|Home | About | Journals | Submit | Contact Us | Français|
Outer retinal injury has been well described in glaucoma. To better understand the source of this injury, we wanted to develop a reliable model of partial retinal ganglion cell (RGC) axotomy.
Endodiathermy spots were placed along the inferior 180° adjacent to the optic nerve margin in the right eyes of four cynomolgus monkeys. Fluorescein angiography, spectral domain optical coherence tomography (SD-OCT), and multifocal electroretinography (mfERG) were performed at various intervals. Two animals were sacrificed at 3 months. Two animals were sacrificed at 4 months, at which time they underwent an injection of fluorescent microspheres to measure regional choroidal blood flow. Retinal immunohistochemistry for glial fibrillary acidic protein (GFAP), rhodopsin, S-cone opsin, and M/L-cone opsin were performed, as were axon counts of the optic nerves.
At 3 months, there was marked thinning of the inferior nerve fiber layer on SD-OCT. The mfERG waveforms were consistent with inner but not outer retinal injury. Greater than 95% reduction in axons was seen in the inferior optic nerves but no secondary degeneration superiorly. There was marked thinning of the nerve fiber and ganglion cell layers in the inferior retinas. However, the photoreceptor histology was similar in the axotomized and nonaxotomized areas. Regional choroidal blood flow was not affected by the axotomy.
Unlike experimental glaucoma, hemiretinal endodiathermy axotomy (HEA) of the RGCs produces no apparent anatomic, functional, or blood flow effects on the outer retina and choroid.
Blindness in glaucoma results from the loss of retinal ganglion cells (RGCs). Histologic examination of eyes from patients with glaucoma, and animals with experimental glaucoma, show decreased numbers of RGCs but relatively intact outer retinal layers. However, careful analysis has shown loss of photoreceptors in human traumatic angle-closure glaucoma1,2 as well as both photoreceptor loss and cone swelling in chronic open-angle glaucoma.3 Although initially controversial,4,5 subsequent quantitative anatomic studies using multiphoton microscopy in human eyes with chronic glaucoma have confirmed cellular losses in both the outer and inner nuclear layers.6,7 Further work at the molecular level in human glaucoma and experimental glaucoma in monkey has demonstrated decreased cone opsin messenger RNA.8 Outer retinal functional deficits have been reported on multifocal electroretinography (mfERG) testing in rhesus and cynomolgus monkeys with experimental glaucoma.9 In the living eye, decreased cone signal is evident using adaptive optics in human patients diagnosed with glaucoma.10,11 Functional and anatomic evidence now suggests that outer retinal injury also occurs in rodent models of glaucoma12–15—including cone loss in rats with laser-induced ocular hypertension (Ortin-Martinez A, et al. IOVS 2012;53:ARVO E-Abstract 2488).
Although outer retinal effects of glaucoma have been documented, the etiology of these changes is unknown, as is the relevance to the cellular mechanisms of RGC loss in this disease. One possibility is that the outer retinal injury is secondary to RGC degeneration and that RGC degeneration is the result of Wallerian-like retrograde deterioration caused by injury at or near the optic nerve head (either mechanical or vascular). The retrograde degeneration hypothesis predicts that regional RGC death and outer retinal injury should occur in close spatial proximity. Indeed, this correlation was reported in postmortem human eyes with glaucoma using a multiphoton approach to quantify cellular nuclei in various retinal layers,6,7 as well as in the living eye with adaptive optics.10,11 Studies of experimental optic nerve transection in rats16 and mice17 have also found outer retinal injury. However, both types of studies have potential pitfalls. Optic nerve transection is complicated by the intimate physical proximity of the ophthalmic artery and the medial and lateral ciliary arteries.18 Thus, any attempt at surgical transection of the nerve has the potential of interfering with the choroidal circulation and injuring the photoreceptors by unintended ischemic effects. In our experience, such vascular interference occurred in half of the animals in which we attempted the procedure, despite having a skilled oculoplastic surgeon to perform an orbitotomy and a neuroophthalmic surgeon to do the nerve transection (Nork TM, et al. IOVS 2003;37:ARVO E-Abstract 2698). Also, in both experimental and human glaucoma, we have found widespread changes in cellular morphology and opsin regulation within photoreceptors in addition to marked local regional variability,8,9 which could weaken spatial correlation between nerve fiber bundle pathology and the outer retina.
