Submacular DL-α-Aminoadipic Acid Eradicates Primate Photoreceptors but Does Not Affect Luteal Pigment or the Retinal Vasculature

Abstract

Purpose.: Macular telangiectasia type 2 (MT2) is a condition of uncertain etiology characterized by retinal vascular abnormalities, depletion of luteal pigment, and photoreceptor loss. To model this condition, the authors recently used a purportedly glial-selective toxin, DL-α-aminoadipic acid (DL-α-AAA), to test the effect of Müller cell disruption on the blood-retinal barrier in rats. In this study, they investigated macular changes after subretinal injection of DL-α-AAA in monkeys.

Methods.: Various doses of DL-α-AAA were injected beneath the macula in eight monkey eyes. Eyes were examined by multifocal electroretinography (mfERG), optical coherence tomography (OCT), fundus autofluorescence, color photography, and fluorescein angiography. Five months after injection, eyes were examined by histology and immunohistochemistry for changes in photoreceptors and the retinal glia. In vitro studies evaluated the effect of DL-α-AAA on 661W cone photoreceptor viability.

Results.: Subretinal injection of DL-α-AAA resulted in virtually complete ablation of photoreceptors in the injected area, as shown by OCT and histology, and severely impaired mfERG responses. Müller cells, albeit activated, survived the injury. Macular pigment remained unchanged in the central fovea. Subretinal injection of DL-α-AAA did not induce vascular leakage, though it increased the fundus autofluorescence. DL-α-AAA had a dose-dependent toxic effect on 661W photoreceptors.

Conclusions.: Submacular injection of DL-α-AAA induced severe damage to photoreceptors but failed to eliminate Müller cells in monkeys. Central macular pigment persisted despite loss of photoreceptors, and the retinal vasculature was unaffected. These observations may have significance in studying the roles of different cellular components in the pathogenesis of MT2.

Macular telangiectasia type 2 (MT2) is a potentially blinding condition of the retina. The predominant and most consistent clinical features include reduced macular pigment optical density, ^{1 –3} parafoveal retinal opacification, ⁴ increased macular autofluorescence, ^2,5 superficial crystalline deposits, ^4,6 foveal atrophy due to disturbance of photoreceptors and inner and outer lamellar retinal cavitation, ^4,7 intraretinal vascular abnormalities including right-angled venules, vascular telangiectasis and leak, and intraretinal and subretinal neovascularization. ^4,6,8 Functional deterioration is predominantly caused by foveal atrophy, hyperplasia of retinal pigment epithelium (RPE), and development of intraretinal or subretinal neovascularization. The morphologic and functional alterations are typically most pronounced temporal to the foveola. ^{6,9 –12} Recently there has been an increasing interest in this disease not only as a model for more common retinal degenerations but also because it appears to be more common than previously thought. There has been great progress in elucidating the clinical phenotype of MT2 ^{1 –3,5,7 –17} ; however, the basic mechanisms leading to the observed retinal alterations have remained largely unexplored and unexplained. It is unclear whether the primary defect in MT2 lies in retinal components such as photoreceptors, Müller cells, or RPE. 

Central reduction or depletion of macular pigment is a common finding in MT2. ^2,3,17 Macular pigment consists of the two different xanthophylls, lutein and zeaxanthin, which are thought to accumulate along the axons of the cone photoreceptors, predominantly in a region between 200 μm and 900 μm diameter around the fovea. ^{18 –20} The main location of macular pigment is believed to be in the Henle's fibers of the fovea and in the perifoveal inner nuclear layer. ^19,21 The foveola is not, however, composed solely of cone photoreceptors. A cone-shaped zone of Müller cells also exists in the central and inner part of the fovea centralis. ^21,22 It has been hypothesized that the Müller cell cone acts as the primary structural and functional support for the fovea and is a reservoir for macular pigment. ²¹ Currently, there is no direct evidence linking MT2 with Müller glial dysfunction and loss of macular pigment. 

