α2-Adrenergic Receptors and Their Core Involvement in the Process of Axonal Growth in Retinal Explants

Abstract

Purpose.: To determine the patterns of α2-adrenergic receptor (α2-AR) subtype expression in normal and degenerated retinas and to analyze the response of these receptors to the α2-AR agonist brimonidine tartrate (BT).

Methods.: The binding characteristics of α2-ARs in the retina were evaluated in experimental and matching sham groups by in vitro quantitative autoradiographic saturation with [³H]-clonidine. Retinal explants from juvenile and adult rats with either elevated intraocular pressure or after optic nerve crush (ONC) were cultured with BT over 96 hours in vitro to analyze the effects of BT on axonal growth by videomicroscopy and axon counting. Changes in retinal protein expression by BT were monitored by two-dimensional gel electrophoresis (2D-PAGE) and matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS).

Results.: The total number of α2-ARs in the retina increased significantly after ONC compared with the sham group. BT supported axonal growth in the juvenile, glaucomatous, and injured retinas (P < 0.004) most effectively at a concentration of 0.001 mg/mL, without influencing the axonal growth rate. Immediate supplementation of BT was more effective than delayed supplementation (P < 0.001). Proteomic analysis revealed treatment-specific expression patterns of glial fibrillary acidic protein (GFAP), glucose-related protein (GRP)58, platelet-activating factor (PAF), and laminin-binding protein (LBP).

Conclusions.: These data are the first to show differences in α2-AR expression in normal and degenerated retinas. BT supports neuronal growth in cultured retinal pieces, suggesting that α2-ARs play a role in retinal metabolism.

Progressive loss of retinal ganglion cells (RGCs) is a hallmark of traumatic or glaucomatous injury of the optic nerve. ^{1 –4} Although glaucoma is a chronic injury, in contrast to acute traumatizing lesions, recent research suggests that the neurodegenerative process in glaucoma exhibits similarities to the degeneration that occurs in the context of traumatic injuries. In addition to the primary injury in neurodegenerative diseases, a secondary degeneration is assumed that is caused by the elevation of toxic substances in the nerve's extracellular milieu. Toxic agents such as glutamate are thought to cause damage of neurons that escaped the primary trauma or were only marginally affected by it. Drugs that are capable of neutralizing the toxicity of such substances, competing with their activities, or increasing the resistance of any remaining viable neurons to the stressful conditions can be considered to have neuroprotective potential. ^5,6  

α2-Adrenoreceptors (α2-ARs), which are involved in vascular autoregulation, ^7,8 are supposed to play a key role in neuroprotection and neuroregeneration. Cellular signaling cascades, such as those responsible for inhibition of adenylate cyclase activity, inhibition of calcium channels, opening of the potassium channels in the cells, and inhibition of proapoptotic mitochondrial signaling, are all activated through α2-ARs and participate in neuroregenerative processes. ^9,10 α2-AR activation is coupled with K⁺ activation, leading to neuronal hyperpolarization, reduced excitability, and release of neurotransmitters during ischemia. ¹¹ Activation of the α2-AR pathway protects RGCs from cell death in various models of acute and chronic injury. ^12,13  

Three different α2-ARs subtypes are expressed in the retina the α2A-, α2B- and α2C-AR. ^14,15 There are lines of evidence that α2-AR agonists exert their neuroprotective effect via the α2_A-AR, ¹⁶ mainly located in the ganglion cells. ¹⁵ The α2A adrenoreceptors in these cells provide a site for the neuroprotective activity of α2 agonists in different models of retinal injury. ^{17 –21} Up to now, there have been limited data on the regulation of injury-induced α2_A-AR expression, which could alter the susceptibility of RGCs to adrenergic agonists. 

The efficacy of α2-AR agonists, such as dexmedetomidine ²² and clonidine, ²³ in reducing the negative effects of brain ischemia and providing neuroprotection has been demonstrated in several experimental studies and is most likely mediated through α2-AR activation, in that it is blocked by corresponding α2-AR antagonists. ¹⁸ The α2-AR agonist BT is used clinically as a topical ocular hypotensive agent in glaucoma patients, ²⁴ as well as in postoperative patients, to control intraocular pressure (IOP). ²⁵ Brimonidine lowers the IOP by reducing the production of aqueous humor and increasing the uveoscleral outflow. ²⁶ After topical application, a sufficient concentration of brimonidine is achieved to activate α2-ARs within the ocular tissues. ²⁷  

