina cribrosa accompanied by mild posterior deformation of the scleral flange and peripapillary sclera in the same EG eyes. 

In the present report, we introduce four new postmortem 3-D histomorphometric cupping parameters and use them to quantify the prelaminar and laminar components of cupping in these same three EG eyes. In addition, we reconstruct both ONHs of one bilaterally normal monkey, so as to characterize physiologic, intra-animal differences for these same parameters. 

Materials and Methods

Animals

All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Three male cynomolgus monkeys, approximately 8 years of age, were used for the studies (see Table 1of our previous report ³⁰ ). 

In addition, we included both eyes of one bilaterally normal male rhesus monkey (laboratory ID 97R0793) age 6 years that was imaged and euthanatized similar to the EG monkeys as outlined later. 

ONH Surface Compliance Testing and Early Glaucoma

We have described our Laser Diagnostic Technologies (LDT, San Diego, CA) CSLT-based ONH surface compliance testing strategy and how we use it to detect the onset of early EG. ^{28 30} Briefly, both eyes of each monkey were imaged on three separate occasions while normal, and then lasering of the trabecular meshwork was begun in one eye of each animal to elevate IOP. CSLT imaging was continued at 2-week intervals until the onset of significant permanent posterior deformation of the ONH surface in the lasered eye (the EG eye), compared with the contralateral eye (the normal eye) using the CSLT-based parameter mean position of the disc. ³⁰ See Table 1and Figure 1in our previous publication ³⁰ regarding the magnitude and duration of IOP elevation experienced by each animal. EG monkeys 2 and 3 were euthanatized 3 weeks and monkey 1, 6 weeks after CSLT detection of ONH surface change. In EG monkeys 1 and 2, IOP elevations were moderate, with only one measurement higher than 30 mm Hg. In EG monkey 3 elevated IOP was not detected. ³⁰  

The bilaterally normal monkey was compliance tested on five occasions and then euthanatized. 

Monkey Euthanatization and Perfusion Fixation at Prescribed IOP

In both the normal and EG monkeys, both eyes were cannulated with a 27-gauge needle with the animals under deep pentobarbital anesthesia, and the IOP was set to 10 mm Hg using an adjustable saline reservoir. After a minimum of 30 minutes, the monkey was perfusion fixed via the descending aorta with 1 L of 4% buffered hypertonic paraformaldehyde solution followed by 6 L of 5% buffered hypertonic glutaraldehyde solution. ³⁰ After perfusion fixation, IOP was maintained for 1 hour, after which each eye was enucleated, all extraorbital tissues were removed, and the anterior chamber was removed 2 to 3 mm posterior to the limbus. By gross inspection, perfusion was excellent for all eight eyes. The posterior scleral shell with intact ONH, choroid, and retina were placed in 5% glutaraldehyde solution for storage. 

Generation of the Aligned Serial Section Images for Each ONH and 3-D ONH Reconstruction

These steps have been described in detail in our previous report. ³⁰ Briefly, the ONH and peripapillary sclera were trephinated (6-mm diameter), pierced with alignment sutures, embedded in paraffin, and mounted on a microtome. Then, the block surface was stained with a 1:1 (vol/vol) mixture of ponceau S and acid fuchsin stains and imaged at a resolution of 2.5 × 2.5 μm per pixel. The sections were serially cut away at 3.0-μm thickness, and the staining and imaging process was repeated after each cut. Imaging began at the vitreoretinal interface and continued approximately 200 μm into the retrolaminar orbital optic nerve. Serial section images were aligned in the anterior-to-posterior direction and stacked at 3.0-μm intervals into a 3-D reconstruction of the ONH and peripapillary scleral connective tissues, which consist of approximately 1080 × 1520 × 400 voxels, each 2.5 × 2.5 × 3.0 μm in size. For both eyes of the normal monkey, a new high-resolution protocol (at 1.5 × 1.5 μm per pixel resolution) was instituted and 1.5-μm sections were cut. 

3-D Delineation of ONH and Peripapillary Scleral Landmark Points

Our 3-D delineation technique has been described in detail in our previous reports. ^{31 32} Briefly, using custom software based on the Visualization Toolkit (VTK, Clifton Park, NY), the 3-D ONH reconstruction is loaded into memory on a remote Linux server. The delineator assigns the approximate center of the neural canal as that reconstruction’s center of rotation, around which 40 seven-voxel-thick, digital radial sagittal slices of the digital 3-D reconstruction are serially served at 4.5° intervals to the delineator’s workstation (Fig. 3) . 

Within each digital section (Fig. 3C) , the delineator marks seven landmark surfaces and six pairs of neural canal landmark points (one point on each side of the canal; Fig. 3D ). In this study, only the anterior lamina cribrosa, the neural canal wall, the BMO, and the anterior laminar insertion (ALI) landmark points (Fig. 3B)were used. 

While marking in the sagittal section view window (Fig. 3D) , the delineator is simultaneously viewing an adjacent window showing the cursor’s 3-D location within a digital transverse section image (Fig. 3E)slaved to the sagittal section view. The delineator also can scroll through the adjacent six one-voxel-thick sagittal section images to locate a section in which the landmark can be clearly identified. The 3-D Cartesian coordinates and category for each mark are saved, generating a 3-D point cloud that represents each of the marked structures (Fig. 3F) . 

Clinical Alignment of the 3-D Reconstruction

A high-resolution reconstruction of the central retinal vessels and the neural canal landmark points ^{31 32} is constructed and three-dimensionally overlaid onto a clinical fundus photograph or CSLT image, to align the 3-D marks accurately to anatomic orientation (superior, inferior, nasal, and temporal; see Fig. 2of our previous publication). ³¹  

BMO-Zero Reference Plane

For each 3-D ONH reconstruction, a least-squares ellipse was fit to the 80 marks defining BMO, creating a BMO-zero reference plane (Fig. 4 , and Figs. 1K and 1Lof our previous report ³¹ ). In the current study, our quantification of post-BMO cup volume and post-BMO total prelaminar volume were made relative to this plane. 