The somata of Müller cells are located within the inner nuclear layer and have processes that envelop all neurons and synapses extending from the inner limiting membrane to the outer limiting membrane. ²³ They play a central role in retinal glucose metabolism and constitute an anatomic and functional link between retinal neurons and blood vessels. ²³ Müller cells are intricately involved in the uptake and degradation of the neurotransmitters glutamate and γ-aminobutyric acid, shuttling energy metabolites from the vasculature to neurons. They act as a siphon for the uptake of extracellular potassium, and they maintain both the ion balance and the pH of the retinal milieu. ²³ Müller cells also play a central role in the formation and maintenance of the inner blood-retinal barrier (BRB). ^23,24 Müller glial dysfunction has been associated with retinal compromise, including neuronal damage and breakdown of the BRB. ^23,25  

DL-α-aminoadipic acid (DL-α-AAA), a purportedly selective glial toxin, is a homologue of the excitatory amino acid L-glutamate, and it has been reported to cause selective glial dysfunction by suppressing glutamine synthetase (GS). ^{26 –29} When exposing α-AAA to the retina in organ culture, Müller cells are the primary site of α-AAA accumulation. ³⁰ We have recently investigated the specificity of DL-α-AAA in vitro and tested its effect on BRB breakdown in adult rats. ³¹ Results from our own investigations showed that DL-α-AAA was toxic to glial cells, including Müller cells and astrocytes, but not to other types of BRB-related cells, including retinal vascular endothelium, pericytes, and RPE in vitro. ³¹ Subretinal injection of DL-α-AAA induced vascular telangiectasis and increased vascular permeability in adult rats. ³¹ Retinal circulation and neuronal anatomy, such as the populations of rods and cones, are different in rats and humans, but the macular anatomy and physiology of monkeys are similar to those of humans ^32,33 ; therefore, information would be more valuable if it were obtained by modeling the disease in monkeys. In this study, we examined changes in the retina after subretinal injection of DL-α-AAA in monkeys and further evaluated the cellular toxicity of DL-α-AAA in 661W cone-derived photoreceptors in vitro. 

Materials and Methods

Preparation of DL-α-AAA Solutions

DL-α-AAA (Sigma, Sydney, Australia), which is not water soluble, was initially dissolved in 1 N HCl to form a 120-mg/mL stock solution. The stock solution was diluted with culture medium or balanced salt solution (BSS), and the pH was adjusted to 7.4 with NaOH and then filtered as described previously. ³¹  

Animal Preparation and Anesthesia

All animal experiments adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The project was supervised by the Animal Ethics Committees at The University of Sydney in Australia and the Peking University in China. Monkeys were anesthetized with intramuscular injection of a mixture containing ketamine (20 mg/kg), acepromazine maleate (0.25 mg/kg), and atropine sulfate (0.125 mg/kg) and the airway, respiration, and pulse were monitored during all procedures. Pupils of monkeys were dilated with 2.5% phenylephrine hydrochloride and 1% tropicamide drops (Alcon, Frenchs Forest, NSW, Australia) before subretinal injection and clinical examinations as described here. 

Subretinal Injection in Monkeys

Eight eyes of four monkeys received subretinal injection of 50 μL BSS (n = 2) or DL-α-AAA at a concentration of 5 (n = 1), 10 (n = 2), or 50 (n = 3) mM, as described previously. ³⁴ The microdelivery system consisted of a 100-μL Hamilton syringe and a 20-cm length of thick-walled polyethylene/PVC extension tube with a Luer lock connecting to a 20-gauge, 25-mm-long subretinal cannula with a 39-gauge, 5-mm-long flexible tip that could be angled after loading the solution (catalog no. E7365; Bausch & Lomb, Rochester, NY). Two ports were made with a 20-gauge V-lance at 1:30 and 10:30 o'clock through the pars plana, 3 mm behind the limbus. A 20-gauge fiberoptic probe was inserted into one port for endoillumination and another for the subretinal cannula insertion. Fifty microliters of solution was injected beneath the macula to create subretinal blebs designed particularly to include the temporal perifoveal region in each eye. 

Clinical Examinations

Monkeys were examined multifocal electroretinography (mfERG), optical coherence tomography (OCT), fundus autofluorescence (FAF), color photography (FCP), and fluorescein angiography (FFA) before injection and at different intervals after subretinal injection. OCT images were recorded using the Stratus OCT (Zeiss Meditec, Dublin, CA). The built-in macular thickness scan program consisting of six radial scan lines measuring 6 mm in length at 30° intervals centered on the fovea was used for OCT imaging. 

