Melanopsin-Expressing Retinal Ganglion Cells Are More Injury-Resistant in a Chronic Ocular Hypertension Model

Abstract

purpose. To investigate the survival of melanopsin-expressing retinal ganglion cells (mRGCs) after the induction of chronic ocular hypertension.

methods. Intraocular pressure (IOP) was elevated in adult Sprague-Dawley rats using an argon laser to photocoagulate the episcleral and limbal veins. IOP was measured with a calibrated tonometer and monitored for a period. Seven days before the animals were killed, a piece of sterile foam soaked with gold fluorescent dye was placed onto the superior colliculus (SC) to label the SC-projecting retinal ganglion cells (scRGCs) retrogradely. mRGCs were visualized by free floating immunohistochemistry on whole-mounted retinas. The number of surviving scRGCs and mRGCs were counted on flatmounted retinas. The branching pattern of dendrites and soma size of mRGCs were examined.

results. An ∼1.7-fold increase of IOP and a significant loss of scRGCs were found in experimental eyes after laser photocoagulation. However, no significant cell loss or morphologic changes on mRGCs and their dendrites after the induction of chronic ocular hypertension are noticed over a 12-week period.

conclusions. Although the degeneration of retinal ganglion cells (RGCs) is a major concern in glaucomatous damage, the findings show that mRGCs are less susceptible to death after the induction of chronic ocular hypertension. This result indicates that mRGCs carry some unique properties that are different from those of other subpopulations of RGCs. The immunohistochemistry approach can be used to distinguish easily these mRGCs from other subtypes. This method provides a useful tool to investigate their injury-resistant properties that are informative for the development of effective neuroprotective treatment for glaucoma.

Glaucoma is an ocular neurodegenerative disease and characterized by progressive damage of optic nerve fibers and retinal ganglion cells (RGCs), ^{1 2} due to the progressive backward bowing and eventually collapse of the lamina cribrosa. ³ Although the major risk factor for development of glaucoma is an elevation of intraocular pressure (IOP), the exact pathogenesis is still under investigation. As glaucoma can lead to irreversible visual damage, understanding the underlying mechanisms is crucial for the treatment and prevention of this ocular disease. 

Nonvisual photoreception is one of the important roles of the retina, including photoentrainment of circadian cycles, photic suppression of activity, acute light-activated suppression of pineal melatonin secretion, and control of pupillary light responses. ^{4 5} These nonvisual functions are retained in mice that are visually blind, with the degeneration of outer retinal layers, ^{6 7} but are abolished after enucleation of the eyes, ^{8 9} which suggests a type of inner retinal cell that is intrinsically photosensitive and contributes to nonvisual photoreception. Recently, a subtype of RGCs containing a novel photopigment, melanopsin, has been discovered in the retina. ¹⁰ These melanopsin-containing (m)RGCs respond to light, independent of input from the rods and cones, and project their axons to the retinohypothalamic tract and suprachiasmatic nucleus (SCN), where the principal pacemaker center for circadian rhythm is located. Indeed, this group of mRGCs has been found to be involved in modulating the circadian rhythm. ^{10 11 12 13 14}  

By using an antibody against melanopsin to distinguish this specific subtype of RGCs, we demonstrated that there was no significant loss of mRGCs and changes in the branching pattern of dendrites and soma size after the induction of ocular hypertension. The findings suggest that some important properties in mRGCs may protect the cell from damage. 

Methods

Animals

Thirty Sprague-Dawley female rats (250–280 g) were reared in a temperature-controlled room on a 12-hour light–dark cycle in the Laboratory Animal Unit of The University of Hong Kong. The animals were divided into five groups. A (n = 8), B (n = 8), C (n = 8), and D (n = 4) were used for the study of RGC loss at 2-, 4-, 8- and 12-week time-points, respectively. Group N (n = 2) served as the control. 

The animals were anesthetized with intraperitoneal injection of a mixture of ketamine (70 mg/kg) and xylazine (7 mg/kg) during the experiments and were killed with an overdose of pentobarbital sodium (150 mg/kg). All the experimental and animal handling procedures complied with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were also reviewed and approved by the Faculty Committee on the Use of Live Animals in Teaching and Research, The University of Hong Kong. 