Parameterization

In this report we introduce four parameters to characterize ONH cupping: post-BMO cup volume, post-BMO total prelaminar volume, prelaminar neural tissue volume, and prelaminar neural tissue thickness. Prelaminar neural tissue thicknesses were calculated identically to laminar and peripapillary scleral thicknesses, as described in Figure 2of our previous report. ³² Briefly, the smoothed anterior laminar surface is used to generate a normal vector at each delineated point. At each delineated point, prelaminar tissue thickness was calculated along the normal vector as the shortest distance from the anterior laminar surface to the internal limiting membrane surface. 

The three volumetric parameters are defined in Figure 4 . Post-BMO cup volume is a measure of the clinical cup that we define to be the volume of empty space beneath the BMO-zero reference plane, but above the internal limiting membrane. Post-BMO total prelaminar volume is designed to detect all forms of neural canal wall and lamina cribrosa connective tissue deformation and is defined to be the total volume (tissue and nontissue), beneath BMO-zero reference plane, above the lamina cribrosa and within the neural canal wall. Prelaminar tissue volume is intended to detect increases or decreases in total prelaminar tissue volume. It is defined to be the total volume of tissues (including vascular and neural) above the lamina cribrosa, below the internal limiting membrane, and within the anterior projection of BMO. 

The calculation and regionalization of the three volumetric parameters are explained by using post-BMO cup volume as an example in Figure 5 . Briefly, within each ONH reconstruction, the internal limiting membrane points were fit to a continuous surface using a thin-plate B-spine (MatLab; The Mathworks, Natick, MA). Serial 3.0-μm digital sections (1.5 μm for the high-resolution normal monkey reconstructions, as described earlier) parallel to the BMO-zero reference plane were generated (Geomagic, Research Triangle Park, NC) and contours of the post-BMO cup within each digital section were created and divided into 24, 15° radial regions, with the BMO centroid (Fig. 5G)as the center. 

The total volume within each 15° region was calculated by adding the area within each contour and multiplying by either 3.0 μm (EG monkeys) or 1.5 μm (normal monkey). Total overall volume was calculated by adding together the regional volumes of all 24, 15° regions. The post-BMO total prelaminar volume and prelaminar tissue volume were calculated in a similar way. 

Data Visualization and Difference Map Generation

For both eyes of each monkey, post-BMO cup volume, post-BMO total prelaminar volume, and prelaminar tissue volume within each 15° radial region were clinically plotted in right eye configuration and clinically overlaid onto the CSLT image of each eye (Figs. 6 7 8) . A difference map comparing two eyes of each monkey (EG − N for the EG monkeys, left eye – right eye for the normal monkey) for each 15° radial volume (Figs. 6 7 8)was then generated. 

Continuous plots of prelaminar tissue thicknesses were generated for each eye by interpolating between the values measured at each delineated anterior laminar surface point (Delaunay-based cubic interpolation; MatLab; The Mathworks) and plotted in right eye configuration (Fig. 9) . Difference maps for each animal were generated by overlaying the BMO centroid for each eye and subtracting the normal data from the EG eye data for the EG monkey or subtracting right eye data from left eye data for the normal monkey (Fig. 9) . 

Variability and Reproducibility Study Design

For this report, interobserver variability was assessed for two delineators by having them delineate the landmark points of both eyes of all three EG monkeys. Intradelineator reproducibility was assessed for the second delineator by having her delineate both eyes of EG monkey 3 on two subsequent occasions (three total) separated by at least 2 weeks. The effect of two delineators (overall) for all four parameters and delineation day for prelaminar tissue thickness (delineator 2 on EG monkey 3 only) were assessed within separate factorial ANOVAs, as outlined in the next section. 

Data Analysis for Volumetric Parameters

Because our volumetric parameters do not lend themselves to multiple measures within individual eyes, a statistical assessment of intra-animal differences for each monkey was not possible. However, because there were two delineators, factorial ANOVAs were used to assess the effects of delineator, treatment (normal versus EG), and region on the three volumetric parameters considered overall for all three EG monkeys (all normal eyes pooled and compared with all EG eyes). 

Prelaminar Tissue Thickness Regionalization and Statistical Analysis

For statistical inference testing, prelaminar tissue thickness data at each delineated point on the anterior surface was assigned to one of 17 subregions, as depicted in our report for the lamina cribrosa position and thickness data. ³² Factorial ANOVAs were used to assess the effects of delineator, region, and treatment (normal or EG) on the prelaminar tissue thickness, both overall and between the two eyes of each monkey for all three EG monkeys. Another factorial analysis was used to assess the effect of eye (left or right) and region to assess the overall and regional physiologic intra-animal difference for the one normal monkey. Finally, a factorial ANOVA then assessed the effects of delineator repetition, region, and treatment within and between the data from both eyes of EG monkey 3. 

Optic Nerve Axon Counting

Optic nerve axon counts were performed in the optic nerve histologic sections from both eyes of the three EG monkeys, obtained approximately 2 to 3 mm behind the ONH, according to a previously published protocol. ³⁶  

Retrospective Topographic Change Analysis Maps

Although our definition of EG was based on our CSLT-based parameter mean position of the disc, as outlined earlier, we retrospectively applied the current Heidelberg Retinal Tomograph Topographic Change Analysis strategy (Artes P et al. IOVS 2006;47:ARVO E-Abstract 3732) ^{37 38} to the normal and the EG eye Laser Diagnostics Technology CSLT data, so as to generate longitudinal topographic change analysis maps for both eyes of each EG monkey (Fig. 2 , all in right eye orientation). Because the bilaterally normal monkey was imaged on only five occasions, no longitudinal change maps are presented. 

