Microvascular Density Is Associated With Retinal Ganglion Cell Axonal Volume in the Laminar Compartments of the Human Optic Nerve Head

Abstract

Purpose: To quantify associations between microvascular density and retinal ganglion cell (RGC) axonal volume in the laminar compartments of the human optic nerve head (ONH).

Methods: Eleven normal human ONHs were evaluated. Antibodies were used to label the vascular endothelium (factor VIII–related antigen/von Willebrand factor antibody) and RGC axons (neurofilament heavy antibody). Three-dimensional analysis of confocal scanning laser microscope images was used to study microvascular density and RGC axonal volume in the prelaminar, anterior lamina cribrosa, posterior lamina cribrosa, and retrolaminar compartments.

Results: Microvascular volume was significantly different between laminar compartments (P < 0.0083) and was greatest in the prelaminar region, occupying 11.7% of tissue volume. Microvascular volume per RGC axonal volume and cumulative capillary length per RGC axonal volume were significantly different between laminar compartments (all P < 0.0083). Both were significantly greater in the posterior laminar cribrosa (27.4% and 2.28 × 10⁻³ μm/μm³, respectively).

Conclusions: Microvascular density is closely coupled to RGC axonal volume in the ONH. The posterior laminar cribrosa is a site of high blood supply as evidenced by a greater ratio of microvascular density to RGC axonal volume. The greater percentage of tissue volume occupied by microvasculature in the prelaminar region may implicate it as a site where significant connections between the central retinal artery and short posterior ciliary arteries occur.

The human optic nerve head (ONH) is composed of approximately 1.2 million retinal ganglion cell (RGC) axons and is one of the most metabolically active sites in the central nervous system.¹ The energy demands of RGC axons along the course of the ONH are not uniform as evidenced by the marked variation in mitochondrial enzymes,² cytoskeletal proteins,^3–5 myelin proteins,^6,7 astrocytes,^8,9 and nitric oxide synthase systems¹⁰ between laminar regions. The highly compartmentalized nature of RGCs¹¹ and the marked heterogeneity in energy supply/demand relationships between laminar compartments are proposed to be an important reason why certain sites within the ONH are more vulnerable to injury and disease than others.¹² 

Within the central nervous system, continual communication between axons and surrounding vascular elements ensures that the supply of nutrients and oxygen is an uninterrupted and organized process.¹³ The optic nerve is part of the central nervous system, and it is likely that the topologic properties of the ONH microcirculation are specialized in accordance with the quantitative properties of RGC axons in each laminar compartment.^14,15 This hypothesis has never been specifically evaluated; however, such information is vital to our understanding of the pathophysiological mechanisms that govern ONH health and disease.¹⁶ Our recent report indicated an interrelationship of arterial supply to the retina, choroid, and ONH in the human eye.¹⁷ An in-depth understanding of the microvascular anatomy of the ONH may expand our knowledge of sites where vascular connections between the optic nerve, retina, and choroid occur. It may also provide valuable insights into biological relationships between microvascular and axonal compartmentalization in the human ONH. In this study we performed three-dimensional (3-D) analysis of confocal scanning laser microscope–derived images of the human ONH to investigate the above hypotheses. 

Methods

This study was approved by the human research ethics committee at The University of Western Australia. All human tissue was handled according to the tenets of the Declaration of Helsinki. 

Human Donor Eyes

A total of 11 human eyes from 11 donors were used for this study. All eyes were obtained from the Lions Eye Bank of Western Australia (Lions Eye Institute, Western Australia). Donor eyes used for this study had no documented history of eye disease. The demographic data and cause of death of each eye donor are presented in the Table. 