Chronic Ocular Hypertension Model

Ocular hypertension was induced unilaterally with an argon laser (Ultima 2000SE Argon Laser; Coherent, Palo Alto, CA) to photocoagulate the limbal and episcleral veins in the animals in groups A, B, C, and D. Only right eyes (experimental eyes) received laser treatment, and left eyes served as the control. Approximately 20 laser spots (power, 1 W; duration, 0.1 s; spot size, 50–100 μm) on two episcleral veins and 70 spots on limbal veins (270° around the limbus, except on the nasal side) were applied to disrupt the aqueous outflow of experimental eyes. A second laser photocoagulation with the same procedures was applied 7 days later to ensure the elevation of IOP was maintained. 

Measurement of IOP

IOP was measured with a calibrated tonometer (Tonopen-XL; Mentor Norwell, MA) on a schedule shown in Table 1 . During measurements, the animals were anesthetized with a mixture of ketamine and xylazine. Before measurement, one drop of proparacaine hydrochloride 0.5% (Alcaine; Alcon, Ltd., Fort Worth, TX) was applied to the eyes to desensitize the cornea. All IOP measurements were taken consistently at approximately 10 AM to avoid diurnal variations and 10 readings were averaged for each measurement. The animals were put back into cages after the measurements and kept warm under light for recovery. 

Retrograde Labeling of Superior Colliculus-Projecting RGCs

The superior colliculus (sc)RGCs were retrogradely labeled by placing a piece of sterile foam (Gelfoam; Upjohn, Kalamazoo, MI) soaked with 6% gold fluorescent gold (FG) tracer (Fluoro-Gold [FG]; Fluorochrome, Denver, CO) over both exposed SC after removing the overlying cortex. Seven days after RGC labeling, the animals were killed with an overdose of pentobarbital sodium. After the removal of corneas, lenses, and vitreous bodies, the eyecups were fixed in 4% paraformaldehyde-phosphate-buffered saline solution for 1 hour. The retinas were then taken out, washed with phosphate-buffered saline (0.01 M PBS [pH 7.4]), and flatmounted onto glass slides for cell counting (except those for immunohistochemical staining of melanopsin). 

Immunohistochemistry of mRGCs

Sixteen pairs of retinas from groups A, B, C, and D (four pairs from each group) and two pairs of retina from Group N were chosen for the investigation of mRGCs survival. Free-floating retinas were washed three times with PBS for 10 minutes each and blocked with 10% normal goat serum in 0.3% Triton X-100 and PBS mixture (PBT) for an hour at room temperature. Then the retinas were incubated with melanopsin primary antibody (polyclonal rabbit anti-melanopsin; 1:1000; Affinity Bioreagents, Golden, CO), which is raised against a synthetic peptide corresponding to amino acid residues 455 to 474 from rat melanopsin protein, in 1% bovine serum albumin in PBT for 3 days at 4°C. To visualize mRGCs, the retinas were incubated with Alexa Fluor FITC or rhodamine goat anti-rabbit IgG secondary antibody (1:400; Invitrogen Corp., Carlsbad, CA) in PBS for 1 hour at room temperature. The free-floating retinas were washed with PBS, flatmounted onto glass slides, and coverslipped for mRGC counting. 