Results

Descriptive data on the onset of glaucoma in the three EG monkeys are reported in Table 1of our previous publication. ³⁰ CSLT topographic change maps demonstrating the onset and progression of ONH surface change in the EG relative to the normal eye of each EG monkey are shown in Figure 2 . 

Regional Volumetric and Continuous Thickness Maps for Each Monkey

Clinically aligned regional maps of the three volumetric parameters along with either EG − N (the EG monkeys) or left eye − right eye (the normal monkey) difference maps are reported for all four monkeys in Figures 6 7 and 8 . Clinically aligned continuous maps of prelaminar tissue thickness along with intra-animal difference maps are reported for all four monkeys in Figure 9 . 

In all three EG monkeys, large increases in post-BMO cup and post-BMO total prelaminar volume were accompanied by substantial increases in prelaminar tissue thickness and mild increases in prelaminar tissue volume. For three of the four parameters (the prelaminar tissue volume excepted), EG − N eye differences in the three EG eyes far exceeded physiologic intra-animal differences (left eye − right eye) in the normal monkey. 

Qualitative examination of the difference maps showed that post-BMO cup volume expansion was most prominent inferiorly and inferonasally and post-BMO total prelaminar volume expansion was greatest within the inferior, inferonasal, and superior regions in all three EG eyes. Prelaminar tissue volume increased within the inferior, inferonasal, and superior regions in two EG eyes and decreased slightly inferotemporally and superonasally in two of the three EG eyes. Prelaminar tissue thickened inferiorly and inferonasally in all three EG eyes, especially in the peripheral regions. 

Overall Volume and Thickness Data by Treatment and Delineator

Overall, normal versus EG eye data for the three EG monkeys (for each parameter and both delineators) are reported in Table 1 . Significant increases in all four parameters were seen within the pooled EG data from both delineators when compared to their contralateral normal controls (P < 0.05, ANOVA). These ranged from a 24% increase for prelaminar tissue volume to a 368% increase in post-BMO cup volume. Overall effects for parameters are similar within the data from each delineator. Within the ANOVA, the effect of delineator was not significant for all four parameters (P = 0.29 for post-BMO cup volume; P = 0.247 for post-BMO total prelaminar volume; P = 0.1212 for prelaminar tissue volume; and P = 0.4102 for prelaminar tissue thickness, ANOVA). 

Regional Volume and Thickness Data by Treatment and Delineator

Statistically significant regional differences in the four parameters between the pooled EG and normal eyes of the three EG monkeys (P < 0.05, ANOVA) are reported for each delineator in Figure 10 . Overall regional effects for all parameters were greatest along the superotemporal to inferonasal axis, being greatest inferonasally, and similar within data from each delineator. 

Overall Volume and Overall Thickness Data for Each Monkey by Delineator

Table 2reports the normal eye and intra-animal difference of all four overall parameters for delineators 1 and 2. Post-BMO cup volume was greatly expanded in all three EG eyes (0.088–0.130 mm³) compared with contralateral normal eyes, and these values far exceeded the intra-animal difference within normal monkey 1 (0.006 mm³). Post-BMO total prelaminar volume was also greatly expanded within each EG monkey (0.122–0.308 mm³) and far exceeded the intra-animal difference within the normal monkey (0.031 mm³). Prelaminar tissue volume expanded modestly within each EG monkey (0.002–0.167 mm³) with only EG monkey 1 dramatically exceeding the intra-animal difference within the normal monkey (0.021 mm³). Prelaminar tissue thickness was substantially increased within each EG eye (47–82 μm, P < 0.05, ANOVA) far exceeding the normal monkey’s intra-animal difference (5 μm). 

Regional Thickness Data for Each Monkey by Delineator

Statistically significant regional differences in prelaminar tissue thickness between the two eyes of each monkey (P < 0.05, ANOVA) are reported for each delineator in Figure 11 . The regional differences of the three EG monkeys reported by two delineators were very similar both in magnitude and pattern. 

Prelaminar tissue thickening was present within most regions, being greatest inferonasally in all three EG eyes, especially within the peripheral regions of EG monkeys 1 and 2 and the midperipheral regions of EG monkey 3. Regional differences between the two eyes of the normal monkey demonstrated significant differences (P < 0.05, ANOVA) within central and inferior regions that did not exceed 30 μm. However, these physiologic, intra-animal differences were far smaller than the treatment differences (EG – N) within all three EG monkeys. 

Overall and Individual Eye Volume and Thickness Data for EG Monkey 3 on Three Different Delineation Days

Overall volume and thickness data for all four parameters within the three delineations of delineator 2 are reported in Table 3 . Overall values for each parameter on different delineation days are remarkably similar. Within the factorial ANOVA of prelaminar tissue thickness the effect of delineation day was not significant (P = 0.2514, ANOVA). 

Overall Regional Difference Maps for All Four Parameters for Both Delineators

Overall regional differences for each parameter for each delineator are reported in Supplementary Figures S1–S4. In this ANOVA, although the delineator was significant, the overall regional difference maps for each delineator were remarkably similar and the magnitude of the interdelineator differences was much smaller than the treatment differences. 

Regional Difference Maps for Monkey 3 from Three Different Delineations Days

Regional differences for each parameter for each delineation day are reported in Supplementary Figure S5. For each parameter, the regional difference maps were remarkably similar on each delineation day, and the magnitude of the interdelineator differences was much smaller than the treatment differences. 

Axon Counts

Overall axon loss in the three EG optic nerves ranged from 16% to 30% compared to the contralateral normal optic nerves. Although there was some tendency for loss to concentrate inferiorly and superiorly, there was substantial nasal involvement, and the regional pattern of axon loss appeared more diffuse than focal (Fig. 12) . 