The mfERG was recorded (VERIS Science 4.0 system; Electro-Diagnostic Imaging, Redwood City, CA). The first-order kernel responses were recorded according to the International Society for Clinical Electrophysiology of Vision guidelines for basic mfERG. ³⁵ Before mfERG recording, the head positions of monkeys were adjusted under indirect ophthalmoscopy so that the macular regions were directed to the center of the stimulus, 33 cm from the monitor screen. Eye movement was restricted by two limbal sutures. After room light adaptation for 15 minutes, the animal was stimulated by a flashing screen pattern of 103 hexagonal elements. The other eye was covered with black tape. Responses were measured using a RETI scan (Roland Consult, Wiesbaden, Germany) connected to an gold foil corneal contact lens electrode (ERG-Jet: Universo SA, La Chaux-De-Fonds, Switzerland) and a subdermal steel reference electrode 1 cm away from the lateral canthus. The signal was amplified 3.25 × 10⁶ times and band-pass filtered, with the low- and high-cut filters set at 3 and 300 Hz, respectively. Each data set was generated as the average of eight consecutive runs. The 103 focal responses were divided into six concentric rings, and an average waveform was generated for each ring. The monkey mfERG trace has three peaks: negative 1 (N1), positive 1 (P1), and negative 2 (N2). The amplitude of N1 was defined as the distance from the highest point of a preceding positive wave to the trough of N1. P1 amplitude was measured from the trough of N1 to the peak of P1, and N2 was measured from the baseline to the trough of N2. The amplitude density of P1 (nanovolts per square degree [nv/deg²]) and the amplitudes and latencies of P1 and N1 waves were used for analysis. 

FAF was imaged with a confocal laser scanning ophthalmoscope (HRA2; Heidelberg Engineering, Heidelberg, Germany). A blue laser was used for excitation at 488 nm, and a barrier filter restricted detection of emitted light to a wavelength range of >500 nm. FCP and FFA were performed using a standard digital imaging system (CR6–45NM; Canon, Tokyo, Japan). FFA was performed by intravenous injection of 10% sodium fluorescein (0.1 mL/kg body weight), with images taken between 10 seconds and 10 minutes after dye injection, as described previously. ^34,36  

Tissue Preparation, Histology, and Immunohistochemistry

At termination of the study 5 months after submacular injection, eyes were enucleated for histology and immunohistochemistry. Monkeys were euthanatized by intravenous injection of an overdose of sodium pentobarbital. Before enucleation, a suture was made on the temporal episclera, 3 mm behind the limbus, to mark the orientation of the macula. Immediately after enucleation, a 1-cm incision was made at 6 o'clock, 3 mm behind the limbus, and the eyes were fixed in 4% paraformaldehyde in 0.1 M PBS for 24 hours. Eyes were washed in PBS and then dissected to retain temporal and nasal sides of the retinas but with intact cornea and lens to support the eye shape for paraffin embedding. The dissected eyes were placed in 70% ethanol and embedded in paraffin. Serial sections of 5-μm thickness were cut, mounted on silanated glass slides, deparaffinized, and rehydrated for histology and immunohistochemistry, as described previously. ³¹  

For histologic analysis, paraffin sections were stained with hematoxylin and eosin and examined under a light microscope. To examine the effect of DL-α-AAA on retinal glial cells, immunohistochemistry for glutamine synthetase (GS) and glial fibrillary acidic protein (GFAP) was performed after antigen retrieval, as described earlier. Sections were incubated in 10% normal goat serum blocking solution for 2 hours and were probed with antibodies to GS (MAB302, 1:100; Chemicon International, Temecula, CA) and GFAP (Z0334, 1:250; Dako, Glostrup, Denmark) in PBS containing 1% BSA and 0.3% Triton X-100 overnight at 4°C. The negative controls omitted the primary antibodies. After incubation with goat secondary antibodies conjugated to AlexaFluor 488 and 594 for 2 hours at room temperature, sections were mounted with an aqueous medium and examined by confocal laser scanning microscopy. 

Cell Culture

661W cells were generously donated by Muayyad Al-Ubaidi (University of Oklahoma Health Sciences Center, Oklahoma City, OK). The 661W cell line was derived from mouse retinal tumors and was shown to be of cone photoreceptor cell lineage. ³⁷ 661W cells were grown in DMEM (Invitrogen, Sydney, Australia) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, 100 μg/mL streptomycin, and 2 mM l-glutamine in a humidified atmosphere of 5% CO₂ and 95% air at 37°C. 