RGC Counting and Morphologic Study

The RGC counting was performed with a fluorescence microscope (Nikon, Kawasaki, Japan) and an eyepiece equipped with a grid (400 × 400 μm²/microscopic field). Samples were taken along the median line of each quadrant from optic disc to the peripheral border of the retina at 500-μm intervals. Each retina was divided into four quadrants, and eight microscopic fields were counted per quadrant, corresponding to approximately 9% of total retinal area. Images of retina showing FG-positive RGCs (i.e., scRGCs) and mRGCs were captured using a UV-385 filter and FITC/rhodamine filters, respectively. The scRGC survival after the induction of chronic ocular hypertension was assessed by counting the number of scRGCs in groups A, B, C, and D animals. To enhance the accuracy of counting a small population of mRGCs using sampling, the number of mRGCs was counted on whole retinas of two randomly selected animals in group C for compatibility. The number of primary dendrites of mRGCs was counted, and the soma areas of mRGCs were measured by a public domain image analysis software ImageJ (available by ftp at zippy.nimh.nih.gov/ or at http://rsb.info.nih.gov/ij; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD). For detailed investigation of dendritic pattern of mRGCs, stacked images of flatmounted retinas (approximately 20 stacks for each retinal scan) with anti-melanopsin immunohistochemistry staining were captured using a laser scanning microscopy (LSM510 META; Carl Zeiss Meditec, Jena, Germany). Stacked images were projected along the plane x to show the branches of dendrites and soma of the cells. The dendrites of each mRGC were reconstructed and the analysis of branching pattern of dendrites was examined using a modification of Sholl’s analysis method. ¹⁵ Briefly, concentric circles of increasing radii 12.50 μm were drawn using the center of the soma as the center of the circles (Fig. 1) . The number of dendrites intersecting each circle was counted and plotted against the distance from the center of soma. The morphology of the dendrites was also investigated in the cross-sectional plane by projecting the stacked images along plane z. A single-blinded approach was used in all cell counting and measurements, to eliminate any subjective bias during the procedures. 

Statistical Analysis

One-way analysis-of-variance (ANOVA; Prism ver. 4.0; GraphPad Software, Inc., San Diego, CA), followed by a post hoc Tukey multiple comparison test, were used for the analysis of the results, including the IOP before and after laser treatment, RGC survival, soma areas, and the data of Sholl’s analysis. To enhance the accuracy of the statistical test for a small population of mRGCs, a Kruskal-Wallis nonparametric test was used to analyze the number of survival mRGC between the experimental and control eyes after the induction of ocular hypertension. Statistically significant difference was set at P < 0.05. 

Results

Intraocular Pressure

Similar to previous studies using the same ocular hypertension model, ^{16 17 18} the preoperation IOPs were approximately 14 mm Hg and there was no difference among groups (P > 0.05). Two hours after laser treatment, the IOP of experimental eyes was dramatically elevated to approximately 29 mm Hg due to the blockage of aqueous outflow. The IOP was further increased to ∼32 mm Hg 6 hours after the treatment and remained at ∼23 mm Hg thereafter (Fig. 2) . The IOP of experimental eyes after laser treatment was significantly elevated compared with those of control eyes (P < 0.001). 

scRGC in Chronic Ocular Hypertension Model

As most RGCs have their terminals projected to SC, retrograde labeling of RGCs at SC provides a useful tool for investigating cell survival. Similar to previous studies, ^{19 20} the mean density of scRGCs in control retinas was approximately 2300 cells/mm² in rats (Fig. 3) . 

Loss of RGCs is one of the characteristics in ocular hypertension models. ^{1 2} In the present study, the loss of scRGCs was progressively increased throughout the time points (Fig. 3)and coherent to those of previous studies. ^{17 18} Comparing the density between experimental eyes and control eyes, a decrease in the number of surviving cells after laser treatment was observed (P < 0.001; Fig. 3 ). 

mRGC in Chronic Ocular Hypertension Model

Melanopsin is a novel photopigment that is expressed only in a small population of RGCs. ^{10 11 12 13 14} Although a consistent loss of scRGCs after the induction of chronic ocular hypertension was noticed, mRGCs of groups A, B, C, and D experimental retinas did not show any significant cell loss after laser treatment compared with those of the control retinas and the control group (group N; P > 0.05; Fig. 4 ). To ensure that our sampling method of cell counting was reliable, the total number of mRGCs was counted on two pairs of whole retinas in group C animals. Comparing those cell count results using sampling strategy, no difference in the number of mRGCs in experimental and control retinas was found (P > 0.05; Fig. 4 ). 