To circumvent some of the problems with investigating correlative effects between RGC loss and outer retinal changes, we first attempted to create a model of RGC axotomy produced by intense laser photocoagulation around the optic nerve.19 The problem with this model approach is that the laser energy is first absorbed by the retinal pigment epithelium and transferred by conduction to the retina, causing coagulative necrosis. To reach the inner retinal layers, intense laser spots of large size are needed that cause unacceptably widespread retinal damage (Ver Hoeve JN, et al. IOVS 2012;53:ARVO E-Abstract 3513). One of the objectives of the current study was to develop a model of partial RGC axotomy that does not interfere with choroidal blood flow and causes minimal collateral retinal damage. Here we report that an endodiathermy procedure achieved a localized axotomy of the RGCs without widespread damage and allowed us to test whether the outer retina is affected by RGC degeneration.
Two female (animal numbers Cy1 and Cy3) and two male (Cy2 and Cy4) cynomolgus macaques (Macaca fascicularis) were used for the collection of data in this study. All of the experimental methods and techniques adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the University of Wisconsin-Madison's Animal Care and Use Committee. Before experiments began, all four animals were determined to be ocularly normal and healthy. The age and weight of the animals at the time of endodiathermy was 5.6 ± 0.8 (mean ± standard deviation) years and 4.5 ± 0.9 kg.
Two adult cynomolgus monkeys (one male and one female) underwent unilateral laser axotomy. A ring of laser burns was placed approximately 0.5 disc diameters from the rim of the optic nerve. A 532 nm diode laser and slit-lamp delivery system (OcuLight GL; Iridex Corp., Mountain View, CA) was employed to deliver laser light through a Kaufman-Wallow contact lens (Ocular Instruments, Inc., Bellevue, WA) applied to the cornea. The spots were all 500 ms in duration and were all either 200 or 500 μm in size. The power settings varied from 200 to 500 mW. Burns were contiguous except for small regions occupied by blood vessels and extended 360° around the nerve head. The clinical appearance corresponded to grade 3 of Wallow, indicating a full-thickness burn.20
All four cynomolgus macaques underwent a single HEA procedure in the inferior portion of the right eye only. The animals were preanesthetized with ketamine (15 mg/kg, intramuscular [IM]) before being intubated and anesthetized with an oxygen/isoflurane mix (3%–5% isoflurane for induction, 1%–3% isoflurane for maintenance) to sustain a deeper level of anesthesia during the procedure. Pupils were dilated using topical 1% tropicamide and 2.5% phenylephrine hydrochloride drops. The corneal surfaces were cleaned and sterilized using a 2.5% betadine solution, and any subsequent rinsing occurred with sterile balanced salt solution. The animal's head was supported in a holding device (not a stereotaxic apparatus) to maintain a stable position. Proparacaine hydrochloride (0.5% ophthalmic solution) was administered as a local anesthetic. A wire speculum was used to retract the eyelids. Two 25-gauge cannulae were inserted 4 mm posterior to the corneal limbus in the 2 o'clock and 10 o'clock meridians through the conjunctiva and sclera using trocars. A fiber optic light was passed through one 25-gauge cannula, and a sharp-tipped endodiathermy probe was passed through the other. A flat contact lens was placed on the cornea. Hydroxypropyl methylcellulose ophthalmic solution (2.5%) served as an optical couple. The retina was then visualized with a stereo operating microscope. Contiguous endodiathermy spots were placed along the inferior 180° adjacent to the optic nerve margin. The individual diathermy spots were created using enough energy to cause retinal whitening. Endodiathermy was not applied directly over the large retinal vessels. Post-HEA, cannulae were removed and application of sutures was not required. Subconjunctival injections of an antibiotic (cefazolin, up to 25 mg/kg) and a corticosteroid (triamcinolone acetonide, up to 20 mg) were given at the end of the procedure. A mild analgesic (flunixin meglumine, 1.5 mg/kg, IM) and an antibiotic (cefazolin, 25 mg/kg, IM) were also administered systemically on the same day as, and daily for 2 days subsequent to, the HEA procedure.
Prior to HEA, baseline fundus photography, fluorescein angiography, and spectral domain optical coherence tomography (SD-OCT) data were collected. Fundus photography and fluorescein angiography data were obtained using the TRC 50EX retinal camera (Topcon Corp., Tokyo, Japan) and captured using an EOS 5D Mark 2 camera (Canon, Tokyo, Japan) connected to the retinal camera. Sodium fluorescein (10%) was used for angiography. SD-OCT data were collected using a Cirrus HD-OCT 4000 (Carl Zeiss, Oberkochen, Germany) and were analyzed using the accompanying software. Immediately following endodiathermy axotomy, fundus photographs were taken. Fundus photographs were also obtained at 3 months postoperation for all four animals and at 4 months postoperation for two of the animals. Fluorescein angiography and SD-OCT measurements were carried out at 3 months for all four animals.