Table

 

Demographic Details of the Optic Nerve Donors

 

Tissue Preparation

Donor eyes were transferred in 4°C Ringer's solution after the removal of corneal buttons for the purpose of transplantation. The eye was dissected to expose the optic disc and 5 mm of optic nerve extending posteriorly from the sclera. The optic disc was dissected away from the globe, sparing a 1-mm rim of tissue that comprised retina, choroid, and sclera. The tissue was immersed in 30% sucrose for cryoprotection. The tissue was then embedded in optimal cutting temperature compound (OCT; Tissue-Tek 4583, product no. 62550-12; Sakura, Tokyo, Japan) in a plastic mold and snap-frozen in liquid nitrogen. 

Thirty-micron-thick sections were cut perpendicular to the long axis of the optic nerve on a cryostat machine (CM1850; Leica, Wetzlar, Germany) at −20°C. The tissue specimen was cut from the retina toward the posterior optic nerve in a consecutive manner. Every section was examined under a light microscope to determine from which laminar compartment it was derived. Using previously described histologic criteria,^4,7,11 each section was assigned to one of four laminar compartments. 

Prelaminar Region

The most anterior portion of the ONH is identified by the paucity of connective tissue structures in this region, and by the presence of retinal layers in adjacent peripapillary tissue. This region corresponds to the superficial nerve fiber layer described by Lieberman et al.¹⁸ 

Anterior Lamina Cribrosa Region

This is situated adjacent to the choroid within which many posterior ciliary arteries are observed around the ONH. This region is equivalent to the prelaminar layer described by Lieberman et al.¹⁸ This is also known as the choroidal part of the lamina cribrosa.¹⁹ 

Posterior Lamina Cribrosa Region

The fenestrated collagen sheets of the posterior lamina cribrosa are continuous with the connective tissue fibers of the sclera and form narrow openings for the transmission of axonal bundles. These dense collagen beams have high autofluorescence characteristics under ultraviolet wavelength. Lieberman et al.¹⁸ previously described this region as lamina cribrosa, and it is also known as the scleral part of the lamina cribrosa.¹⁹ 

Retrolaminar Region

This is situated posterior to the posterior lamina cribrosa region. The absence of dense collagen beams and the relative increase in ONH diameter from myelinated RGC axons characterize this region. Connective optic nerve (ON) peripheral tissues are continuous with the pia mater sheath of the meninges, and subarachnoid space around the ON is observed. This region is equivalent to the retrolaminar region described by Lieberman et al.¹⁸ 

All sections were organized in numerical fashion within each laminar compartment to identify their relative location along the longitudinal axis of the ON. We used only sections that demonstrated ON tissue from a single laminar region for immunohistochemical studies. Oblique sections that demonstrated tissue from multiple laminar regions were excluded. We selected four to six sections per laminar region from each eye for immunohistochemistry. 

Immunohistochemistry

All tissue sections were labeled for vascular endothelium and RGC axons using indirect immunofluorescence methods. The vascular endothelia were labeled using a rabbit polyclonal factor VIII–related antigen/von Willebrand factor (vWF) antibody-1 (RB-281-A, 1:80; Thermo Fisher Scientific, City, State, Country). Von Willebrand factor antibody was used because its staining pattern correlated with vessel lumens and was therefore ideal for quantitating vascular volume. Retinal ganglion cell axons were labeled using a chicken polyclonal anti-200 kD neurofilament heavy (NFH) antibody directed against the phosphorylated and nonphosphorylated NFH subunit (1:4000, ab4680; Abcam, Cambridge, UK). Vascular endothelium and RGC axons were visualized using a goat anti-rabbit Alexa Fluor 633 (1:200, A21071; Life Technologies, Rockford, IL, USA) and a goat anti-chicken Alexa Fluor 546 (1:200, A11040, Life Technologies), respectively. 