The morphology and the branching pattern of dendrites of mRGCs were also examined in the present study. We used immunohistochemistry to stain the cell bodies, dendrites, and axons of mRGCs with an antibody against melanospin. ¹⁴ Some processes of mRGCs could not be shown using ordinary fluorescence microscopy as the branches were not ramified at a single focal plane (Figs. 5A 5B 5C 5D) . Stacked images of scanning laser microscopy allowed viewing of all the processes by putting the images at different focal planes into one image. The dendrites of each mRGC were highly branched and exhibited an extensively overlapped pattern with adjacent mRGCs (Figs. 5E 5F 5G 5H) . The branching pattern of dendrites was investigated using Sholl’s concentric circle method. ¹⁵ Results showed that induction of ocular hypertension did not lead to any pathologic changes on the branching pattern of the mRGC dendrites (Fig. 6) . A cross-sectional plane of view (stacked images constructed along axis z) illustrated the pattern of stratification of mRGC dendrites in the inner plexiform layer (IPL). No obvious alternation of stratification was observed in the ocular hypertensive retina compared with the control (Fig. 7) . The average mean soma area of the mRGCs was approximately 170 μm², which was similar in the two groups, and there was no difference in the number of primary dendrites after the induction of ocular hypertension (P > 0.05; Fig. 8 ). 

Discussion

Elevation of IOP in a Chronic Ocular Hypertension Model

The major risk factor of glaucoma is the elevation of IOP, and hence lowering IOP is a primary goal of treatment. Different approaches for inducing ocular hypertension in animal have been proposed, to aide in understanding the underlying mechanism of glaucoma. ^{21 22 23 24 25} The extent of IOP elevation normally varies with the induction approaches and animal species. In the present study, laser photocoagulation at the episcleral and limbal veins was used to impede the aqueous outflow, and the induced IOP elevation was comparable to that in previous studies in the same animal model. ^{16 17 18} The IOP was dramatically increased by approximately 2.2-fold in the experimental eyes several hours after laser treatment. This acute IOP elevation was probably due to the sudden blockage of aqueous outflow by laser photocoagulation at its major drainage system. Thereafter, the IOP was maintained at approximately 23 mm Hg (from day 1 to week 12), approximately 1.7-fold elevation was induced in our model which was similar to a previous study using unilateral cauterization of episcleral veins. ²³  

Loss of scRGC in a Chronic Ocular Hypertension Model

The loss of RGCs or their axons is an important sign of glaucomatous damage. ^{1 2} Similar to the approaches used to create elevated IOP, the extent of cell death depends on different ocular hypertension models and animal species. Certain populations of RGCs are shown to undergo apoptosis after IOP elevation, ^{21 26} by triggering a cascade of caspases and proapoptotic Bax activation, and decreasing the level of Bcl-2. ^{27 28 29} Cauterization of episcleral veins in rats created a 4% RGC loss per week after the induction of ocular hypertension. ²¹ Using laser photocoagulation at episcleral veins and limbal veins in rat ocular hypertension model, the scRGC loss was approximately 13% 2 weeks after laser treatment. ^{17 18} To exclude the possibility of RGC damage due to laser treatment itself during induction, we conducted a pilot study by performing laser photocoagulation unilaterally in two groups of animals. We killed the animals immediately and 24 hours after laser treatment, and no significant cell loss was found (data not shown). 

Effect of Induced Chronic Ocular Hypertension on mRGCs

Mammalian circadian rhythm is regulated by a principal pacemaker located in SCN. Recently, a subtype of RGCs was shown to send axons into the SCN and regulate the rhythm in photoentrainment. These RGCs are intrinsically photosensitive independent of input from the rods and cones. ^{11 30} Intriguingly, as our results demonstrated, within this unique subtype of RGCs, mRGCs did not show any significant cell loss compared with the scRGCs, even after 12 weeks of induction of chronic ocular hypertension. The finding was supported by a recent study on mRGC and optic nerve axotomy in mice that yielded comparable results. ³¹ In optic nerve crush and axotomy models, a population of RGCs survives even after a short period, but most RGCs die thereafter. ^{32 33 34} A similar phenomenon was observed in glaucoma or ocular hypertension. ^{35 36 37} Our result suggests that these RGCs, at least in part, are melanopsin containing. The underlying mechanism for the resistance to injury in this group of cells is uncertain. We do not know at this stage whether the neural circuit within the retina or the unique functionalities of this group of cells associated with this photopigment contributes to the survival after injury. The issues are currently under investigation. 