Discussion

Accurately characterizing ONH and peripapillary scleral neural and connective tissue architecture has been the goal of an important body of literature, which has been extensively reviewed in our previous reports. ^{28 30 31 32} In the present study, we introduced a strategy for 3-D histomorphometric quantification of ONH prelaminar tissue architecture that separates the clinical phenomenon of cupping into laminar (deep) and prelaminar (shallow) pathophysiologic components (Fig. 1) . We then used this strategy to report the presence of laminar deformation in the setting of prelaminar neural tissue thickening, not thinning, in three EG monkey eyes. 

The principal findings of this report are as follows. First, in the young adult monkey eye exposed to moderate levels of chronic IOP elevation, early glaucomatous cupping was principally the result of fixed deformation of the connective tissues of the neural canal wall and lamina cribrosa, rather than a manifestation of prelaminar neural tissue thinning due to compression or axon loss. Because each animal had IOP lowered to 10 mm Hg in both eyes for at least 30 minutes before perfusion fixation, these changes probably represent permanent ONH and peripapillary connective tissue deformation. To our knowledge, the fact that in ocular hypertension, the onset of CSLT-detected ONH surface change occurs in the setting of laminar deformation and prelaminar tissue thickening, not thinning, has not been described (Fig. 13) . 

All three EG eyes demonstrated a large laminar component of cupping without an important prelaminar component (Fig. 13) . By this, we mean that post-BMO total prelaminar volume (the parameter and the entity) was much larger in each of the three EG eyes than in the contralateral normal eyes, and these differences far exceeded the physiologic intereye difference for this parameter within the normal monkey. We propose that an increase in this parameter cannot occur except in the setting of pathologic expansion of the connective tissue of the neural canal wall and lamina cribrosa. We further propose that pathologic expansion of these tissues occurs only under the influence of a level of IOP-related stress and strain that exceeds the yield point (or elastic limit) of the connective tissues whether IOP is normal or elevated. 

That an increase in post-BMO total prelaminar volume can only be due to connective tissue deformation follows from its definition (Fig. 4) . We have reported expansion of the neural canal wall and posterior deformation of the lamina cribrosa relative to BMO-zero reference plane in these ^{30 31 32} and other ²⁸ EG eyes. The volumetric parameter post-BMO total prelaminar volume captured all components of this deformation in a single parameter, and suggests the principal axis of that deformation is superotemporal to inferonasal when the volumetric expansion is looked at regionally. 

That the presence of permanent or fixed deformation of these tissues proves that they are experiencing IOP-related strain which exceeds their elastic limit follows from basic engineering principles. ²⁹ First, although it is possible that new scar tissue or activated myofibroblasts can contract and pull the tissues into deformation, we believe that until this phenomenon can be demonstrated in a model of glaucomatous damage that does not use elevated IOP, this explanation for ONH connective tissue deformation is unlikely. Hayreh et al. ³⁹ has reported thickening of the lamina cribrosa in monkey EG and separately proposed that retrolaminar ischemia and fibrosis could pull the lamina posteriorly and cause a glaucomatous appearance to the ONH. However, although the lamina was thickened in all three EG eyes in this report, ³² we saw no evidence of retrolaminar fibrosis in these eyes. 

Second, although clinical cupping was evident in the TCA change maps and post-BMO cup volume was enlarged in all three EG eyes, the prelaminar neural tissues increased in both volume and thickness in all three EG eyes. Thus, even in the setting of 16% and 30% axon loss, there was no evidence of prelaminar cupping in any of the three EG eyes. Most strategies for ONH surface changed detection presume “thinning” or “loss” of prelaminar neural tissue. Our data suggest that EG ONH surface change reflected a combination of underlying connective tissue deformation and overlying prelaminar neural tissue thickening. Therefore, in ocular hypertensive eyes with robust connective tissues that are less likely to deform, counterintuitive changes in ONH and peripapillary retinal parameters may reflect true pathologic swelling of the prelaminar neural tissues and should not be discounted. 

The prelaminar tissue thickening is probably the result of orthograde axoplasmic transport blockage within the RGC axons combined with prelaminar tissue edema and gliosis. Acute and chronic IOP elevations in experimental animals have been shown to inhibit both orthograde and retrograde axonal transport in the lamina cribrosa. ^{40 41 42 43} Gliosis within the prelaminar and laminar tissues in glaucomatous eyes has also been reported. ^{43 44 45 46}  

Third, the pattern of increase in prelaminar neural tissue volume and thickness closely followed the pattern of connective tissue deformation in two of the three (for volume) and all three (for thickness) EG eyes. This relationship is particularly striking for the axes of greatest prelaminar neural tissue thickness increase in the three EG eyes compared with the predominant axes of connective tissue deformation as captured by post-BMO total prelaminar volume (Fig. 12) . 

Fourth, there is no obvious correlation between retrospective ONH surface change (by TCA mapping), posteuthanatization subsurface structural change (by 3-D histomorphometry) and posteuthanatization retrolaminar optic nerve axon loss in these three EG eyes (Fig. 12) . These relationships will be best elucidated in future studies that combine prospective TCA and subsurface structural change characterization (using next-generation ocular coherence tomography [OCT] imaging) ⁴⁷ followed by posteuthanatization 3-D histomorphometry and retrolaminar axon counting in a larger group of EG eyes. 