Tissue sections were fixed on glass slides with 4% paraformaldehyde for 10 minutes. All slides were washed with 0.1 M phosphate buffer three times for 30 minutes. Tissue was permeabilized and blocked with the mixture of 0.5% Triton X-100 (T8787; Sigma-Aldrich Corp., St. Louis, MO, USA) and 10% goat serum (G9023; Sigma-Aldrich Corp.) for 30 minutes. All primary antibodies were made into solution with 10% goat serum and 1% bovine serum albumin (A9647; Sigma-Aldrich Corp.), which was applied to each specimen on the glass slide and left overnight at 4°C. On the next day, the slides were washed three times over 30 minutes in 0.1 M phosphate buffer before secondary antibody incubation for 1 hour in room temperature. Following secondary antibody incubation, all specimens were washed three times over 30 minutes in 0.1 M phosphate buffer. The slides were then mounted in Hydromount (National Diagnostics, Atlanta, GA, USA) and immediately viewed under the confocal microscope. 

Confocal Scanning Laser Microscopy

Confocal images of antibody-labeled specimens were collected using the Nikon C1 confocal scanning laser microscope (Nikon, Tokyo, Japan), coupled with the EZ-C1 (v.3.20) software. As all the sections were labeled with vWF antibody-1 (Alexa Fluor 633) and NFH antibody (Alexa Fluor 546), fluorochromes on the sections were excited by laser beams of wavelength 635 and 546 nm, respectively. The laser power, iris, and duration of sampling time were unchanged during data acquisition. 

Separate images were captured from the prelaminar, anterior lamina cribrosa, posterior lamina cribrosa, and retrolaminar regions of the ONH for each antibody. Prior to acquisition of confocal stacks, autofluorescence images were captured for each section using the 405-nm laser excitation coupled with 445-nm emission filter. The autofluorescence properties of the ON section together with the histologic criteria described previously^4,7 were used to confirm the correct apportionment of each ON section to a specific laminar compartment. 

All the images were collected using Plan Apo x10 (NA 0.45; Nikon, Tokyo, Japan) dry lens (dimension of image being 1270 × 1270 μm or 1024 × 1024 pixels) at 1.15-μm z-interval steps and using a ×2 frame-scan average. Collection of four to eight separate confocal stacks was required to cover the entire diameter/area of the ON in cross section (Fig. 1). 

Figure 1

 

Methodology for three-dimensional image preparation. Multiple confocal image stacks were collected to ensure coverage of the entire laminar compartment. Each image stack comprised 27 slices with a step size of 1.15 μm between slices. In this example, four image stacks are required to cover the entire optic nerve head (A). All the stacks were subsequently montaged to create one large stack of the entire optic nerve head (B). The region of interest (ROI) was outlined and the choroid, sclera, dura, central retinal artery, and vein were manually traced and removed from the stack (C). Using isosurface function on Imaris software, a 3-D volume stack expressing all elements was then rendered (D). This stack could also be used to study the microvasculature and axons separately (Di). Inset in (D) denotes the region shown in (Di).

Volumetric Analysis

Z-stacks were montaged using the pairwise stitching plugin with the linear blending fusion method²⁰ available on ImageJ (v.1.49n; National Institutes of Health, Bethesda, MD, USA, http:/rsb.info.nih.gov/ij; in the public domain). This resulted in one large z-stack file for each ON cross section (Fig. 1B). Once montaged, structures outside the ON such as choroid, sclera, and pia mater were removed after manually tracing the outline of the ON. The central area of the ON devoid of neural tissue containing the central retinal artery and vein was also manually traced and removed (Fig. 1C). The following measurements were then calculated from the resultant stack: 

1.  
   Volume of region of interest (ROI): total volume of the stack after removal of the choroid, sclera, pia mater, central retinal artery, and central retinal vein.
2.  
   Microvascular volume: defined as the volume of the voxels occupied by vWF antibody staining (false-colored red in Fig. 1D).
3.  
   Axonal volume: defined as the volume of the voxels occupied by NFH antibody staining (false-colored blue in Fig. 1D).