Other studies have demonstrated the efficacy of brimonidine in increasing the survival of RGCs after various types of injury in vivo. Intraperitoneal pretreatment with BT significantly enhances RGC survival and retinal function after ONC. ^11,28,29 In addition, early RGC loss was prevented after topical administration of brimonidine before transient retinal ischemia induced by vessel ligation. ^{30 –33} However, although retinal regeneration assays monitoring axonal regrowth after various types of RGC injuries have been conducted, ^34,35 the effects of BT on neuronal growth have yet to be demonstrated. The mechanisms underlying the neuroprotective effect remain to be elucidated, but treatment with brimonidine appears to result in the inhibition of glutamate and aspartate accumulation, ¹⁸ the upregulation of antiapoptotic genes such as bcl-2 and bcl-xl, and the production of neurotrophic molecules such as fibroblast growth factor. ^13,29  

A tool suitable for the study of treatment-associated changes in tissue metabolism by analyzing differences in proteomics is 2-D PAGE with subsequent MALDI-MS. ³⁶ Proteomics allows identification of protein–protein interactions and disease or treatment-associated proteins encoded by the genome. Differences in protein expression in the retinas of normal rats and diseased rats can be shown, ³⁷ and the proteomic tool helps to identify novel pharmacologic targets. ³⁸  

The present study was undertaken to explore the distribution, binding, and saturation properties of α2-ARs within the retina of normal rats and rats in which the optic nerves have been injured. The cellular localization of the α2_A-ARs was assessed with immunohistochemistry. The effects of the α2-AR agonist BT on axonal growth were evaluated in retinal pieces that were cultured in vitro to avoid microcirculation-mediated effects. Finally, we used a proteomic approach to show that BT specifically regulates retinal proteins, thereby showing that α2-ARs are probably involved in the vascularization-independent regulation of retinal metabolism. 

Materials and Methods

Animals and Drugs

All experiments were conducted in strict accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Sprague-Dawley (SD) rats were housed in a standard animal room in a 12-hour light–dark cycle, with food and water provided ad libitum. A total of 168 SD rats were used (Table 1). Surgical procedures were performed on rats that weighed 180 to 250 g or were 16 days old (P16) under general anesthesia induced by a mixture of 2 mg/kg ketamine and 2 mg/kg xylazine (Ceva-Sanofi, Düsseldorf, Germany), administered intraperitoneally. After surgical intervention, gentamicin (Gentamytrex; Dr. Mann Pharma, Berlin, Germany) was applied topically. All surgical manipulations were unilateral, with the contralateral eye serving as the corresponding control. BT, which was provided by Allergan Laboratories (Irvine, CA), was dissolved in S4 medium (Astrocyte Microglia Growth Medium; Promocell, Heidelberg, Germany) or Hanks' balanced salt solution (HBSS; Merck, Darmstadt, Germany) in specific concentrations. 

Table 1.

 

Experimental Approaches

Experimental Approach	n	SD Rat Type	In Vivo		In Vitro
			Experimental Manipulation	Time	Experimental Manipulation	Time
α2-AR autoradiographic saturation	4	150–250 g	Left ON exposure without crush (sham)	5 d, 15 d	[3H]-clonidine (0.8–8 nM)
			Left ONC	5 d
			Left ONC	15 d
α2-AR localization	4	150–250 g	Left ON exposure without crush (sham)	5 d, 15 d	[3H]-clonidine (0.8–8 nM)
			Left ONC	5 d
			Left ONC	15 d
α2A-AR, Rhodopsin immunohistochemistry	4	150–250 g	Left ON exposure without crush (sham)	5, 15 d
			Left ONC	5 d
			Left ONC	15 d
Effect of BT concentrations on axonal growth	8	P16			Sham, 1, 0.1, 0.005, 0.01 0.005, 0.001 mg/mL BT	96 h
Effect of time point of BT on axonal growth	8	P16			0.001 mg/mL BT after 3 h, 1/2 h and immediate	96 h
Effect of BT on growth of glaucomatous tissue	8	150–250 g	Left eye cauterization of 2 episcleral veins	14 d	Sham, 0.001 mg/mL BT after 3 h, 1/2 h and immediate	96 h
Effect of BT on axonal growth of injured retina	8	150–250 g	Left ONC + LI	3 d	Sham, 0.001 mg/mL BT immediate	96 h
Effect of BT on axonal growth rate	3	P16			Sham, 0.001 mg/mL BT immediate	96 h
2D-PAGE, MALDI-MS	3	P16			Sham, 0.001 mg/mL BT immediate	96 h
PAF, GRP58, GFAP immunohistochemistry	3	P16

Unilateral Optic Nerve Crush

To analyze the expression pattern of the various α2-AR subtypes in the normal and injured retinas of female adult rats (weighing 180–250 g), we crushed the optic nerve of the left eye (ONC) intraorbitally in the experimental animals (n = 8), sparing the retinal artery in vivo. In brief, after a skin incision was made near the superior orbital rim, the optic nerve was exposed and crushed with jeweler's forceps, approximately 0.5 mm behind the eye cup. A group of sham-surgery animals (n = 8) in which the optic nerve was exposed but not crushed served as the control in experiments analyzing α2-AR expression. 