Assessments of Cell Viability and Cellular Metabolic Activity

The toxicity of DL-α-AAA in 661W cells was studied after 16-hour treatment with DL-α-AAA at concentrations of 0, 0.1, 1.0, 5.0, 10, and 50 mM. Cell viability was assessed by staining the treated cells with 2 μg/mL calcein-AM (Molecular Probes, Invitrogen, Carlsbad, CA) for live cells and 10 μg/mL propidium iodide (Sigma, St. Louis, MO) for dead cells, as described previously. ³¹  

Cellular metabolic activity after DL-α-AAA treatment was evaluated by Alamar blue assay (Biosource, Camarillo, CA), a proprietary assay designed to quantify cell proliferation, cytotoxicity, and viability, as described previously. ³¹ In brief, Alamar blue dye was diluted into the culture medium to 10% (vol/vol). For fluorescence measurement, 661W cells were transferred into a sterile flat-bottomed, multiwell cell culture plate. Cell density was adjusted to 1 × 10⁵ cells/mL for 48-well plates, and 5 × 10⁴ cells in 500 μL medium were added to each well. The cells were then allowed to grow for 2 to 3 days to form an 80% confluent cell monolayer. After 16-hour treatment with DL-α-AAA, the culture medium was aspirated, and 350 μL of 10% Alamar blue working solution was added to each well. Cells were further incubated for 3 hours to allow them to take up the dye. Fluorescence measurements were performed with a fluorescence multiwell plate reader (Safire2; Tecan, Männedorf, Switzerland) with excitation/emission wavelengths at 530/590 nm, sensitivity gain at 40, and temperature at 37°C. 

Statistical Analysis

Results are expressed as mean ± SEM. Data were analyzed using ANOVA for multiple comparisons (GB-STAT version 9.0; Dynamic Microsystems, Silver Spring, MD), with P <0.05 accepted as statistically significant. Bonferroni correction for multiple comparisons was performed when required. 

Results

Subretinal Injection of DL-α-AAA Reduced Macular Function in Monkeys

To investigate the effect of DL-α-AAA on retinal neuronal function, mfERG was performed before and 1 and 3 months after injection. Figure 1 shows mfERG trace arrays and the 3D response density plots before and 1 month after subretinal injection of BSS or different doses of DL-α-AAA. Compared with the ERG response recorded before injection, injection of BSS into the subretinal space appeared to slightly reduce the mfERG response (Figs. 1A–D). However, pronounced reductions in mfERG responses occurred in eyes receiving 5, 10, or 50 mM of DL-α-AAA (Figs. 1E–J). The reduced mfERG responses had not recovered when the injected eyes were examined again at 3 months after injection (data not shown). 

Figure 1.

 

Impaired macular function as measured by mfERG 1 month after subretinal injection. (A, C, E, G, I) Trace arrays. (B, D, F, H, J) 3D response density plots of A, C, E, G, and I, respectively. (A, B) Before injection. (C, D) Same eye as (A) and (B) after injection of BSS. (E, F), 5 mM DL-α-AAA. (G, H) 10 mM DL-α-AAA. (I, J) 50 mM of DL-α-AAA.

Values of averaged amplitude densities of the P1 waves, amplitudes, and latencies of the P1 and N1 waves are shown in Figure 2. Compared with the mfERG responses recorded before injection and in eyes receiving BSS, subretinal injection of DL-α-AAA significantly reduced the P1 amplitude density from ring 1 to ring 5 (Fig. 2A), the P1 amplitude from ring 1 to ring 4 (Fig. 2B), and the N1 amplitude from ring 1 to ring 3 (Fig. 2C). In addition, DL-α-AAA extended the N1 latency from ring 1 to ring 4 (Fig. 2D) but not the P1 latency (data not shown). In general, after subretinal injection of DL-α-AAA in the macula, changes in the P1 and N1 waveforms were predominantly confined to the macular regions from ring 1 to ring 4. No statistically significant changes were observed in the region of ring 6 after DL-α-AAA injection. 

Figure 2.

 

Reductions in amplitude densities of the P1 waves, amplitudes, and latencies of the P1 and N1 waves after subretinal injection of DL-α-AAA. (A) Reduced P1 amplitude density in eyes injected with DL-α-AAA was confined primarily to an area between ring 1 and ring 5. (B) Reduced P1 amplitudes from ring 1 to ring 4 after injection of DL-α-AAA. (C) Reduced N1 amplitude from ring 1 to ring 3 in eyes receiving DL-α-AAA. (D) DL-α-AAA extended the N1 latency from ring 1 to ring 4. *P < 0.05 and †P < 0.01, DL-α-AAA (n = 5) versus baseline (n = 5), respectively; n = 2 in BSS-injected group.