In addition, earlier studies showed a significant decrease in soma size and the branching pattern of dendrites of RGCs in monkeys and cats after the induction of ocular hypertension. ^{35 38} In the present study, we sought to find out any morphologic change of mRGC before and after the induction of ocular hypertension in our rat model. Previous reports categorized mRGCs to type III RGCs according to the dendritic pattern and soma size, ¹⁴ which fits with our observation. After the induction of chronic ocular hypertension, no observable changes in the soma size were noticed. Apart from shrinkage of soma, RGCs also undergo a pathologic degeneration of dendrites in ocular hypertension cases. ^{35 38} The concentric circle method of Sholl ¹⁵ was adopted to investigate the change of the branching pattern of mRGC dendrites in our model. In certain circumstances, the difficulty encountered in using the concentric circle method ¹⁵ was the extensive overlapping of distal dendrites of mRGC with the neighboring cells. Therefore, dendrites extending up to 100 μm from the center of soma were reconstructed and analyzed. Results showed no significant changes in the branching pattern after the induction of ocular hypertension. A previous report demonstrated that most mRGCs send their dendrites to the outer/OFF substratum of the IPL, whereas some of the neurites bifurcated in the inner/ON substratum of the IPL. ³⁹ A cross-sectional plane of view provided a clear image of the stratification of mRGC dendrites into the IPL, which is in line with the report. The unique morphologic and physiological properties of mRGCs, having a large soma size, long and sparsely branching dendrites covering the whole retina, containing intrinsic photosensitive property, ¹⁴ are suggested to be the factors contributing to the resistance in cell injury. ³¹  

Cellular resistance to injury-induced damage is extensively studied. Previous studies proposed that cells expressing a high level of cytochrome oxidase are invulnerable to injury. ⁴⁰ Heat shock proteins are also well known to increase in expression after pathologic insults for initiating neuroprotection on retina and RGCs. ^{41 42 43 44} Similarly, delayed cell death has been observed after optic nerve cut, which is due to the pathologic insult-inducing elevation of Akt phorphorylation and hence prevents caspase activation. ⁴⁵ Generally, in addition to the morphologic and physiological properties of the cell, expression of proteins or enzymes in a normal or injured cell helps the cells to cope with the change after injury. It may also be attributable to a combination of intrinsic and extrinsic properties, as discussed. 

Neurons in the retina are diversified. There are approximately 12 types of RGCs according to the physiological and anatomic classification. ^{46 47} Although glaucoma is characterized by the death of RGCs, different types of RGCs may have a different extent of resistance to the damage or injury, which indicates a differential effect on RGC death. ^{35 36 37} Morphologically large RGCs with thick axons are selectively damaged in a greater extent compared with those small RGCs under glaucomatous stress. ^{36 37} Understanding what types of RGCs are more susceptible to death in glaucoma condition is informative to neuroprotective strategies. However, the lack of cell-type–specific markers of RGCs ^{46 47} still hinders the investigation of selective cell death in glaucoma. In the present study, melanopsin antibody can specifically mark one subtype of RGCs, the mRGC. Thus, this antibody can be used as a tool for further study of this group of cells which is resistant to injury. Further investigation will allow us to gain an insight into the protective mechanism of this type of cell and may provide us cues for the development of novel strategies for neuroprotection. 

Conclusions

Our findings showed that mRGCs were resistant to death after the induction of chronic ocular hypertension. This result indicates that mRGCs carry some unique characteristics that are different from other populations of RGCs. Although much evidence has shown that these physiological factors are protective to certain RGCs in the glaucoma condition, the identification of these cells remains controversial. That mRGCs can be identified readily using the antibody targeting melanopsin provides an opportunity to conduct further investigations into the unique properties of mRGCs and the underlying mechanism of glaucoma. 