For analyzing the impact of α2-AR agonist BT supplementation of cell cultures on axonal growth, the rats underwent ONC, together with additional lens injury inflicted by a capillary tube in vivo, as it has been shown that axotomy and lens injury enable axonal growth in vitro. ³⁹ The rats were allowed to survive for 3 days after this treatment. 

Autoradiographic Studies

For in vitro receptor autoradiographic saturation and localization studies, sham-surgery and experimental (left ONC; n = 12) animals were deeply anesthetized with ketamine (40 mg/kg body weight) and killed by decapitation 6 or 16 days after the ONC. Their eyes were rapidly removed, frozen at −40°C in dry-ice–cooled isopentane and kept at −80°C until further use. Sagittal, 10-μm-thick sections were cut with a cryostat (CM 1500; Leica, Bremen, Germany), mounted on chrome alum/gelatin-coated slides, and left to dry before the autoradiographic binding experiments. 

The α2-AR agonist [³H]-clonidine (24 Ci/millimole; Amersham Biosciences Europe, Freiburg, Germany) was used to determine the ligand-specific binding sites by means of in vitro autoradiographic techniques. For the saturation studies, the concentration of [³H]-clonidine ranged from 0.8 to 8 nM. For the quantitative localization studies, the titrated ligand was used at a concentration of 4 nM ([³H]-clonidine). Nonspecific binding was determined in the presence of 10 μM clonidine (Tocris Cookson, Bristol, UK). Adjacent tissue sections were prewashed in 50 mM Tris-HCl buffer (pH 7.7) at 4°C for 30 minutes before incubation at room temperature (RT) with [³H]-clonidine (60 minutes), at the aforementioned concentrations. Free radioligand was removed by washing in ice-cold Tris buffer for 5 and 10 minutes, and the sections were dried in a cold-air stream. Slides, along with appropriate radioactive standards ([³H]-Microscales; Amersham Biosciences Europe), were exposed at 4°C for 8 to 12 months to tritium-sensitive film (Hyperfilm; Amersham Biosciences Europe) developed (D-19; Eastmak Kodak, Rochester, NY), fixed (Acidofix; Agfam Leverkusen, Germany), and cleared in running water. Quantitative image analysis of the autoradiograms was performed with a computerized image-analysis system comprising a black-and-white CCD camera (XC-77CE; Sony, Tokyo, Japan) that was connected to a computer (G4 Macintosh; Apple Computer, Cupertino, CA) via a frame grabber (LG-3; Scion Corp., Frederick, MD). NIH-Image software (NIH Image, ver 1.61, developed by Wayne Rasband, National Institutes of Health, Bethesda, MD; available at http://rsb.info.nih.gov/ij/index.html) was used for densitometry of various brain nuclei and subsequent expression of specific binding, which is expressed as femtomoles of receptor per milligrams of tissue. The equilibrium dissociation constant (K _d) and the maximum binding sites (B _max) for [³H]-clonidine were determined (Radlig-Ebda-Ligand; Biosoft, Ferguson, MO) with an iterative fitting program based on a multisite binding model, expressed by the Scatchard-Rosenthal equation: B/F = (B/K _d) + (B _max/K _d), where B is the density of bound receptors, B _max is the total receptor density, and F is the free-ligand concentration. 

Induction of Glaucoma and IOP Measurement

IOP was elevated in the rats (weighing 180–250 g) by cauterizing two episcleral veins in the left eye, as described previously, ⁴⁰ resulting in blockage of 50% of the venous outflow. IOP measurement was begun 7 days after cauterization and was determined with the rats under light anesthesia by ether inhalation, as anesthetics cause a reduction in IOP, ³⁷ and topically applied 0.5% proparacaine (URSA-Pharm, Saarbrucken, Germany). All measurements were performed daily between 9 AM and 12 PM over 7 days with a handheld tonometer (Tono-Pen XL; Mentor, Norwell, MA). Ten tonometer readings were taken directly from the display of the instrument for each eye measurement, recorded, and averaged. “Off” (or outlier) readings and instrument-generated averages were ignored. Animals in which IOP returned to normal during the 7-day measurement period were excluded from the data analysis. Eyes in which the IOP had stayed elevated for 7 days were enucleated and used for further experimental steps. Thirty-two animals were used for this part of the study. The baseline IOP (15.0 ± 0.5 mm Hg) in nontreated rats (n = 8) served as the control. Cauterization elevated IOP to a mean value of 20.3 ± 1.8 mm Hg (n = 32). An IOP elevation of 1.4-fold is accepted as a model for glaucoma. ⁴⁰ The mean IOP in rats appeared to be nearly identical with that recorded in awake humans, rabbits, and anesthetized monkeys. 