The mfERG testing procedure was performed as previously described.21 Briefly, the animals were preanesthetized with ketamine (15 mg/kg, IM). Pentobarbital sodium was used to induce each animal into a deeper level of anesthesia. The induction dose of pentobarbital sodium was 15 mg/kg intravenous (IV), and was followed by supplemental doses within the range of 1 to 10 mg/kg IV as needed in order to continue the restriction of eye movements. The intraocular pressures for each eye were measured with a rebound tonometer (TonoVet; Paragon Medical, Coral Springs, FL) prior to and following induction with pentobarbital sodium to ensure that pressures remained within a physiologically normal range. Heart rate and blood oxygen saturation were monitored continuously with a pulse oximeter, and body temperature and respiratory rate were also regularly monitored to ensure physiological stability during testing. Wire specula were used for lid retraction. Topical 1% tropicamide and 2.5% phenylephrine hydrochloride were used to induce pupillary mydriasis and cycloplegia. ERG-jet contact lens electrodes were placed on the corneas with 2.5% hydroxypropyl methylcellulose. Reference electrodes for each corneal contact lens electrode consisted of subdermal needles that were inserted at the ipsilateral outer canthus. The VERIS Science 4.9 system (Electro-Diagnostic Imaging, Inc., San Mateo, CA) was applied for generating stimuli, collecting data, and preliminary analyses. The VERIS auto-calibration system was applied for calibration of the stimulus monitor at the outset of the mfERG studies. The visual stimulus consisted of 241 unstretched hexagonal elements that were displayed on a Philips model MGD403 monochromatic monitor (Koninklijke Philips Electronics NV, Eindhoven, Netherlands). The VERIS fast sequence (binary maximum-length sequence cycle of 215-1) was used with a frame rate of 75 Hz (13.3 ms per frame). Maximum and minimum luminances of the display were 200 cd/m2 and ~1 cd/m2, with a mean luminance of ~100 cd/m2. The sampling rate of the signal was 1200 Hz (0.83 ms). The eyes were refracted for the 20 cm viewing distance. A refracting lens was positioned in front of the tested eye, and an opaque occluder was employed to obstruct the untested eye. A reversing ophthalmoscope with corner cube was used to align the stimulus display with the visual axis. The visual angle subtended by the entire stimulus was ~80°, and each hexagonal element subtended ~5.2°. Testing order of the eyes was counterbalanced; that is, the eye tested first in a given session was the one tested second in the previous session. Multifocal ERG recordings were collected weekly for 1 month (four separate test periods) postoperatively, then again at the 2- and 3-month time points. For two animals (Cy3 and Cy4), ERG data were also collected at the 4-month time point.
Each 241-element first-order kernel (K1) trace array was visually inspected to locate the hexagonal element with the largest N1-P1 response density amplitude (referred to subsequently as “amplitude”). Two seven-element hexagon groups were then delineated at locations immediately superior and immediately inferior to the foveal stimulus element (VERIS Science 4.1). The processed data (from the VERIS Science 4.1 software) were exported for further analysis to a task-specific routine using MATLAB (The MathWorks, Inc., Natick, MA). Root-mean-square (RMS) determinations were made for the 9 to 35 and 40 to 70 ms epochs of the K1 as well as for the 1 to 80 ms epoch of the second-order (first slice) kernel (K2.1) for the two separate hexagon groups. RMS values were then averaged for each hexagon group, and compared as a ratio between eyes for the three epochs using paired t-tests. The amplitudes for the N1, P1, N2, and P2 components of the K1 and for the p1, n2, and p2 components of the K2.1 averaged waveforms for each of the hexagon groups were derived from the same hexagon groups as for the RMS values. The calculated waveform components were compared for each hexagon group between baseline and postendodiathermy recordings using a paired t-test. The level of significance for all statistical analyses was defined as P < 0.05.