The limitations of our method of 3-D reconstruction have been discussed ³⁰ and include: (1) For the three EG monkeys in this study, anterior-to-posterior resolution was limited to 3 μm by the fact that we cut 3-μm-thick histologic sections. However, we have now accomplished 1.5-μm serial sectioning as outlined in the methods for the bilaterally normal monkey of this report. Thus, the resolutions of three EG monkeys and one normal monkey are different, but the higher resolution of the normal monkey only serves to characterize more accurately the full-range of intra-animal difference; (2) the stain is applied by hand to the block face with a cotton-tipped swab, and the excess is manually removed with lens paper; thus, staining variation between section images can be substantial; (3) there are tissue shrinkage effects (from both fixation and embedding) associated with this technique, but since all eyes were treated identically, comparisons between the two eyes of each monkey and within treatments should be valid; and (4) in this study we have characterized only physiologic intra-animal differences for these parameters in one bilaterally normal monkey—five bilaterally normal monkeys will be characterized in a future report. 

We choose 15° radial regions for our volumetric parameters because the accuracy of clinically overlaying the vessel reconstructions onto clinical photographs is approximately 3° (data not shown). In comparing the two eyes of a monkey, alignment errors of 3° in opposite directions in each eye still ensures 9° degrees of overlapping data within each 15° region. 

Whereas the normal monkey reconstructions were aligned to photos, for the three EG monkeys, clinical alignment of the ONH reconstructions was performed to false-colored CSLT images that may not provide as clear an image of the clinical optic disc margin. However, in a subsequent alignment of both eyes of EG monkey 3 to clinical photographs (the only animal for which they were available), no shift in alignment over that achieved with CSLT images was necessary (Fig. 13) . ³¹  

Our volumetric parameters do not lend themselves to assessments of statistically significant differences between the two eyes of a monkey. Our strategy for quantification is time consuming and, in practice, will be based on a single delineation of both eyes of each monkey by a single delineator. For a given ONH, the landmark point clouds produced by a single delineation generate one measurement of each volumetric parameter for each 15° radial volume. Without a true replication of this measurement for each ONH, no statistical comparison to the single measurement of the contralateral ONH is possible. 

However, for our 3-D histomorphometric studies, we are currently delineating both eyes of five normal IOP monkeys, perfusion fixed with both eyes set to 10 mm Hg by manometer before euthanatization, from which we will begin to characterize a 95% confidence interval for the physiologic intereye differences for each cupping parameter within each 15° radial volume. In the future we will add additional normal animals to this characterization and for a given EG monkey, only those treated-eye volume differences that exceed the 95% confidence interval will be considered significant for these normal eyes. 

Finally, it is possible that the acute IOP lowering to 10 mm Hg 30 minutes before perfusion fixation induced the prelaminar tissue thickening seen in all three EG eyes. However, we think this is not likely for the following reasons. First, IOP was lowered to 10 mm Hg (slowly, over 1–2 minutes) and was not that high in the EG eye of each monkey on the day of euthanatization. IOP was 32 mm Hg in the EG eye of monkey 1, 18 mm Hg in the EG eye of monkey 2, and in the EG eye of monkey 3 was not checked on the day of euthanatization but was never detectably elevated. ³⁰ Although it is possible that the change from 32 to 10 mm Hg caused thickening in EG monkey 1, we doubt that the pre-euthanatization IOP change in EG monkeys 2 and 3 accounts for this finding. We predict that longitudinal OCT imaging of prelaminar tissue volume and thickness (once clinically available), will confirm this finding in hypertensive human and monkey eyes. 

We propose that clinical quantification of the four prelaminar parameters described in this report should become an important goal of clinical glaucoma imaging. ³⁵ While histologically characterized in this report, these parameters may soon be clinically quantifiable by next-generation, high-resolution OCT imaging. ⁴⁷ In that setting, if multiple measures are made initially, longitudinal change in these parameters may be detectable in longitudinal images of the same eye acquired over time. 

The data in this report suggest two important concepts for imaging all forms of optic neuropathy. First, detection of ONH surface change (clinical cupping) suggests the presence of an optic neuropathy but does not confirm that IOP is an active etiologic agent. Thus, regardless of clinical circumstance, but particularly when IOP is not detectably elevated, ONH surface change (clinical cupping) without connective tissue deformation should not be an absolute indication for lowering IOP. 

We have previously proposed that in patients with robust ONH connective tissues, IOP-related stress and strain can cause a prelaminar form of cupping by causing axonal degeneration without damaging the underlying connective tissues. ^{29 35} Having proposed this concept, we now emphasize that at present, without direct evidence of ONH connective tissue damage, the role of IOP in an individual optic neuropathy cannot be certain. 

However, in contrast to surface change detection (clinical cupping), clinical detection of post-BMO total prelaminar volume expansion will serve to confirm ONH connective tissue damage (i.e., a laminar contribution to cupping) and will do so without having to image the connective tissues themselves. In so doing, clinical detection of post-BMO total prelaminar volume expansion, once possible, should become an absolute indication for lowering IOP, regardless of the level of IOP or the etiology of the primary connective tissue insult (ischemia, autoimmune, inflammatory, or IOP-related strain). ^{29 35}  

Finally, clinical characterization of the four parameters introduced in this report should provide quantitative definitions for common clinical descriptions of clinical cupping such as “shallow,” “deep,” “senile sclerotic,” “thinned,” “tilted,” and “excavated.” We propose that such quantification should underlie the characterization of all forms of clinical cupping by the regional presence and magnitude of constituent laminar and prelaminar components. 

 Presented in part at the annual meeting of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, May 2007.

 Supported in part by National Eye Institute Grant R01EY011610 (CFB); a grant from the American Health Assistance Foundation, Rockville, Maryland (CFB); a grant from The Whitaker Foundation, Arlington, Virginia (CFB); a Career Development Award (CFB) and an unrestricted departmental grant (LSU Eye Center) from Research to Prevent Blindness, Inc.

 Submitted for publication June 28, 2007; revised July 24, 2007; accepted September 12, 2007.

 Disclosure: H. Yang, None; J.C. Downs, None; A. Bellezza, None; H. Thompson, None; C.F. Burgoyne, None

 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.