Preparation and Dissection of the Retina

For testing the effects of BT on axonal growth in vitro, the rats were killed using an atmosphere of CO₂. The eyecups were taken and placed in HBSS, where all subsequent preparation steps were conducted under sterile conditions. Retina was isolated, flat-mounted on a nitrocellulose filter with fine forceps, divided into eight wedge-shaped pieces, and placed with the ganglion cell layer on a substrate consisting of polylysine (Sigma-Aldrich, Munich, Germany) and laminin (Roche, Mannheim, Germany). S4 growth medium containing 1% gentamicin and BT was added (3 mL). Cultures were maintained for 96 hours at 37°C in a humidified atmosphere containing 5% CO₂. The number of outgrowing axons was determined after 48, 72, and 96 hours in culture, with the aid of an inverted phase-contrast microscope (Axiovert 135; Carl Zeiss, Jena, Germany) at a magnification of ×200. Axons encountered up to an imaginary line of 50 μm from the explant's margin were counted. Fasciculation occurred in areas in which there were many fibers, sometimes preventing the clear distinction of individual axons within the bundles. In such cases, a single bundle was counted as one axon throughout all groups. Each experimental group comprised eight eyes, yielding 64 slices. 

Application of BT In Vitro

The effect of BT on axonal regeneration in vitro was analyzed in normal juvenile (P16 rats) retinal tissue after IOP elevation and ONC. To ascertain the most effective BT concentration, we tested the growth-promoting activity of BT on normal retinal tissue of P16 rats. BT, dissolved at concentrations of 1.0, 0.1, 0.05, 0.01, 0.005, and 0.001 mg/mL in S4 medium, was added to the cultures. The cultures from one group were not given BT and served as the control. The most effective BT concentration thus defined (0.001 mg/mL) was tested for its axonal growth-promoting activity on rats after IOP elevation and ONC+LI in vitro. Again, one group was left without BT and served as the control group. 

To establish whether the time at which BT was added to the culture plays a role in axonal growth, we added the most effective concentration of BT (0.001 mg/mL) to the retinal cultures at three different time points. In brief, 0.001 mg/mL BT was dissolved in HBSS and S4 medium and added to the retinas immediately, 30 minutes, or 3 hours after eye enucleation. This experimental series was conducted on retinal samples from P16 rats, rats with elevated IOP, and rats after ONC+LI. 

Growth Rate

The most effective concentration of BT (0.001 mg/mL) was used to assess the influence of this α2-ΑR agonist on the growth rate of axons in culture. After explantation, the axonal growth of the RGCs of normal retinal tissue, cultured for 96 hours with BT given immediately or without BT, was monitored by computer-assisted direct microscopy. The growth rate was calculated by measuring the advancing axonal growth cones of a random selection of 30 fibers per explant (n = 3) during several 3-hour periods. A total of 90 axons in each group were examined, to calculate the rate of growth, which is expressed in micrometers per hour. 

Two-Dimensional Gel Electrophoresis and Proteomics

2D-PAGE was performed on retinal samples of P16 rats, cultured with and without 0.001 mg/mL BT, after 96 hours in vitro. Explants were harvested and used for proteomic analysis by 2D-PAGE and MALDI-MS. 2D-PAGE was conducted according to a method initially described by O'Farrell. ³⁸ The retinal explants were boiled in 10% sodium dodecylsulfate (SDS; Sigma-Aldrich, Taufkirchen, Germany) and homogenized in 2D-PAGE lysis buffer (7 M urea, 2 M Thiourea; Merck), 4% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propan sulfonate (USB, Cleveland, OH), 40 mM Tris base (Roth, Karlsruhe, Germany), 1 mM phenylmethylsulfonyl fluoride (Sigma-Aldrich), and 10 mM dithiothreitol (Roche, Mannheim, Germany). The final SDS concentration was 0.25%. Soluble protein (200 μg, according to the Bradford test) together with 2% immobilized pH-gradient (IPG) buffer (pH 3–10; Amersham Biosciences Europe) and 20 mM dithiothreitol were loaded on strips (pH 4–7, 18 cm; Immobiline DryStrip; Amersham Biosciences Europe) and rehydrated overnight. The rehydrated strips were focused on an electrophoresis unit (Multiphor II; Amersham Biosciences Europe) at approximately 80 kVh. Focused IPG strips were incubated twice for 15 minutes in equilibration solution (50 mM Tris-HCl [pH 8.8], 6 M urea, 30% glycerol, 2% wt/vol SDS, and a trace of bromophenol blue [Merck]), with 1% β-mercaptoethanol added to the first and 2.5% iodo-acetoacetamide added to the second equilibration step. 