DL-α-AAA Injection Induced Retinal Neuronal Damage on OCT

Representative OCT images across the macular regions in monkeys are presented in Figure 3B. In standard resolution OCT, the nerve fiber layer (NFL), the inner plexiform layer (IPL), and the outer plexiform layer (OPL) are more optically reflectant than the inner nuclear layer (INL) and the outer nuclear layer (ONL) and are seen as red, yellow, or bright green false color in the OCT images. ⁷ The first highly reflecting layer is the nerve fiber layer. The three low-reflectance intraretinal layers, presented as blue or black false color in the OCT images, are the ganglion cell layer (GCL), INL, and ONL. The junction between the photoreceptor inner segment (IS) and outer segment (OS) is visualized as a thin, highly reflecting (red) feature in the outer retina, anterior to the RPE layer. 

Figure 3.

 

Photoreceptor injury revealed by OCT after subretinal injection of DL-α-AAA. OCT was performed before injection as baseline (A, C) and 12 days after subretinal injection of BSS (B) or different doses of DL-α-AAA (D–F). (B) Representative OCT image showing the retinal structure across the macula in monkeys. DL-α-AAA remarkably reduced retinal thickness with particular damage to the ONL, but the backscattering RPE layer (D–F, arrows) appeared intact in areas receiving DL-α-AAA.

Compared with OCT images before surgery, subretinal injection of BSS did not induce obvious retinal damage in the macula (Figs. 3A, 3B). However, DL-α-AAA caused marked thinning of the macula when examined at 12 days after injection of 5, 10, or 50 mM DL-α-AAA (Figs. 3C–F). DL-α-AAA particularly damaged the ONL, resulting in severe injury or complete loss of photoreceptors and the IS/OS (Figs. 3D–F). The highly reflectant RPE layer appeared unaffected (Figs. 3D–F, arrows). Damage to the ONL by DL-α-AAA in OCT at 3 months after subretinal injection remained the same as we observed at 1 month after injection (data not shown). 

Changes in FCP, FFA, and FAF

FCP, FFA, and FAF were performed at 12, 30, 60, and 135 days after injection. Subretinal injection created blebs in the macular region, the aftereffects of which were easily identified with FCP (Figs. 4A–D, 4I, 4K). No obvious retinal damage was observed in eyes injected with BSS (Fig. 4A). However, injections of DL-α-AAA as low as a 5-mM concentration induced retinal whitening that persisted for the whole period of observation (Figs. 4B–D, 4I, 4K). The central foveal macular pigment, however, appeared only slightly affected in eyes receiving DL-α-AAA (Figs. 4B–D, 4I, 4K, arrows) compared with eyes injected with BSS (Fig. 4A). FFA did not reveal any obvious vascular leak in eyes injected with DL-α-AAA though patchy fluorescein staining was observed at the sites of needle penetration (Figs. 4F–H, 4J, 4L, asterisks). 

Figure 4.

 

Changes in FCP and FFA after subretinal injection of BSS (A, E) or DL-α-AAA (B–D, F–L). (A–D, I, K), FCP. (E–H, J, L), FFA of eyes in (A–D, I, K) respectively. (B–D, F–H) An eye receiving 5 mM DL-α-AAA at 12, 60, and 135 days after injection, respectively. (I, J) 10 mM DL-α-AAA 135 days after injection. (K, L) 50 mM DL-α-AAA 135 days after injection. Asterisks: sites of needle penetration; arrows: macular pigment in the central fovea.

Assessment of FAF demonstrated that subretinal injection of BSS slightly increased autofluorescence signals in some areas at the edge of the bleb in BSS-injected eyes (Fig. 5A, arrowheads). Somewhat more evenly increased FAF signals were commonly observed in eyes receiving 5, 10, and 50 mM DL-α-AAA (Figs. 5B–D). Noticeably, the usual attenuation of FAF, which is attributed to the macular pigment in the fovea, was persistently observed in eyes receiving BSS or DL-α-AAA (Figs. 5A–D, arrows). 

Figure 5.