The area of the ROI for each slice of the z-stack was determined using the Analyze function on ImageJ, which was then multiplied by the thickness of the stack (μm) to calculate the volume of the ROI (μm³). All other volume measurements were determined using the Imaris software (v.7.6.2; Bitplane, Zurich, Switzerland). A single composite 3-D model was reconstructed from the montaged stack of confocal images using the 3D surpass view on Imaris (Fig. 2). The voxel size of 1.24 × 1.24 × 1.15 μm was applied to all the stacks. Isosurface was built for microvasculature and axons from the 3-D reconstruction of the stack. Consistent range of parameters for smoothing and thresholding was used for each structure throughout the isosurface construction to ensure an accurate rendering of the original tissue staining pattern. The total volume of staining present in the rendered cube of tissue was then calculated in μm³. The volume of voxels stained with vWF antibody (microvascular volume) was then divided by the ROI volume to determine the percentage of the ROI volume that was occupied by vWF antibody. The microvascular volume was also divided by the axonal volume (the volume of voxels stained with NFH antibody) to determine the ratio of microvascular volume to axonal volume in the stack. These calculations can be summarized as follows: 

Figure 2

 

Two-dimensional and three-dimensional characteristics of the microcirculation and retinal ganglion cell axons in the different laminar compartments of the human optic nerve head. Blood vessels were labeled with antibodies to von Willebrand factor antigen and false-colored red. Retinal ganglion cell axons were labeled with antibodies to neurofilament heavy subunit and false-colored blue. Two-dimensional, low-magnification images of each laminar compartment are shown in (A–D). Three-dimensional reconstruction of microvascular and axonal structures (merged and individual images) for each laminar compartment are provided (Ai–Di). Scale bar: 300 μm. Insets (A–D) denote the regions illustrated in (Ai–Di), respectively. LC, lamina cribrosa.

1.  
   Microvascular density per tissue volume (Fig. 3A) = volume of voxels stained with vWF antibody/ROI volume × 100 (%).
2.  
   Microvascular density per RGC axonal volume (Fig. 3B) = volume of voxels stained with vWF antibody/volume of voxels stained with neurofilament antibody × 100 (%).

Figure 3

 

Comparisons of the ratio of microvascular volume to ROI volume (A) and the ratio of microvascular volume to the RGC axonal volume (B) between laminar compartments. For the ratio of microvascular volume to ROI volume (A), the differences between laminar compartments were significant (all P < 0.0083). Similarly, the ratio of microvascular volume to the RGC axonal volume (B) showed significant differences between all laminar compartments (all P < 0.0083) except for the comparison between the prelaminar and the anterior lamina cribrosa compartments (P = 0.284). LC, lamina cribrosa.

We selected four to six sections from each laminar compartment for 3-D reconstruction and volume rendering. This resulted in approximately 120- to 180-μm thickness of each laminar compartment being studied. Sections from the posterior lamina cribrosa were subanalyzed by quadrants. The nerve cross section was divided into four quadrants (superior, inferior, temporal, and nasal) by intersecting diagonal (45°) lines using Oblique Slicer on the Imaris software. Microvascular density per RGC axonal volume was calculated in each quadrant. All images for the manuscript were prepared using Adobe Photoshop CS4 Extended (version 11.0; Adobe Systems, Inc., San Jose, CA, USA) and Adobe Illustrator CS4 (version 14.0; Adobe Systems, Inc.). 

Microvascular Length Analysis

The length of all blood vessels was measured in three dimensions to account for the contribution from smaller capillaries that could have a relatively small representation in the total microvascular volume. Montaged z-stacks were rendered to produce a 3-D reconstruction as described above. Using Filament Tracer's advanced tracing mode on Imaris, all vessels were directly traced onto the 3-D volume image with automatic z-depth placement. The use of this AutoDepth function on Imaris allowed more accurate length measurement in a 3-D image because blood vessels in the ONH travel longitudinally as well as transversely, forming a complex microvascular network (shown as red structures in Fig. 4). The length of each vessel segment was automatically summated on Imaris. This total vessel length measurement (μm) was expressed as a ratio of the RGC axonal volume (μm³) in the 3-D stack, as shown in Figure 5. 