Two of the four animals (Cy3 and Cy4) were euthanized at 4 months following endodiathermy, at which time they underwent an injection of fluorescent microspheres into the left ventricle of the heart to determine the effects of axotomy on regional choroidal blood flow. Approximately 10 million fluorescent polystyrene spheres (15.5 ± 0.42 μm, mean ± SD diameter) (Invitrogen, Carlsbad, CA) were injected. The process by which this was accomplished is described in a previous paper.22 Briefly, the animals underwent preanesthesia with ketamine (15 mg/kg, IM) before being intubated and anesthetized with an oxygen/isoflurane mix (3%–5% isoflurane for induction, 1%–3% isoflurane for maintenance) in order to sustain a deeper level of anesthesia during the procedure. Buprenorphine (0.01–0.03 mg/kg, IM) also was administered in advance of incisional procedures. A bilateral cut down of the femoral arteries was performed to place two intra-arterial catheters; one catheter was used to collect reference blood samples, while the other was connected to a continuous intra-arterial blood pressure monitor and used to collect samples for analyzing blood chemistry in assessing physiological stability. Blood chemistry samples were collected immediately after the intra-arterial catheters were placed, after the chest cavity was opened, and after the completion of the fluorescent microsphere injection but before euthanasia. All blood chemistry samples were analyzed using an i-Stat blood gas analyzer (Abbott Point of Care, Princeton, NJ). An incision was made along the ventral midline of the animal from approximately the area of the sternum toward the level of the diaphragm. A rib spreader was used to retract the rib cage to permit visualization of the heart. A direct injection of the microspheres into the left ventricle of the heart was carried out over a period of approximately 40 seconds. During the injection and for 2 minutes immediately following it, an arterial reference blood sample (1 mL/min) was collected. The animals were immediately euthanized with an overdose of pentobarbital sodium (>50 mg/kg) injected directly into the heart or intra-arterially. Following sacrifice, the animals' eyes were immediately removed and immersion fixed in cold 4% paraformaldehyde in 0.1 M sodium phosphate buffer for histology.
The anterior segment of the globe was removed and the retina carefully dissected away from the choroid and retinal pigment epithelial layer. Potassium permanganate and oxalic acid were used to bleach the choroidal and RPE melanin.22 Six radial incisions were made through the peripheral choroid and sclera, and the tissue was coated with 2.5% hydroxypropyl methylcellulose before flattening between two glass slides. An epifluorescence microscope was used to obtain digital images. The x- and y-coordinates of each sphere were obtained using NIH ImageJ (National Institutes of Health, Bethesda, MD). The microspheres were then counted in 0.25 mm (0.0625 mm2) bins using a task-specific routine written in MATLAB (The MathWorks, Inc.). Three-dimensional plots of microsphere density, wherein density is proportional to choroidal blood flow, were created. A three-dimensional smoothing function was applied (locally weighted scatterplot smooth [LOESS], fourth-order polynomial) to produce a contour graph using Sigma Plot 11.0 (Systat Software, Inc., San Jose, CA). Comparisons were made between eyes by subtracting the plot values of the HEA eyes from those of the control eyes.
Following euthanasia and enucleation, the eyes of all four animals were fixed (4% paraformaldehyde) overnight at 4°C. They were then placed in 0.1 M sodium phosphate buffer (pH 7.6) for long-term storage at 4°C. Segments of retina from fellow control and HEA eyes of the four animals, and from the axotomized and nonaxotomized retinal locations of the HEA eyes, were collected for embedding in either glycol methacrylate or Durcupan epoxy resin (Fluka; Sigma-Aldrich, St. Louis, MO). Sections (1 μm) were cut with the use of a retracting microtome. They were cut either in a standard (radial) orientation or tangentially with respect to the plane of the retina at the level of the photoreceptor inner segments. Some sections were stained with Richardson's solution.23 On the unstained sections, immunohistochemical (IHC) procedures were performed as described previously.24 Prior to immunoreaction, the Durcupan resin was removed according to an established procedure.25 Once rehydrated, the sections were incubated in trypsin at 37°C for 8 minutes. After cold double deionized water rinses, the sections were incubated overnight (4°C) with primary antibodies for glial fibrillary acidic protein (GFAP) (1:2000, rabbit polyclonal; Dako Corp., Carpinteria, CA), for rhodopsin (1:2000, mouse monoclonal; MAB5316 Chemicon/Millipore, Billerica, MA), or for M/L-cone opsin (1:2000, rabbit polyclonal; AB5405 Chemicon/Millipore). After washing, the sections were incubated with biotinylated anti-mouse IgG and biotinylated anti-rabbit IgG, as appropriate, for 30 minutes and then incubated for 1 hour at room temperature with avidin–biotin complexes (Vectastain ABC Kit Elite PK-6102 with the rhodopsin primary and Vectastain ABC Kit Elite PK-6101 with the GFAP and cone opsin primaries; Vector Laboratories, Burlingame, CA). 3, 3′-Diaminobenzidine tetrahydrochloride (Sigma-Aldrich) served as the chromogen.