For the second dimension, the equilibrated IPG strips were fixed with 0.5% wt/vol melted agarose (Merck) on homogeneous 12.5% SDS gels (Rotiphorese Gel 30; Carl Roth, Karlsruhe; Germany). The proteins were separated by vertical SDS-PAGE (BioRad, Munich, Germany) according to Laemmli. ⁴¹ Protein spots were initially labeled with colloidal Coomassie Blue G 250 (Merck). Spots were manually excised from the gel, tryptically digested in the gel, extracted, purified (Zigtips, microbed C18; Millipore, Bedford, MA), and then subjected to MS analysis. Peptide maps were generated by MS (TOF-Spec-2E; Micromass, Manchester, UK), and selected retinal peptides were sequenced by nano-high-performance liquid chromatography-MS/MS (Ultimate; LC Packings, Amsterdam, The Netherlands; Esquire 3000; Bruker Daltonics, Bremen, Germany). Six gel replicates were compared. National Center for Biotechnology Information (National Institutes of Health, Bethesda, MD, available in the public domain at www.ncbi.nlm.nih.gov) and SWISS-PROT (Swiss Institute of Bioinformatics, Geneva, Switzerland, http://www.expasy.ch/sprot) databases were searched using Mascot software (Matrix Science, London, UK). Additional image analysis was performed on gels stained with silver nitrate. 

Immunohistochemistry

For the cellular localization of α2_A-ΑRs, sham surgery and experimental (left ONC) rats (n = 12 in both groups) were deeply anesthetized with intraperitoneal chloral hydrate (7%) injection and then perfused transcardially with phosphate-buffered saline (PBS) followed by a fixative solution (4% paraformaldehyde in 0.1 M phosphate buffer [PB]; pH 7.4). The eyes were removed, postfixed in the same fixative containing 10% sucrose for 4 hours at 4°C, refrigerated overnight in 0.1 M PB with 20% sucrose for cryoprotection, and then frozen in isopentane at −40°C. Free-floating cryosections were cut at a thickness of 10 μm and collected in 0.1 M PB solution (pH 7.4). The sections were thoroughly washed in 0.01 M PBS (pH 7.4) and then incubated in a blocking solution containing 1.5% normal horse serum, 5% bovine serum albumin (BSA), and 0.1% Triton X-100 in 0.01 M PBS for 60 minutes at room temperature. 

For single labeling, sections were washed for 10 minutes at RT with 3% H₂O₂ in PBS to suppress endogenous peroxidase activity and then rinsed in buffer before blocking. They were then incubated overnight at 10°C with a polyclonal antibody, goat anti-α2_A-AR (diluted to 1:100 with 0.1% Triton X-100 in 0.01 M PBS and 2% BSA; Santa Cruz Biotechnology, Santa Cruz, CA). After being rinsed in PBS, the sections were incubated with a biotinylated anti-goat antibody solution (diluted 1:200 in 0.01 M PBS; Vector Laboratories, Peterborough, UK) for 2 hours at RT, followed by an avidin-biotin-peroxidase solution (ABC Elite Kit, A and B both diluted to 1:100 in 0.1% Triton X-100 in 0.01 M PBS; Vector Laboratories) for 1 hour in the dark at room temperature. Immunoreactive cells were visualized via the peroxidase-catalyzed polymerization of 0.05% 3,3′-diaminobenzidine tetrahydrochloride (Vector Laboratories) in 0.01% H₂O₂ buffer solution (pH 7.5). The sections were washed repeatedly in cold buffer, mounted on chrome alum/gelatin-coated slides, and left to dry. To facilitate the identification of retinal layers and to enhance the contrast, we used methyl green as a counterstain (1% methyl green in double-distilled H₂O; Sigma-Aldrich). Slides were dehydrated and coverslipped with a rapid mounting medium (Entellan; Merck). 