 

Figure 1.

 

All clinical cupping, regardless of etiology, is a manifestation of underlying prelaminar and laminar pathophysiologic components. (A) Normal ONH. To understand the two pathophysiologic components of clinical cupping, start with (B) a representative digital central horizontal section image from a postmortem 3-D reconstruction of this same eye (white section line in A): vitreous (top), orbital optic nerve (bottom), lamina cribrosa between the sclera and internal limiting membrane (ILM) (green dots). (C) The same section is delineated into principle surfaces and volumes (black, ILM; purple, prelaminar neural and vascular tissue; cyan blue line, BMO-zero reference plane cut in section; green outline, post-BMO total prelaminar area or a measure of the space below BMO and the anterior laminar surface (see Fig. 4 ). (D) Regardless of the etiology, clinical cupping can be shallow (E) or deep (F; these clinical photographs are representative and are not of the eye in A). A prelaminar, or shallow, form of cupping (G, arrows) is primarily due to loss of prelaminar neural tissues without important laminar or ONH connective tissue involvement. Laminar or deep cupping (H, small white arrows) follows ONH connective tissue damage and deformation that manifests as expansion of the total area beneath BMO, but above the lamina. Notice in (H) that whereas a laminar component of cupping predominates (white arrows), there is a prelaminar component as well (black arrows). Although prelaminar thinning is a manifestation of neural tissue damage alone, we propose that laminar deformation can occur only in the setting of ONH connective tissue damage followed by permanent (fixed) IOP-induced deformation.

Figure 2.

 

Retrospectively generated TCA maps of both eyes of the three EG monkeys in right eye configuration. Red: posterior surface change; green anterior surface change relative to the baseline surface topology (before lasering). Images are arranged in temporal order (from initial test to the last one before death) for each eye. Note that although there is sporadic change in the contralateral normal eye, the onset and progression of ONH surface change is present in all three EG eyes.

Figure 3.

 

3-D delineation of ONH and peripapillary scleral landmark points within colorized, stacked-section, 3-D ONH reconstructions. ³¹ 3-D ONH reconstruction from aligned serial section images ³⁰ and 3-D delineation within each ONH reconstruction ³¹ are explained in detail in previous publications. (A) Sagittal histologic section through a representative normal monkey ONH showing the anatomy and (B) the associated neural canal and surface landmarks: Bruch’s membrane opening (BMO, red), anterior laminar insertion (ALI, yellow), anterior laminar/scleral surface (white), and the neural boundary (dark gray). (C) Forty serial digital radial sagittal slices (seen in D, each seven voxels thick) are served to the delineator at 4.5° intervals. (D) A representative digital sagittal slice, showing the marks for seven landmark surfaces and six pairs of landmark points, which are 3-D delineated by using linked simultaneous colocalization of the sagittal slice (shown) and the (E) transverse section image. (F) Representative 3-D point cloud showing all delineated points in a normal monkey ONH relative to the last serial section image (orbital optic nerve bottom, vitreous top). ^{31 32}

Figure 4.

 

Parameter definitions. (A) Representative horizontal digital and sagittal section images from the 3-D reconstruction of the left eye of monkey 1 (right, temporal; left, nasal) with the major structures delineated for this study: ILM (green dots); anterior laminar surface (white dots); neural canal wall (dark gray dots); BMO (red dots). BMO-zero reference plane (cyan line). (B) Continuous surfaces are fit to internal limiting membrane (pink line), neural canal wall (yellow line), and anterior laminar surface (dark blue). A cylindrical surface generated by projecting the BMO points up to the internal limiting membrane surface (red line) defines the outer border of the prelaminar tissue volume above the BMO reference plane. (C) Post-BMO total prelaminar volume (light green: a measure of the laminar or connective tissue component of cupping) is the volume beneath the BMO-zero reference plane, above the lamina cribrosa and within the neural canal wall; (D) post-BMO cup volume (pink: a measure of the clinical cup) is the volume (of the clinical cup) beneath BMO-zero reference plane but above the ILM; (E) prelaminar tissue volume (purple) is the volume above the lamina, inside the neural canal and below the internal limiting membrane within the cylinder defined by the BMO projection; and (F) prelaminar tissue thickness (purple with arrows) which is defined as the distance along a normal vector from each delineated anterior laminar surface point (white dots), to the internal limiting membrane surface (pink line).

Figure 5.

 

Construction and regionalization of one representative volumetric parameter: post-BMO cup volume. (A) ILM delineating points (green) are fit to a continuous surface (B), rendered in Geomagic software (Geomagic Research, Triangle Park, NC) with the BMO-zero reference plane (yellow line; C). Then the volume posterior to BMO-zero reference plane is isolated to generate the post-BMO cup (D). To quantify this volume, it is digitally sectioned at 3-μm (1.5 μm in the normal monkey) intervals (E) parallel to the BMO reference plane, generating parallel contour lines (F). Area within 15° radial regions is calculated for each contour and added to generate color-coded 15° regional volume (G). The color-coded volume along with BMO points (red) and ALI (yellow) are projected onto the BMO-zero reference plane and overlaid onto the CSLT reflectance image for clinical alignment (H).

Figure 6.

 

Regional post-BMO total prelaminar volume and difference maps by monkey. The 15° radial region maps for the normal eye and EG eye of the three EG monkeys are all presented in right eye configuration. Data for the right (OD) and left (OS) eyes of the normal monkey are mapped. Difference maps (EG − N in the three EG monkeys; left eye − right eye of the normal monkey) are plotted in right eye configuration on the EG or normal monkey eye. Colocalized BMO (red) and ALI (yellow) points are present in each eye map and are overlaid onto the difference maps on the right (solid line for EG eyes of EG monkeys or right eye of the normal monkey and dotted line for normal eyes of the EG monkeys or left eye of the normal monkey). The color-coded volume unit is cubic millimeters (scale at bottom). Note that volume expansion in all three EG monkey difference maps exceeds the physiologic intra-animal difference, as quantified (requiring lighter colors) in the normal monkey difference map (bottom right). These data, taken together, confirm the presence of a laminar component to cupping in all three EG eyes.