Figure 4

 

Microvascular length analysis using three-dimensional images. Blood vessels traced with automatic z-depth placement onto a 3-D reconstructed image in different laminar compartments are provided (A–D). The morphology and appearance of the microcirculation in the prelaminar (A), anterior LC (B), posterior LC (C), and retrolaminar (D) regions after filament tracing are provided. LC, lamina cribrosa.

Figure 5

 

Comparisons of the ratio of cumulative microvascular length to the RGC axonal volume between laminar compartments. Significant difference was demonstrated between all laminar regions (P < 0.0083) except for the comparison between prelaminar and anterior lamina cribrosa (P = 0.858).

Statistical Analysis

All statistical testing was performed using commercial software (R; R Foundation for Statistical Computing, Vienna, Austria). The residuals of all data from the linear models were tested for normality with Kolmogorov-Smirnov test. One-way analysis of variance (ANOVA) testing was performed using linear mixed-effects modeling to test differences in volume ratios (volume/volume or length/volume ratios) between all laminar compartments. In the fixed effects part of the models, the volume ratio was the response variable, and lamina compartment, sex, and age were the explanatory variables. The random factor was donor eye to account for correlation between multiple measurements from the same eye. The models testing between four laminar compartments generated six comparisons, so a P value of <0.0083 was considered significant following Bonferroni correction. Results are expressed as mean ± SE. 

Results

General

The mean age of donors was 55.6 ± 11.1 years. We examined five right eyes and six left eyes from a total of nine male and two female donors. The average postmortem time before eyes were enucleated was 14.8 ± 6.7 hours (Table). The total length of ON analyzed from 11 donor eyes was 6.51 mm. The total volume of ON analyzed was 10.2 mm³, from which 217 3-D reconstructions and 651 volume renderings were generated. This included 150 volume measurements (0.8 mm³) from the prelaminar compartment, 174 volume measurements (1.5 mm³) from the anterior lamina cribrosa compartment, 177 volume measurements (3.1 mm³) from the posterior lamina cribrosa compartment, and 150 volume measurements (4.8 mm³) from the retrolaminar compartment. 

Microvascular and Axonal Volume Measurements

The organization of the microvasculature in the different laminar compartments of the ONH is shown in Figure 2. The microvascular density per tissue volume was greatest in the prelaminar compartment (11.7 ± 0.3%) followed by the anterior and posterior lamina cribrosa compartments (10.3 ± 0.3% and 8.4 ± 0.2%, respectively) and lowest in the retrolaminar compartment (3.3 ± 0.1%). The differences between laminar compartments were significant (all P < 0.0083) (Fig. 3A). 

The microvascular density per axonal volume was greatest in the posterior lamina cribrosa compartment (27.4 ± 1.2%) followed by the anterior lamina cribrosa (20.0 ± 0.6%) and the prelaminar compartment (18.8 ± 0.5%). The retrolaminar compartment showed the least microvascular density per axonal volume (6.7 ± 0.2%). Significant differences between all laminar compartments (all P < 0.0083) were demonstrated except for the comparison between the prelaminar and the anterior lamina cribrosa compartments (P = 0.284; Fig. 3B). 

There was no difference in microvascular density per axonal volume between any of the four quadrants (all P > 0.449): superior (23.2 ± 1.5%), inferior (23.1 ± 1.5%), temporal (23.4 ± 1.4%), and nasal quadrants (23.8 ± 1.6%). 

Microvascular Length Measurement

A representative example of blood vessel tracing of a 3-D volume stack for each of the laminar compartments is shown in Figure 4. Quantitative analysis of cumulative capillary length for each laminar compartment is provided in Figure 5. The ratio of total vessel length to RGC axonal volume was greatest in the posterior lamina cribrosa compartment (2.28 ± 0.10 × 10⁻³ μm/μm³) followed by the anterior lamina cribrosa (1.17 ± 0.08 × 10⁻³ μm/μm³) and the prelaminar compartment (1.12 ± 0.06 × 10⁻³ μm/μm³). The retrolaminar compartment showed the lowest vessel length to RGC axonal volume ratio (0.66 ± 0.02 × 10⁻³ μm/μm³). The ratio of total vessel length to RGC axonal volume was significantly different between laminar compartments (all P < 0.0083) except for the comparison between the prelaminar and the anterior lamina cribrosa compartments (P = 0.858). 