To characterize the co-localization of the α2_A-AR with cellular retinal structures, we performed double labeling with the α2_A-AR antibody (Santa Cruz Biotechnology) detected by a rhodamine-labeled secondary antibody (Sigma-Aldrich) and a rhodopsin antibody as a cell-specific marker for photoreceptors (Chemicon), detected by a Cy-2 labeled secondary antibody (Dianova, Hamburg; Germany), serving as the second antibody. A detailed examination of the localization and co-localization of α2_A-AR has been performed in detail by WoldeMussie et al. ¹⁵  

Further characterization of the protein expression of GFAP, GRP58, and PAF in the retina was preformed by immunohistochemistry. Retinal flat mounts obtained from groups of immediately BT-treated (0.001 mg/mL) and untreated, sham-surgery P16 rats (cultured for 96 hours) were fixed overnight in 4% paraformaldehyde, embedded (Tissue-Tek; Sakura Finetek, Torrance, CA), and cut into 10-μm-thick sections (Kryostat 2800 Frigocut E; Reichert-Jung, Wetzlar, Germany) or stained without cutting for immunohistochemistry of growing axons from retinal explants. The immunohistochemical procedure just described was then conducted with a monoclonal rabbit anti-GFAP antibody (Sigma-Aldrich) diluted in fetal calf serum (FCS) to 1:100; a polyclonal rabbit anti-GAP-43 antibody (Chemicon) diluted in FCS to 1:100; or a polyclonal goat anti-PAF acetylhydrolase antibody (Santa Cruz Biotechnology) or polyclonal anti-goat STAT3 (Sigma-Aldrich) as a cross-reacting marker for GRP58, ⁴² diluted in FCS to 1:100, as the primary antibody; and anti-rabbit/anti-goat Cy-2 antibody (Dianova) or anti-rabbit/anti-goat rhodopsin antibody (Chemicon) diluted in FCS to 1:200 and 1:1000 as the secondary antibody. Nuclei were counterstained with Dapson. Finally, the slides were coverslipped with antifade mounting medium (Mowiol; Hoechst, Frankfurt, Germany) and viewed with the appropriate filter on a microscope equipped with epifluorescence (Axiophot; Carl Zeiss Meditec). Table 2 provides a list of the immunohistochemical markers used in the study. 

Table 2.

 

Markers Used in Immunohistochemical Approaches

Antibodies	Isotype	Dilution	Company
α2A-Adrenoreceptor	Pc anti-goat IgG	1:100	Santa Cruz Biotechnology, Santa Cruz, CA
GFAP	Mc anti-rabbit IgG	1:100	Sigma-Aldrich, St. Louis, MO
Neurofilament	Mc anti-rabbit IgG	1:200	Sigma-Aldrich
PAF	Pc anti-goat IgG	1:100	Santa Cruz
Rhodopsin	Mc anti-mouse IgG	1:100	Chemicon, Temecula, CA
STAT 3	Pc anti-goat	1:100	Sigma-Aldrich

 Pc, polyclonal antibody; Mc, monoclonal antibody.

Data Evaluation

All data regarding the number of axons per retinal explant are presented as the mean ± SD. Data were analyzed with a test for two independent samples (SPPS, SPSS, Inc., Chicago, Il; Statistica ver. 7; StatSoft, Tulsa, OK), to examine Gaussian distribution, and processed using the independent-samples t-test (Gaussian) and Kruskal-Wallis H test (not Gaussian). The remaining quantitative data were analyzed by two-way ANOVA (Statistica ver. 7). If the distribution was not Gaussian, the Kruskal-Wallis H test was used. Post hoc analyses with the Tukey HSD test were performed to identify possible differences among the experimental groups. 

Selected images from film autoradiograms were digitized with the aforementioned black-and-white image-analysis system. High-resolution photomicrographs were taken with a color CCD camera (DXC-950P; Sony) on a microscope (Eclipse E800; Nikon, Tokyo, Japan) connected to a computer via a frame grabber (CG-7; Scion Corporation). Figures were prepared with image management software (Photoshop; Adobe Systems, San Jose, CA), and the overall brightness and contrast were adjusted without retouching. 

Results

Autoradiographic Saturation Studies of Sham-Surgery and ONC Retinas

Scatchard analysis of [³H]-clonidine binding showed saturable high-affinity binding in the retina, with K _d values ranging between 0.75 and 1.36 nM (Table 3, Fig. 1). Sham-surgery retinas exhibited an average affinity of 1.33 nM and moderate densities (B _max ≈ 42 femtomoles/mg tissue) of α2-ARs (nonspecific binding ranged between 5% and 10% in all cases). Radioligand binding at 5 days after ONC (short-term) showed no significant differences of either binding-site affinity or density. However, 15 days after the crush (long term) there was a significant increase in the total number of binding sites (two-way ANOVA, P < 0.05; Tukey HSD = 0.023) in the ipsilateral retina, compared with the levels in the sham-surgery animals. 