Figure 7.

 

Regional post-BMO cup volume and difference maps by monkey. The 15° radial region maps for the normal eye and EG eye of the three EG monkeys are all presented in right eye configuration. Data for the right (OD) and left (OS) eyes of the normal monkey (both in right eye configuration) are shown. Difference maps (EG − N for the three EG monkeys; left eye − right eye for the normal monkey) are plotted in right eye configuration on the EG (EG monkeys) or right normal monkey eyes. Colocalized BMO (red) and ALI (yellow) points are present in each eye image and are overlaid onto the difference maps (solid line for EG eyes of EG monkey or right eye of normal monkey and dotted line for normal eyes of EG monkey or left eye of normal monkey). The color-coded volume unit is cubic millimeters (scale at bottom). Note that volume expansion in all three EG monkey difference maps, exceeds physiologic intra-animal difference as quantified (requiring lighter colors) in the normal monkey difference map (bottom right). These data, taken together, confirm clinical expansion of the cup, (without identifying laminar or prelaminar components), in all three EG eyes.

Figure 8.

 

Regional prelaminar tissue volume and difference maps by monkey. The 15° radial region maps for the normal eye and EG eye of the three EG monkeys are all presented in right eye configuration. Data for the right (OD) and left (OS) eye of the normal monkey (both in right eye configuration) are mapped. Difference maps (EG − N for the three EG monkeys; left eye − right eye for the normal monkey) are plotted in right eye configuration on the EG (EG monkeys) or right normal monkey eye. Colocalized BMO (red) and ALI (yellow) points are present in each eye image and are overlaid onto the difference maps on the right (solid line for EG eyes of EG monkeys or right eye of normal monkey and dotted line for normal eyes of EG monkey or left eye of normal monkey). The color-coded volume unit is cubic millimeters (scale at bottom). Note that although prelaminar tissue volume expansion in EG monkey 1 exceeds the intra-animal difference as quantified in the normal monkey difference map (requiring lighter colors), EG monkeys 2 and 3 are indistinguishable from the normal monkey. These data, taken together, suggest no reduction in regional prelaminar tissue volume (and no prelaminar component to cupping) in any of the three EG eyes.

Figure 9.

 

Continuous prelaminar tissue thickness and difference maps by monkey. Continuous prelaminar tissue thickness maps for the normal eye and EG eye of the three EG monkeys, all presented in right eye configuration. Data for the right and left eyes of the normal monkey (both in right eye configuration) are mapped. Continuous difference maps (EG − N for the three EG monkeys; left eye − right eye for the normal monkey) are plotted in right eye configuration on the EG (EG monkeys) or right normal monkey eye. Colocalized BMO (red) and ALI (yellow) points are present in each eye image and are overlaid onto the difference maps (solid line for EG eyes of EG monkey or right eye of normal monkey and dotted line for normal eyes of EG monkey or the left eye of the normal monkey). The color-coded thickness unit is micrometers (scale at bottom). Note that although there is thinning of the central prelaminar tissues in all three EG eyes, its magnitude does not exceed the physiologic intereye difference demonstrated by the single normal eye, except very focally (midperipherally at the 2 o’clock position) in EG monkey 1. These data, taken together, confirm that with the exception noted in EG monkey 1, focal prelaminar tissue thinning is not present at the onset of CSLT-detected ONH surface change in the young adult monkey eye. These data further confirm (again, with the exception noted in EG monkey 1) that there was no focal prelaminar component to cupping in any of the three EG eyes.

Table 1.

 

Overall Volume and Thickness Data for Pooled Normal and EG Eyes by Delineator

Parameters	N			ΔEG
		D1	D2	D1	D2
Post-BMO cup volume (mm3)	0.026	0.025	0.092*	0.092*
Post-BMO total prelaminar volume (mm3)	0.183	0.185	0.207*	0.214*
Prelaminar tissue volume (mm3)	0.263	0.269	0.063*	0.068*
Prelaminar tissue thickness (μm)	102	102	67*	68*

 ΔEG, the EG eye measure minus the N eye measure; D1, delineator 1; D2, delineator 2.

*   P < 0.05, ANOVA.

Figure 10.

 

Overall (pooled) regional treatment effects (EG − N) for all four parameters by delineator. Statistically significant regional treatment effects (pooled EG eyes versus pooled normal eyes of the three EG monkeys) for each parameter are color mapped (for magnitude of effect) for delineator 1 (top) compared to delineator 2 (bottom). Volumetric parameters (cubic millimeters) are pooled into 24, 15° radial regions and prelaminar tissue thickness data (in micrometers) are pooled into 17 concentric regions. All data are plotted in right configuration. S, superior; SN, superonasal; N, nasal; IN, inferonasal; I, inferior; IT, inferotemporal; T, temporal; ST, superotemporal. See Figure 5in our previous report. ³² Only the colored regions achieved statistical significance (P < 0.05, ANOVA, Bonferroni correction for multiple comparisons). For prelaminar tissue thickness (in micrometers) the magnitude of treatment effect for each region is noted by number. Note that the regional treatment effect maps for the two delineators are virtually identical. These data confirm that although individual monkey behavior was variable, the principle axes of early EG pathophysiology in these three EG eyes were slightly counterclockwise to vertical (both for the connective tissues (post-BMO total prelaminar volume) and the neural tissues (prelaminar tissue volume and thickness).

Table 2.