 

Increased FAF after subretinal injection of DL-α-AAA. FAF was measured 12 days after subretinal injection of BSS (A) or DL-α-AAA (B–D). Asterisks: sites of needle penetration; arrows: areas in which FAF signals were masked by the existing macular pigment in the central fovea. (A, arrowheads) Increased FAF in borders between normal and BSS-covered areas. Note: the usual attenuation of the foveal FAF by macular pigment (arrows) persists in eyes injected with BSS or DL-α-AAA, and increased FAF signals are commonly observed in areas injected with DL-α-AAA (B–D).

Photoreceptor Loss in Histology after DL-α-AAA Injection

Histologic examinations revealed that subretinal injection of BSS did not induce obvious structural damage to the neuroretina (Fig. 6A). However, subretinal injections of 5, 10, or 50 mM DL-α-AAA resulted in severe loss of photoreceptors and nearly complete absence of the inner and outer segments (Figs. 6B–D). There was no obvious dose-dependent difference in photoreceptor damage caused by different concentrations of DL-α-AAA (Figs. 6B–D). Photoreceptor loss was observed in a broad area, ranging from the central fovea to the paramacular region, consistent with the entire region having been treated with DL-α-AAA (Fig. 6E). 

Figure 6.

 

Photoreceptor damage after subretinal injection of DL-α-AAA in histology. (A–E) Light microscopic images after hematoxylin and eosin staining in the retinas receiving BSS (A), 5 mM (B, E), 10 mM (C), or 50 mM (D) DL-α-AAA 5 months after subretinal injection. Photoreceptor damage was evident in eyes injected with DL-α-AAA (B–D, arrows) but not in the BSS-injected eye (A). (E) Montage showing nearly complete loss of photoreceptors in the macula injected with 5 mM DL-α-AAA. Scale bars, 100 μm.

Müller Cell Changes after Subretinal Injection of DL-α-AAA

To investigate the effect of DL-α-AAA on retinal glial cells after subretinal injection, double-label immunohistochemistry was performed for GS and GFAP immunoreactivity (Fig. 7). In BSS-injected retinas, the GS antibody labeled astrocytes and Müller cells along their entire cellular extent, particularly on the somata within the inner nuclear layer (Fig. 7A, inset, M) and their endfeet at the inner and outer limiting membranes. Although DL-α-AAA induced severe damage to photoreceptors, GS⁺-cells were persistently visible in eyes receiving 5, 10, or 50 mM DL-α-AAA (Figs. 7D, 7G, 7J, insets), despite the nearly complete loss of photoreceptors in the ONL. In BSS-injected retinas, GFAP staining was confined only to filamentous structures in the innermost retina (Fig. 7B), which were likely to represent astrocytes because GFAP staining did not colocalize with GS staining in normal eyes. Increased GFAP immunoreactivity that colocalized with GS staining was commonly observed in the GCL and INL in eyes injected with DL-α-AAA (Figs. 7E, 7H, 7K). 

Figure 7.

 

Changes in the retinal glia after subretinal injection. Retinas were double labeled for glutamine synthetase (GS: A, D, G, J) and glial fibrillary acidic protein (GFAP: B, E, H, K) at 5 months after injection. (A–C), BSS. (D–F) 5 mM DL-α-AAA. (G–I) 10 mM DL-α-AAA. (J–L) 50 mM DL-α-AAA. In the BSS-injected eye (A–C), GS antibody labels Müller cells with the somata (indicated by “M” in the inset in A) within the INL and their endfeet at the inner and outer limiting membranes. GFAP staining is confined to filamentous structures in the innermost retina (NFL and GCL). DL-α-AAA induced photoreceptor damage, but most Müller cells (indicated by “M” in insets in D, G, J) survived although increased GFAP immunoreactivities were observed in these eyes (E, H, K). (C, F, I, L) Merged images. Scale bars, 100 μm.

Toxicity of DL-α-AAA in 661W Photoreceptors

Our previous data have shown that DL-α-AAA is toxic to glial cells, particularly to Müller cells, but not to pericytes, retinal vascular endothelial cells, or RPE cells. ³¹ In this study, we further tested the cell viability and metabolic activity of DL-α-AAA in 661W, a cone-derived photoreceptor cell line (Fig. 8). Treatment of 661W cells with 5, 10, and 50 mM DL-α-AAA for 16 hours resulted in 8.1% ± 1.9%, 35.6% ± 2.4%, and 86.4% ± 3.5%, respectively, of cells dying (Figs. 8D–F; P < 0.01 vs. 0 mM; n = 8 in each group). Changes in cell metabolic activity were quantitatively measured using the Alamar blue assay (Fig. 8G). A significant inhibition of cell metabolic activity was observed with 5, 10, and 50 mM DL-α-AAA, and the reductions in metabolic activity seemed to be dose dependent (Figs. 8A–G). 