Segments (2 mm) of the optic nerves centered 3 mm posterior to the globe were removed from all of the eyes. A single reference cut was made radially at 12 o'clock and a double cut at the temporal side of the nerves for the purpose of orientation following embedding and sectioning. All axon counts were carried out on paraphenylene diamine (PPD)-stained optic nerve sections that had been embedded in Durcupan and were cut 1 μm thick. Counts were done only on the superior hemisection of both HEA and fellow optic nerves. Using cellSens Dimension software (Olympus, Center Valley, PA), lines were drawn from approximately the center of the optic nerve section to the outer edge along approximately the 10, 11, 1, and 2 o'clock line positions, using the 12 o'clock cut as a reference. The lengths of these four lines were then divided into equal-length thirds, and photographs under a ×60 microscope were taken at the points joining the first and second segments and the second and third segments. The photographs were then analyzed using ImageJ. The region of interest (counting area) was 61 μm in diameter (2.92 × 103 μm2), in the exact center of the photograph. Each individual axon that lay within the region of interest was manually counted. Partial axons that bisected the border of the circle were not included. The counting procedure was done by a masked observer. A paired t-test was then used to compare counts between eyes for each individual animal.
As part of a previous study, axotomy was performed in two cynomolgus monkeys by placing large intense laser spots surrounding the optic nerves (Ver Hoeve JN, et al. IOVS 2012;53:ARVO E-Abstract 3513). Both animals developed serous retinal detachments within a few minutes of completion of the laser photocoagulation (Fig. 1A). Additionally, laser axotomy caused widespread thermal damage, destroying underlying distal retinal cell layers surrounding the optic nerve head. However, it was successful in causing nearly complete degeneration of the retinal axons (Figs. 1B–D). Axotomy by endodiathermy caused minimal collateral damage (Fig. 2A). Specifically, the spots needed to cauterize the axons were small; they could be placed closer to the optic nerve head, and there were no secondary retinal detachments. One of the four animals (Cy1) that underwent endodiathermy experienced moderate postoperative inflammation that cleared with administration of a subconjunctival steroid.
At 3 months post-HEA, there was marked thinning of the inferior nerve fiber layer of the HEA eye evident both on fundus photography (Fig. 2B) and by SD-OCT (Fig. 3), which was later confirmed histologically (Fig. 4) in all four animals. Also seen by SD-OCT were small microcystoid spaces in the inner retina, which were also found on the histologic sections (Fig. 4). No outer retinal injury was apparent on the SD-OCT scans for any animal. Fluorescein angiography, carried out 3 months after HEA in all of the animals, showed no sign of retinal or choroidal vascular disturbance distal to the endodiathermy spots (not shown).
A typical mfERG trace array and the superior and inferior retinal locations of the seven-element hexagon groups used for data analysis are shown in Figure 5. The N1, P1, N2, and P2 amplitudes of the K1 mfERG waveform at 3 months postendodiathermy were present in the axotomized inferior retinas of the HEA eyes. No statistically significant differences in amplitude were seen for N1 (P = 0.449) when compared to the corresponding baseline amplitudes. This was also true for the fellow uninjured eye N1 response (P = 0.792). The P2 component amplitude of the K1 mfERG waveform, however, was found to be significantly decremented (P < 0.0007) postaxotomy as compared to the baseline amplitudes in the HEA eye. P2 decrement did not occur in the fellow eye (P = 0.743). The p1 and p2 features of the K2.1 mfERG waveform were also seen to be significantly decremented (P < 0.00005 and P < 0.002, respectively) as compared to the baseline responses following HEA. A similar reduction was not seen in the fellow control eye. The averaged traces from the inferior retinal hexagon groups of the HEA eyes, both at baseline and following HEA, are shown in Figure 6.
The RMS measures of the seven-element hexagon groups from superior and inferior retinas (Fig. 5) were compared between the HEA (right) and fellow control (left) eyes and expressed as ratios. This ratio was calculated by dividing the RMS of the HEA eye by the RMS of the fellow control eye for a given seven-hexagon (superior or inferior) area. For example, in Figure 7A, the “superior” plot of the “baseline” section is the collection of ratios of the superior area of the right eye (K1 9–35 ms epoch) divided by the superior area of the left eye on each testing day—the distribution of the individual test-day ratios for all of the baseline testing days is shown as a box plot.