 Supported by the award of The University of Hong Kong, the donation in memory of the late Mr. and Mrs. Ma Yip-Seng and Research Grant P-G73 of The Hong Kong Polytechnic University.

 Submitted for publication September 29, 2005; revised March 2, 2006; accepted May 19, 2006.

 Disclosure: R.S. Li, None; B.-Y. Chen, None; D.K. Tay, None; H.H.L. Chan, None; M.-L. Pu, None; K.-F. So, 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.

 

Table 1.

 

The Schedule of Laser Treatment and IOP Measurements

	Group A (Week 2)	Group B (Week 4)	Group C (Week 8)	Group D (Week 12)
Week 0	←Prelaser IOP measurement and 1st laser treatment→
2 & 6 hours	←IOP measurement→
Day 1	←IOP measurement→
Week 1	←IOP measurement before 2nd laser treatment and 2nd laser treatment→
Week 2	IOP measurement and death	IOP measurement	—	—
Week 3	—	—	IOP measurement	IOP measurement
Week 4	—	IOP measurement and death	—	—
Week 5	—	—	IOP measurement	IOP measurement
Week 6	—	—	—	—
Week 7	—	—	IOP measurement	IOP measurement
Week 8	—	—	Death	—
Week 9	—	—	—	IOP measurement
Week 12	—	—	—	IOP measurement and death

Figure 1.

 

Sholl’s analysis of the branching pattern of dendrites. Concentric circles with increasing radii of 12.50 μm were drawn on top of a reconstructed mRGC.

Figure 2.

 

Change of IOP (mean ± SD) before and after laser treatment at different time points. The IOP of experimental eyes was significantly elevated after the first laser treatment compared with those of control eyes. The elevated IOP remained at a steady level throughout the experimental period.

Figure 3.

 

Survival of scRGCs labeled with FG, before and after the induction of chronic ocular hypertension. Density of scRGCs was expressed as the number of scRGCs/mm² (mean ± SD) in groups N, A, B, C, and D. There was a loss of scRGCs in experimental eyes of group A, B, C, and D animals after treatment (***P < 0.001) but not in group N (P > 0.05). No difference was found in control eyes among the groups (P > 0.05).

Figure 4.

 

Survival of mRGC stained with anti-melanopsin antibody before and after the induction of chronic ocular hypertension. Density of mRGCs was expressed as the number of mRGCs/mm² (mean ± SD) in groups N, A, B, C, and D and whole retina (cell counting on two pairs of whole retinas). No difference in the number of mRGCs was found between experimental and control eyes among all groups after treatment (P > 0.05).

Figure 5.

 

Images of representative retinas from group N and C animals captured with a fluorescence microscope (A–D). (A) A right retina in group N. (B) A left retina in group N. (C) An experimental retina in group C. (D) A control retina in group C. Stacked images representative retinas from groups B and C captured with a laser scanning microscope (E–H). (E) An experimental retina in group B. (F) A control retina in group B. (G) An experimental retina in group C. (H) A control retina in group C. Surviving mRGCs were visualized with Alexa Fluor 488 or 568 goat anti-rabbit IgG secondary antibody. Scale bars: (A–D) 50 μm; (E–H) 25 μm.

Figure 6.

 

The branching pattern of the mRGC dendrites according to Sholl’s analysis method. No difference was noticed in the branching pattern between the experimental and control eyes (P > 0.05). Error bars, SEM.

Figure 7.

 

Cross-sectional view of the stratification of mRGC dendrites in the IPL. (A) Experimental retina from group C. (B) Experimental retina from group D. (C) Control retina. Outer IPL, outer substratum of the IPL; inner IPL, inner substratum of the IPL; GCL, ganglion cell layer. Scale bars, 15 μm.

Figure 8.

 

Morphologic study of cells from groups B, C, and D. (A) Number of primary dendritic branches (mean ± SD) of mRGCs. (B) Soma area (mean ± SD) of mRGCs. No morphologic changes were noticed in the mRGCs after the induction of ocular hypertension (P > 0.05).

The authors thank Tak-Ho Chu and Wui-Man Lau for their generous help and discussion of the experiments.