 

Overall Volume and Thickness Data for Both Eyes of Each Monkey by Delineator

Parameters	Monkey 1							Monkey 2							Monkey 3							Normal*
		N			ΔEG			N			ΔEG			N			ΔEG			OD	OS-OD
		D1	D2	D1	D2	D1	D2	D1	D2	D1	D2	D1	D2	D2	D2
Post-BMO cup volume (mm3), †	0.054	0.053	0.127	0.13	0.013	0.012	0.058	0.057	0.010	0.009	0.090	0.088	0.014	0.006
Post-BMO total prelaminar volume (mm3), †	0.231	0.245	0.308	0.308	0.159	0.151	0.122	0.132	0.160	0.159	0.191	0.202	0.187	0.031
Prelaminar tissue volume (mm3), †	0.268	0.287	0.167	0.159	0.232	0.229	0.025	0.031	0.289	0.289	0.002	0.013	0.254	0.021
Prelaminar tissue thickness (μm)	110	110	82, ‡	81, ‡	100	100	47, ‡	51, ‡	95	96	58, ‡	61, ‡	95	−5, ‡

 ΔEG, the EG eye measure minus the N eye measure; D1, delineator 1; D2, delineator 2.

*  The normal monkey is only delineated by delineator 2.

†  No statistical testing within each monkey is done for these measurements because there are no repeated measures within a single delineation of one delineator.

‡   P < 0.05, ANOVA.

Figure 11.

 

Regional treatment differences in prelaminar tissue thickness (in micrometers) by monkey and delineator. Colored regions: statistically significant differences between the normal and EG eyes of each monkey for delineator 1 (top) and delineator 2 (bottom; P < 0.05, ANOVA). Color intensity represents the magnitude of difference as illustrated in the color bars and is also noted by number for each region. Note that regional treatment effects by monkey are virtually identical in pattern and magnitude for each delineator. Note also that the magnitude of physiologic, intra-animal difference in the normal monkey (bottom right; delineated by delineator 2 only) do not exceed 30 μm and are substantially less than the regional treatment effects in all three EG eyes.

Table 3.

 

Overall Volume and Thickness Data for Monkey 3 on 3 Delineation Days

Parameters	Day 1			Day 2			Day 3
		N	ΔEG	N	ΔEG	N	ΔEG
Post-BMO cup volume (mm3)*	0.009	0.088	0.010	0.085	0.011	0.083
Post-BMO total prelaminar volume (mm3)*	0.159	0.202	0.155	0.197	0.156	0.2
Prelaminar tissue volume (mm3)*	0.289	0.013	0.259	0.029	0.251	0.04
Prelaminar tissue thickness (μm)	96	61, †	95	60, †	95	61, †

 ΔEG, the EG eye measure minus the N eye measure.

*  No statistical testing within each day is done for these measurements because there are no repeated measures within a single delineation of one delineator.

†   P < 0.05, ANOVA.

Figure 12.

 

Summary of TCA surface change, cupping parameter difference maps, and regional axon loss maps for each monkey. Retrospectively generated, preeuthanatization TCA maps of ONH surface change for the three EG monkeys are shown. EG eye difference maps of post-BMO cup volume, post-BMO total prelaminar volume, and prelaminar tissue volume (in cubic millimeters) and prelaminar tissue thickness (in micrometers) for all three EG monkeys and the normal monkey are presented. Right: Total and regional axonal difference maps for all three EG monkeys. See Figures 2 6 7 8 and 9for an explanation of color coding within the TCA and cupping parameters. Within the axon count data, the magnitude of intra-animal difference (in percent) is shown in red. Regions that achieve statistical significance by ANOVA are shaded gray or black according to degree of significance. ³⁷ Taken together, these data suggest that the relationship between ONH surface change (left column), subsurface structural change (four middle columns), and optic nerve axon loss (right column) in the three EG eyes is complicated.

Figure 13.

 

Clinical cupping in the early glaucomatous monkey eye is laminar in origin without a significant prelaminar component: the clinical and pathologic importance of post-BMO total prelaminar volume expansion. Top: normal lamina cribrosa (nonhatched), scleral flange (hatched), prelaminar tissue (beneath the ILM; brown line), Bruch’s membrane (solid orange line), BMO-zero reference plane (dotted orange line), border tissue of Elschnig (purple line), and choroid (black circles). Bottom: changes in EG. We have reported the bowing of the lamina and peripapillary scleral flange and thickening of the lamina in these same EG eyes (gray shading). ³² These connective tissue changes underlie posterior deformation of the ONH and peripapillary retinal surface (dotted brown to solid brown ILM) that occurs in the setting of thickening (arrows), not thinning, of the prelaminar neural tissues (brown shading) in EG. EG eye expansion of all three volumetric parameters is depicted at the bottom. The interactions between these parameters are important. Although expansion of the cup and deformation of the surface are clinically detectable at this early stage of the neuropathy, because they occur in the setting of prelaminar tissue thickening, (not thinning), we believe that expansion of post-BMO total prelaminar volume (due to ONH connective tissue deformation) drives these findings. Thus, cupping in these three EG eyes is laminar in origin, without a significant prelaminar component (see Fig. 1and Discussion for details).

Supplementary Materials

Supplementary Figure S1 - (PDF) 

Supplementary Figure S2 - (PDF) 

Supplementary Figure S3 - (PDF) 

Supplementary Figure S4 - (PDF) 

Supplementary Figure S5 - (PDF) 

The authors thank Jonathon Grimm, Juan Reynaud, and Budd Hirons for assistance with software for volumetric and thickness quantification; Mike Roberts for consistent help with algorithm and programming; Grant Cull, Lin Wang, and Jack Cioffi for performance of optic nerve axon counting in the three EG monkeys; and Christopher Girkin for critically reading the manuscript.