The difference in the ratios between the baseline and post-HEA mfERG data were statistically evaluated using t-tests. For K1, the respective superior and inferior responses in the HEA and fellow control eyes were similar for the 9 to 35 ms epochs (Fig. 7A). The mean of the ratios comparing the responses of the K1 9 to 35 ms epoch approximated a numerical value of 1 for both the HEA (P = 0.881) and fellow control (P = 0.401) eyes. Ratios were also compared for the K1 40 to 70 ms epoch responses (Fig. 7B). RMS ratios of the inferior/axotomized region of the HEA and fellow control eyes were found to be significantly decreased (P < 0.00004) in comparison to baseline responses. The ratios comparing the responses of the superior/uninjured region of the HEA and fellow control eyes were found to also be significantly decreased (P < 0.036) in comparison to baseline responses. However, when compared within individual animals, no significant difference was found between baseline and post-HEA RMS ratios in the superior (uninjured) region (P = 0.585, 0.090, 0.208, 0.136). Significant decreases in RMS ratios were seen postendodiathermy as compared to baseline response ratios in the inferior (HEA) region in three of the four individual animals (P < 0.004, 0.011, 0.639, 0.008). The averaged RMS ratio (HEA versus fellow control eyes) of the K2.1 (1–80 ms epoch) response for the seven-element hexagon groups was not significantly different (P = 0.761) between baseline and post-HEA responses from the superior (uninjured) retina (Fig. 7C). The RMS ratio of the K2.1 mfERG of the inferior retina, however, was found to be significantly decreased (P < 0.008).
Radial retinal sections stained with Richardson's23 solution (Fig. 8) showed a marked reduction in numbers of nuclei in the ganglion cell layer and thinning of the nerve fiber layer in the axotomized inferior retinas of the HEA eyes. Nonstaining microcystoid spaces were present in the inner retina in the axotomized areas of all four of the HEA eyes. Photoreceptor shape, distribution and opsin IHC staining of the rods, S-cone, and M/L-cones were similar in the axotomized and nonaxotomized retinas. No disruption of the normal regular matrix of photoreceptors in the axotomized retinas was evident on tangential sections cut at the level of the inner segments (not shown). There was little, if any, GFAP staining of the Müller cells in both the uninjured and axotomized regions of the HEA eyes (Fig. 9). GFAP was prominent in the inner retinas (presumably the astrocytes) and illustrated the thinning of the nerve fiber layer in the axotomized areas.
Cross sections of optic nerves stained with PPD showed a loss of greater than 95% of axons in the inferior segment of the nerves, corresponding to the area of the HEA (Fig. 10). Some patches of axons did remain relatively undisturbed in the inferior portion of the optic nerves in each axotomized eye, which may have occurred because endodiathermy was not applied directly over the large retinal vessels.
There was no apparent difference in the amount of axonal loss for the two animals that were sacrificed 3 months postaxotomy (Cy1 and Cy2) and the two animals sacrificed at 4 months postaxotomy (Cy3 and Cy4). Likewise, no differences in either Richardson's histology or IHC (for either GFAP or photoreceptor opsins) were seen between the 3- and 4-month sacrifice times.
For the two animals that underwent injection of fluorescent microspheres, the three-dimensional density plots (proportional to blood flow) of each eye were compared subjectively. Plots were first overlaid, visually compared, and found to not contain any major differences with respect to shape or total number of microspheres. The plot of the averaged differences between eyes was relatively flat. That is, there was no obvious decrement or enhancement of blood flow in the superior and inferior regions of the HEA eyes (Fig. 11).
The average density of axons in the superior portion of the optic nerves was not significantly different between HEA (right) and fellow control (left) eyes (P = 0.70, Table). In other words, there was no apparent secondary degeneration in the portions of the nerves on the opposite side of the axotomies. Axonal densities in the normal (nonaxotomized regions) were similar to what has been previously reported for this species.26
Unlike what is observed in chronic human glaucoma or experimental glaucoma in nonhuman primates, no apparent anatomic or functional injurious effects were seen in the outer retina with HEA. Therefore, the hypothesis that the photoreceptor alterations found in glaucoma are the result of dying or absent RGCs is not supported in this model.
HEA produces much less collateral damage than laser axotomy (Figs. 1, ,2).2). Because the endodiathermy is applied directly to the proximal surface of the retina, only enough energy need be administered to destroy the underlying nerve fibers, whereas laser requires large spots of high intensity so that the heat generated at the retinal pigment epithelium will extend up to and include the nerve fiber layer. In our experience, not only does the intense laser destroy a large area of adjacent retina but additionally generates an acute serous retinal detachment (Fig. 1), which, although it resorbs quickly, may affect the subsequent experimental results.