Discussion

This study investigated the association between microvascular density and RGC axonal volume in the different laminar compartments of the human ONH. The major findings are as follows. (1) There is significant variation in microvascular density and RGC axonal volume between laminar compartments; (2) microvascular density per tissue volume is the greatest in the prelaminar compartment; (3) the posterior laminar cribrosa contains the greatest microvascular density per RGC axonal volume as well as the greatest microvascular length per RGC axonal volume. 

The concept of biological “microcompartmentation” was introduced by Friedrich²¹ in 1984 following the observation of restricted molecular mobility inside living cells. This concept was expanded by the work of Kellermayer et al.,²² who showed that K⁺ and [³⁵S]-methionine–labeled proteins remained compartmentalized to the cytoskeleton despite detergent-based removal of intracellular lipid membranes. Minaschek et al.²³ and Zaccolo and Lefkimmiatis^24,25 provided further evidence that enzyme systems and signaling pathways are compartmentalized within eukaryotic cells, thereby consolidating a paradigm shift in our understanding of cellular and molecular biology. Biochemical and structural compartmentalization is expected to confer a functional advantage to RGC axons: a neuronal cell of unique morphology that traverses a range of tissue pressures and metabolic environments during their trajectory (approximately 50 mm) from the eye to the brain.²⁶ The notion that RGCs are compartmentalized, however, is not new. Quigley²⁷ suggested that the RGC in toto should be considered as six distinct compartments while the work by Yu and colleagues¹² proposed a four-compartment model composed of an intraretinal compartment and three laminar compartments. Whitmore et al.²⁸ postulated that neuronal death programs that underlie glaucomatous optic neuropathy are also compartmentalized within the ONH. 

The microcirculation is a critical source of energy for RGC axons, and our understanding of if and how the microcirculation is organized to meet the specific energy demands of each laminar compartment is limited. In this study we demonstrate a strong association between RGC axonal volume and microvascular density in each laminar compartment of the human ONH. Both measures of microvascular density (microvascular volume per axonal volume and cumulative capillary length per axonal volume) evaluated in this study were greatest in the posterior lamina cribrosa. The posterior laminar cribrosa is a critical site for RGC axonal injury in glaucoma²⁹ and also for venous endothelial injury in central retinal vein occlusion.^30,31 Our group previously demonstrated the pressure gradient between the intraocular and intracranial compartment is greatest in the posterior lamina cribrosa, and proposed that it may act as one of the factors causing damage to RGC axons.^26,32 Other factors are also likely to be drivers of RGC axonal injury in the posterior lamina cribrosa such as activation of microglial cells and macrophages.^33,34 Changes in the concentration of myelin basic proteins may also contribute to RGC axonal injury as the posterior lamina cribrosa represents the point of transition between myelinated and unmyelinated ON.^28,35 Dense collagenous plates,^19,36 a relatively higher density of mitochondrial cytochrome c oxidase enzyme,² and a relatively greater concentration of endothelial cell nitric oxide synthase enzymes¹⁰ are thought to be required to tolerate the greater forces within the posterior lamina cribrosa. This study demonstrates that a greater concentration of microvasculature may be another anatomic specialization that supports the relatively greater energy demands of the posterior laminar cribrosa. Our previous work showed that the density of radial peripapillary capillaries is correlated to nerve fiber layer thickness in the human eye.³⁷ Taken together with the present study, it appears that coupling between RGC axons and the microcirculation is an anatomic specialization that involves the intraretinal and ON portions of RGC axons. Our findings concerning the ON are consistent with what has previously been reported in the brain. Cavaglia et al.³⁸ evaluated neuronal–vascular relationships in rodent brains and demonstrated that regions of highest synaptic activity and metabolic demand, such as the parietal cortex and hippocampal CA1 region, were characterized by higher levels of vascularization. Zhang et al.³⁹ measured cerebral microvascular plasma perfusion of rats in three dimensions using a similar methodology to the present report. They found that 2% to 3% of cerebral tissue volume is occupied by vasculature, a finding that is similar to our vascular density measurement of the myelinated retrolaminar region of the ONH (approximately 3.3%). However, we found much greater vascular density in the unmyelinated laminar regions (prelaminar, anterior, and posterior lamina cribrosa), ranging between 8.4% and 11.7%. These unmyelinated regions have greater energy demands for signal transmission and axonal transport, particularly in the lamina cribrosa where pressure gradient is predicted to be greatest.²⁶ 