Figure 8.

 

Cell viability and metabolic activity of 661W photoreceptors after treatment with DL-α-aminoadipic acid (DL-α-AAA). (A–F) 661W cells were treated with different concentrations of DL-α-AAA for 16 hours and then stained with calcein-AM for live (green) and propidium iodide for dead (red) cells. Scale bars, 100 μm. (G) Cellular metabolic activities as quantitatively measured by Alamar blue assay. 661W cell were treated with DL-α-AAA for 16 hours and then incubated in culture media contained 10% Alamar blue for 3 hours Each column represents the mean ± SEM (n = 8). *P < 0.01 versus normal medium without DL-α-AAA.

Discussion

Although DL-α-AAA is considered to be a glia-specific toxin in the central nervous system, we found in this study that the predominant effect of submacular injection of DL-α-AAA was to destroy the photoreceptors, as reflected by changes in OCT and histology, together with pronounced attenuation of mfERG responses from the treated area. Immunostaining of GS and GFAP demonstrated that, although they were activated, most Müller cells survived the injury. Moreover, despite the loss of photoreceptors, macular pigment did not change remarkably in the central fovea, nor was the loss of photoreceptors induced by DL-α-AAA associated with changes in the retinal vasculature when examined with fluorescein angiography. These observations provide further information on the contributions of different cell components to the pathogenesis of MT2. 

It is believed that the macular pigment in the fovea is located in the Henle fiber layer, which corresponds anatomically to photoreceptor axons and Müller cell processes. ^19,21 Recent studies have shown that a unique reduction of macular pigment optical density within the central retina is a common finding in patients with MT2. ^{1 –3} The mechanism of deposition of macular pigment in the central retina is unclear. A number of functions have been proposed for macular pigment. ^18,20,38 Its absorption maximum at 460 nm results in blue light filtration, which may reduce photic damage and glare, minimize the effects of chromatic aberration on visual acuity, improve the distinction of fine detail, and enhance contrast sensitivity. ^18,20,38 Moreover, its neutralization of reactive oxygen species may protect the outer retina, retinal pigment epithelium, and choriocapillaris from oxidative damage. ²⁰ Recent studies have suggested that degenerative processes causing impairments in transport and storage of macular pigment may play a causative role in the pathogenesis of MT2. ^2,3,17  

The functions of Müller cells in the fovea centralis have been largely overlooked. The human fovea is marked by a pit with elevated sides in which the inner retinal layers are piled up so that only photoreceptors and surrounding Müller glial cell processes remain. ²¹ In 1969, Yamada ²² reported that the foveola contains an inverted cone-shaped zone of Müller cells (Müller cell cone). The truncated apex of the cone was located at the outer limiting membrane where there were no receptor nuclei. The base of the cone formed the floor of the fovea centralis and extended into the area of the clivus in the perifoveolar region. In this study, subretinal injection of DL-AAA in the macular region nearly completely eliminated photoreceptors in the central fovea. The damage to photoreceptors was confirmed by OCT, mfERG, and histologic studies. If the macular pigment were stored mainly in cone photoreceptors in monkeys, we would expect to see dramatic changes in the macular pigment in regions in which DL-α-AAA was injected. However, the macular pigment remained relatively unchanged in eyes injected with DL-α-AAA, whereas most Müller cells, albeit activated, survived the injury. This suggests that the storage of macular pigment may predominantly reside in the processes of Müller cell cones rather than the cone photoreceptors. Observations from several recent studies suggest that Müller cell dysfunction may play a critical role in the pathogenesis of MT2. ^1,8,15,16,39  