An advantage to performing an HEA is that comparisons can be made not only between the control and the treated eye but also between the axotomized inferior and uninjured superior retinas of the same eye. By comparing the effects within the same eye and to the fellow eye, nonspecific effects related to the surgical procedure per se can be detected. This was especially useful for the mfERG analyses in comparisons between baseline and postaxotomy conditions (Figs. 6, ,7).7). No significant differences were seen for the early waveform components (9–35 ms epoch RMS ratio) of the K1 mfERG. Although driven by the photoreceptors, most of the N1 and P1 signal strength in nonhuman primates comes from bipolar cell contributions.27 In our experience, these early wave features are exaggerated in experimental glaucoma (Nork TM, et al. IOVS 2012;53:ARVO E-Abstract 6607).9 The later K1 waveforms (40–70 ms epoch) are thought to be largely inner retinal (i.e., RGC) in origin—especially in the cynomolgus monkey.9 These later K1 waveforms were significantly reduced in amplitude in the axotomized portions of the treated eyes, but not the nonaxotomized regions (Figs. 6, ,7).7). Unlike the K1 mfERG, which is not influenced by responses to preceding stimulus flashes, the K2.1 mfERG represents the effect on a flash response from a response occurring in the immediately preceding base period interval.28 As such, the K2.1 is thought to reflect rapid local retinal (nonlinear response) adaptation. However, there are significant species differences in the physiology of these responses. For example, we found that in advanced chronic experimental glaucoma, the K2.1 was markedly reduced in the cynomolgus monkey but little changed in the rhesus.9 In the cynomolgus monkeys used in the present study, there was significant reduction in the K2.1 mfERG but only in the axotomized portions of the retinas (Figs. 6, ,77).
There was no apparent trend in the mfERG data over time. That is, the waveforms at 1 week post-HEA were similar to those at 3 months. In an earlier study, we also found this to be the case for surgical transection of the optic nerve.9 As in this earlier study, there was considerable day-to-day variability in the responses from any given animal. There were also differences between animals. However, by combining data from several sessions, we have been able to demonstrate statistically significant experimental changes that would not otherwise be obvious. A larger number of animals would be required to pick up a relatively subtle trend in the data over time.
In vivo, there was SD-OCT evidence for nearly complete elimination of the inner retinal layers of the HEA eyes (Fig. 3), yet the outer layers appeared to be intact (Fig. 4). This was confirmed with histological analysis of the enucleated eyes. Marked dropout in the ganglion cell and nerve fiber layers was evident but only in the inferior portion of the retinas that had the inferior peripapillary endodiathermy (Figs. 4, ,8).8). Unlike what is seen in human eyes with glaucoma, there was no cone swelling or loss. By cutting the sections tangentially at the level of the photoreceptor inner segments, any loss of either rods or cones is easily spotted as a breakup of the highly regular distribution of rods and cones.3 No such changes were evident. The pattern of IHC staining for the rod, M/L-, and S-cone opsins was also normal in the axotomized retinas. Upregulation of the intermediate filament, GFAP, has been described in naturally occurring glaucoma in humans29 and dogs,30 as well as in experimental glaucoma in rats31 and cynomolgus monkeys.32 On histologic sections, this upregulation is most apparent in the Müller cells, which normally produce GFAP at levels not easily detected by IHC. In glaucomatous eyes, though, the radial nature of the Müller cells is made obvious by their strong staining with antibodies to GFAP. However, in our axotomized retinas, there was no apparent increase in Müller cell GFAP immunostaining (Fig. 9).
A phenomenon that, to our knowledge, has not been previously described in the context of experimental axotomy is the presence of small microcystoid spaces in the inner nuclear layer in the region of degenerated RGCs (Figs. 4, ,8).8). This occurred in all four of the treated eyes and only in the axotomized regions. Since the animals were sacrificed at 3 and 4 months, this appears to be a chronic phenomenon. Because the microcystoid spaces were seen both with SD-OCT in vivo and on histologic sections, they cannot be the result of a processing artifact. The microcystoid spaces occur far distal to the endodiathermy spots and are therefore not likely to be a direct result of the local injury. One possibility is that they are a general residual effect that is related to missing RGCs. Preliminary scanning by SD-OCT in some of our experimentally glaucomatous cynomolgus monkeys with advanced disease has shown the presence of similar-appearing microcystoid spaces (Nork TM, et al. IOVS 2013;54:ARVO E-Abstract 4818). Another possibly related phenomenon called “microcystic macular degeneration” or “microcystic macular edema” has recently been described in multiple sclerosis33–35 and optic neuropathy36–38 in human patients.