Table 3.

 

α2-AR Binding

	B max (femtomoles/mg)	K d (nM)
Sham surgery
Ipsilateral	44.25 ± 1.72	1.33 ± 0.10
Contralateral	39.90 ± 1.84	1.32 ± 0.02
5 Days after ONC
Ipsilateral	36.98 ± 3.59	1.03 ± 0.02
Contralateral	32.65 ± 0.64	0.64 ± 0.11
15 Days after ONC
Ipsilateral	69.06 ± 6.68* †	1.36 ± 0.21
Contralateral	52.50 ± 0.75*	0.75 ± 0.14

 Receptor-ligand saturation of α2-ARs in the rat retina after ONC. Data are expressed as the mean ± SD.

*  Statistically significant difference compared to 5 days after ONC.

†  Statistically significant difference compared to sham-operation animals (2-way ANOVA. P < 0.01; post-hoc Tukey HSD.

Figure 1.

 

Typical Scatchard and saturation plots of α2-AR binding using the specific binding of 3H RX821002. (A) Scatchard and saturation (inset) plots of 3H-RX821002–specific binding in the ipsilateral retina of a sham-surgery rat. (B) Scatchard and saturation (inset) plots of 3H-RX821002–specific binding in the ipsilateral retina 15 days after ONC.

Quantitative Laminar Pattern of α2-AR Binding Sites in Sham-Surgery and ONC Retinas

The distribution of α2-ARs in the different layers of the retina, as determined by [³H]-clonidine binding, was heterogeneous. The expression level was highest in the external layers (29 ± 1.16 femtomoles/mg tissue protein) and lowest in the inner layers (13 ± 1.13 femtomoles/mg tissue protein) of the retina. Five days after ONC, there was no alteration in the levels of α2-AR binding sites bilaterally (two-way ANOVA, P < 0.05; Tukey HSD = 0.09 ipsilateral; P = 0.12, contralateral). In long-term survival experiments, 15 days after the ONC, both ipsilateral (two-way ANOVA, P < 0.05; Tukey HSD = 0.032) and contralateral (Tukey HSD = 0.048) retinas exhibited elevated α2-AR binding levels compared with sham-surgery retinas (Fig. 2). 

Figure 2.

 

Quantification of α2-ARs in the rat retina, 5 (short-term reaction) and 15 (long-term reaction) days after unilateral ΟΝC. *Statistically significant difference compared with 5 days FTER ONC; #statistically significant difference compared with sham-surgery animals (two-way ANOVA, P < 0.01; post-hoc Tukey HSD).

Cellular Localization of α2_A-AR

Immunohistochemical evaluation of α2_A-AR subtypes in the retina revealed expression in the ganglion cell layer (GCL) in the inner nuclear layer (INL) bilaterally after sham-surgery and 5 days after ONC and in the INL in long-term survival experiments 15 days after ONC. 

Double labeling with the photoreceptor specific-rhodopsin antibody using confocal microscopy better visualized the localization within the retina (Fig. 3). 

Figure 3.

 

Expression of α2_A-AR in the rat retina. (A) Expression of α2-ARs in the ipsilateral retina 15 days after ΟΝC using confocal imaging. Double labeling for α2_A-ARs (red; A, D) and rhodopsin (green; B, E) and merged images of both (C, F) are shown. In addition, each is shown at a higher magnification in (D–F). Scale bar: (A–C) 20 μm; (C–E) 12 μm.