The studies by Radius¹¹ and Radius and Gonzales³⁶ demonstrated a regional variation of the ONH anatomy wherein the density of connective tissue elements was greater in nasal–temporal as compared with inferior and superior quadrants of the ONH. This variation was contrasted with preferential susceptibility patterns of ONH in glaucoma, supporting the hypothesis that greater structural integrity may yield relatively more resistance to a given pressure elevation. In this study, we did not find significant variation of the microvascular density per axonal volume between quadrants at the level of posterior lamina cribrosa. 

There is marked interindividual variation in human ONH vascular anatomy. From previous microvascular corrosion casting studies, it has been well established that branches of short posterior ciliary arteries (SPCAs) are main contributors to the blood supply of the ONH.^40,41 The central retinal artery (CRA) mainly supplies the prelaminar region, commonly referred to as the superficial nerve fiber layer.^18,40 Olver et al.⁴¹ first reported branches of the CRA anastomosing with ON capillaries in the retrolaminar region approximately 1.5 mm posterior to the lamina cribrosa, and no such anastomosis between the CRA and the SPCAs was observed anterior to this point. Apart from this observation in the relatively posterior portion of the retrolaminar region, traditionally, the CRA and the SPCAs have been considered independent vascular systems that do not communicate with each other in the ONH.^42,43 Our recent study using dual microvascular cannulation of the CRA and the SPCAs found that the vascular beds of the CRA and SPCAs may potentially communicate in the anterior ONH.¹⁷ 

In conclusion, glaucoma,⁴⁴ central retinal vein occlusion,⁴⁵ and the spectrum of ischemic optic neuropathies⁴⁶ are major causes of visual morbidity globally. An in-depth understanding of the quantitative properties of the ONH microcirculation is critical to defining pathogenic mechanisms that underlie these diseases.^16,47,48 This report illustrates that the quantitative properties of the microcirculation are closely coupled to RGC axonal load in each laminar compartment. Therefore, the results of this report provide important new information about the physiological mechanisms that govern RGC axonal function and the vulnerability of specific laminar compartments to injury. An important limitation of this study is the restricted sample size that underwent postmortem analysis. Another limitation of this study lies in the uneven representation of neural tissues in posterior lamina cribrosa sections due to posterior bowing of lamina cribrosa. Despite our careful sectioning and our studying sections only from a single laminar region, the bowed morphology of the posterior lamina cribrosa may have resulted in greater variability of measurements from this region. Addressing these limitations as well as evaluating optic nerves of diseased eyes using similar techniques will further enhance the findings of this report. 

Acknowledgments

The authors thank staff from the Lions Eye Bank of Western Australia, Lions Eye Institute for provision of human donor eyes, and staff from Donate Life, the Australian agency for organ and tissue donation, who facilitated the recruitment of donors into the study by referral and completion of consent processes. 

Supported by the National Health and Medical Research Council of Australia and the Ophthalmic Research Institute of Australia. 

Disclosure: M.H. Kang, None; M. Suo, None; C. Balaratnasingam, None; P.K. Yu, None; W.H. Morgan, None; D.-Y. Yu, None