Optic nerve cross sections stained with PPD showed greater than 95% loss of axons in the portion of the nerves corresponding to the axotomized (inferior) areas (Fig. 10). The superior portion of the nerves was unaffected as determined by quantitative sampling. Previous work with surgical partial transection of the optic nerves in cynomolgus monkeys39 and Wistar rats40–43 has shown that there is a secondary degeneration that occurs in the nontransected areas. A potential explanation could be associated with the tight bundling of the optic nerve axons. Partial optic nerve transection could result in the release of neurotoxic agents as the cut axons die and thereby affect the nearby, uncut axons. A similar penumbral effect has been well described in stroke.44 Why secondary degeneration should not occur in our model is not immediately apparent. Endodiathermy may have the advantage that, being applied to the peripapillary retina where the axons of the superior and inferior retina are no longer in close proximity, any such penumbral effect is avoided. A second possible explanation for the absence of secondary degeneration in our model is that the surgical procedure of partial optic nerve transection, of necessity, is carried out near the vessels supplying the optic nerve—especially the branches of the ophthalmic artery. The ophthalmic artery and its branches are known to be in close proximity to the nerve.18 It is reasonable to suppose that the process of tissue dissection to expose the optic nerve for partial transection could also interfere with its blood supply. HEA by contrast is done downstream from the optic nerve blood supply and cannot interfere with it. There may have been secondary effects within the region of the HEA, that is, the inferior retina. The current experiment was not designed to answer that question. The endodiathermy spots were placed such that the areas of coagulative necrosis of the inner retinal regions (as evidenced by whitening, Fig. 2A) were contiguous. Thus, in this case, all of the axonal death could be reasonably explained by a direct effect of the endodiathermy and not due to secondary degeneration.
Whenever an ocular experiment is designed such that the fellow eye serves as control, the concern arises that there could be a systemic effect from the HEA eye that affects the fellow. This is much more likely in drug studies but cannot be absolutely ruled out a priori even in a surgical study. However, in this case, we have both baseline anatomic and functional measurements of the fellow eye, and these did not change following axotomy to the HEA eye. Also, the axon counts of the fellow eye were similar to what has been reported in the literature. So, there is no evidence for a contralateral effect.
One difficulty with the hypothesis that RGC damage produces a retrograde effect on the photoreceptors in glaucoma is that it does not explain the phenomenon of decreased choroidal blood flow throughout the eye seen in both acute45,46 and chronic (Munsey KM, et al. IOVS 2009;50:ARVO E-Abstract 5215) experimental ocular hypertension. Consistent with this, histologic studies show widespread alterations of the choroidal vessels in human47 and monkey48 eyes with chronic glaucoma. In the present study, we found no alterations in choroidal blood flow when comparing HEA and non-HEA retinal regions to the fellow control eyes in the two animals injected with fluorescent microspheres (Fig. 11). This suggests that decreased choroidal blood flow is a distinguishing feature between glaucoma and optic nerve axotomy.
In conclusion, the hypothesis that the outer retina is affected by RGC degeneration is not supported in the case of HEA. If true for all cases of RGC degeneration, this would mean that the outer retinal injury seen in glaucoma is an independent phenomenon. A logical place to look for a cause of the outer retinal effects in glaucoma is the choroidal blood flow, which may be decreased in glaucoma and ocular hypertension but not in axotomy. The choroid accounts for 85% of the blood flow to the eye.46,49 It was once thought that this high rate was unnecessary except perhaps as a heat sink to stabilize retinal temperature. We now know that the photoreceptors are perhaps the most metabolically active cells in the body50 and need a high blood flow system for proper physiologic functioning.51 Moderate elevations in intraocular pressure produce reductions in choroidal blood flow.46 The changes observed in glaucoma (photoreceptor swelling, loss and functional deficits) could be explained by outer retinal ischemia, which would not be expected to occur in HEA. Although perhaps an independent phenomenon, this does not rule out the possibility that injury to the outer retina somehow contributes to a toxic intraretinal environment leading to RGC death in glaucoma.
Supported by National Institutes of Health Grants P30 EY016665 and R01 EY014041, the BrightFocus Foundation, Research to Prevent Blindness, and The Wisconsin National Primate Research Center (P51RR000167 and P51OD011106).
Disclosure: R.J. Dashek, None; C.B.Y. Kim, None; C.A. Rasmussen, None; E.A. Hennes-Beean, None; J.N. Ver Hoeve, None; T.M. Nork, None