The results obtained from our study may not exclude the contribution of photoreceptor damage to MT2. Little is known about the processing of macular pigment. If the storage of macular pigment is localized primarily to the Müller cells, disturbance of binding and transport of lutein and zeaxanthin in photoreceptors may also contribute to the pathologic processes of MT2. Bhosale et al. ^40,41 have shown that the uptake and transport of lutein and xanthophyll carotenoids into the foveal region are mediated by specific binding proteins, such as a membrane-associated lutein-binding protein (LBP) and a Pi isoform of glutathione-S-transferase (GSTP1), a zeaxanthin-binding protein. Immunolocalization with antibodies directed against either carotenoid-binding protein or GSTP1 showed specific labeling of rod and cone inner segments, especially in the mitochondria-rich ellipsoid region in humans and monkeys. ^40,41 MT2 is a chronic and slowly progressive disease that evolves over many years, whereas the intervention used in this study is an acute event with a subsequent follow-up period of only few months. Slow processing of macular pigment could leave its density virtually unaffected within several months even if the binding and transport of lutein and zeaxanthin in photoreceptors are interrupted. 

To investigate whether DL-α-AAA induces vascular abnormalities in the macular region, FFA was performed periodically from 2 weeks to 5 months after submacular injection of DL-α-AAA. No vascular leakage was detected by FFA, nor did subsequent histologic examination show vascular abnormalities in eyes receiving DL-α-AAA. This is different from what we observed in rats in which subretinal injection of DL-α-AAA, similar to siRNA-targeting glutamine synthetase, induced retinal vascular leak and telangiectasis. ³¹ In the rat study, DL-α-AAA severely disrupted Müller cells, but most cells in the ONL survived. ³¹ In the present study, by contrast, the subretinally delivered DL-α-AAA nearly completely eliminated cone photoreceptors, whereas most Müller cells survived the injury. Our in vitro results confirmed that DL-α-AAA is highly toxic to cone photoreceptors after 16 hours of incubation in vitro. The discrepant changes in retinal vasculature observed in rats and monkeys may be due to the anatomic differences between these two species. The rat retina contains <2% of cones, ³³ but the fovea in humans and primates contains the highest density of cone photoreceptors in the retina (199,000/mm²), with sparse rod photoreceptors in the central 250 μm. ³² It is possible that DL-α-AAA was predominantly absorbed by cone photoreceptors in the monkey study; thus, its effect on the somata of Müller cells in the INL was dramatically attenuated. The different patterns of cone photoreceptor injury and Müller glial disruption between rats and monkeys may account for the variations of vascular responses in these two species. There is evidence that, under pathologic conditions, neurons and glial cells interact with blood vessels by releasing specific neurotropic factors such as nerve growth factor, brain-derived neurotrophic factors, and glial cell line-derived neurotrophic factor, which may also couple with other angiogenic factors, such as vascular endothelial growth factor (VEGF), to influence vascular function. ^{42 –45} In the rat study, we did observe overexpression of VEGF in photoreceptors after subretinal injection of DL-α-AAA. ³¹ The photoreceptor toxicity of DL-α-AAA in monkeys may virtually eliminate dysfunctional photoreceptors as candidates contributing to the disease phenotype in MT2. For instance, if the vascular alterations in MT2 are secondary to a chronic photoreceptor alteration with subsequent VEGF expression, this would not be modeled by the intervention described in the present study, in which the outer nuclear layer was obliterated. Barthelmes et al. ⁴⁶ have reported a case of MT2 in which vascular changes, including vascular telangiectasis, right-angled vessels, and diffuse-late hyperfluorescence, developed in both eyes many years after diagnosis with cone dystrophy. 

Overall, our data showed that subretinal injection of DL-α-AAA in the macular region of the monkey induced severe photoreceptor damage, but Müller cells survived the injury. Photoreceptor injury induced neither central macular pigment depletion nor vascular telangiectasis. Further research making use of more specific approaches, such as viral vectors and cell-specific promoters to interfere with different retinal components, including Müller cells and photoreceptors in the primate macula, is warranted to establish the precise role of each cell type in MT2. ^{47 –50}  

Footnotes

 Supported by a grant from Lowy Medical Research Institute, Australia and an International Program Development Fund, the University of Sydney. Mark C. Gillies is a Sydney Medical Foundation Fellow.

Footnotes

 Disclosure: W. Shen, None; J. Zhang, None; S.H. Chung, None; Y. Hu, None; Z. Ma, None; M.C. Gillies, None

The authors thank Jing Shi and Hong Shen for animal care and anesthesia, Changguan Wang for advice on subretinal injection, Xinrong Lu, Fang Qian, and Wei Wang for assistance in clinical examinations, and Ying Li for histologic preparations.