Effect of BT on Axonal Growth In Vitro

Addition of BT to retinal cultures had no visible effect on the attachment of the retinal pieces. BT exhibited dose-dependent growth-promoting activity in vitro (concentrations of 1.0, 0.1, 0.05, 0.01, 0.005, and 0.001 mg /mL were tested; Figs. 4A–D). The control group exhibited 28.3 ± 10.3 fibers after 48 hours, 107.3 ± 24.8 after 72 hours, and 192.7 ± 28.2 after 96 hours. No outgrowth was observed in the 1-mg/mL BT condition. BT concentrations of >0.01 mg/mL (0.1, 0.05, and 1 mg/mL) reduced the number of fibers in vitro, compared with the control group, and concentrations of <0.01 mg/mL significantly (P < 0.04) increased it. The most effective concentration of BT was 0.001 mg/mL, which increased outgrowth to 41.75 ± 8.0 fibers after 48 hours, 197.44 ± 15.8 after 72 hours, and 307.4 ± 17.89 after 96 hours (P < 0.004; Fig. 4E). In adult ONC-treated retinal tissue, administration of 0.001 mg/mL BT resulted in enhanced axonal growth to 21.3 ± 2.4 fibers after 48 hours, 40.6 ± 6.3 after 72 hours, and 101.5 ± 5.4 after 96 hours; the corresponding numbers for the sham-surgery group are 10.5 ± 1.1, 19.5 ± 1.4, and 36.6 ± 2.9, respectively (P < 0.009; Fig. 5A). In glaucomatous retinal explants, 0.001 mg/mL BT increased the number of fibers similarly, to 8.0 ± 0.32 after 48 hours, 36.4 ± 6.4 after 72 hours, and 50.4 ± 1.6 after 96 hours, compared with the sham-surgery group, which exhibited 5.5 ± 1.6, 15.2 ± 4.0, and 31.0 ± 12.9 fibers after 48, 72, and 96 hours, respectively (Figs. 5B, 6). The growth-promoting activity of BT was enhanced significantly by early application (P < 0.001), which increased axonal outgrowth in P16 rats to 78.8 ± 30.6 fibers after 48 hours, 259.3 ± 40.2 after 72 hours, and 502.4 ± 86.5 after 96 hours (Fig. 7A) and in glaucomatous animals to 8.84 ± 5.53 fibers after 48 hours, 33.25 ± 7.16 after 72 hours, and 85.82 ± 13.8 after 96 hours (Fig. 7B). A 3-hour delay before adding 0.001 mg/mL BT to cultures reduced the effect of the drug, such that axonal outgrowth was not significantly increased compared with that in the sham-surgery group in both normal (P16) and glaucomatous samples (P = 0.4). 

Figure 4.

 

Growth of axons in retinal cultures prepared from normal retinal tissue (P16 rats). (A, B) The same retinal explant photographed at 3 and 4 days in vitro (div) exhibited a marked increase in the number of axons and was used for axon counts in untreated control explants. (C, D) The same retinal explant obtained from normal retinal tissue exhibited a marked increase in the number of axons between 2 and 4 div when treated with 0.001 mg/mL BT. Quantification of axons in retinal explants from normal retinal tissue (P16 rats) between 2 and 4 div. Testing of different concentrations of BT (C) revealed that 1 mg/mL BT had adverse effects on axon growth, whereas lower concentrations had growth-promoting effects. (E) The most effective concentration of BT was 0.001 mg/mL. *Statistically different P < 0.05; **statistically different P < 0.005. Scale bar, 100 μm.

Figure 5.

 

Quantification of axons in retinal explants from adult retinal tissue after ONC and lens injury (A) and in glaucomatous retinal tissue (B). A BT concentration of 0.001 mg/mL significantly increased axonal outgrowth in the crushed and glaucomatous tissues compared with the untreated control group. *Statistically different P < 0.05.

Figure 6.

 

Axonal growth of retinal samples after 96 hours in vitro (magnification, ×20) obtained from glaucomatous retinal tissue. (A) Axonal growth in the untreated samples; (B) growth in 0.001 mg/mL BT-treated samples. Scale bar, 50 μm.

Figure 7.

 

Quantification of axons in retinal explants from normal and glaucomatous retinal tissue according to the times of BT application. Immediate placement of the eye in BT-containing buffer (time 0) resulted in a greater degree of axon growth compared with delaying the addition of BT by either 30 minutes or 3 hours in glaucomatous (A) and normal retinal (B) samples. *Statistically different P < 0.05; **statistically different P < 0.005.

The rate of axonal growth was analyzed in the presence of BT and in a non-BT-treated (control) group. In both groups, retinal axons started to grow after 30 hours in vitro. Apart from a slight difference in the initial growth rate (20 μm/hour in the control group; 40 μm/hour in the BT-treated group) both groups displayed similar growth rates, with a maximum growth rate of 80 μm/hour after 78 hours in culture, followed by a steady decrease after 80 hours of observation. 

2D-PAGE and MALDI- MS

BT-induced alterations in the pattern of protein expression were characterized using 2D-PAGE. The gels revealed distinctive differences between the 0.001-mg/mL BT-treated samples and the control (untreated) group (Figs. 8A, 8B). Modified spots were excised and identified by MALDI-MS, GFAP, GRP58, and PAF, which were highly expressed in the control group, were lacking in the BT-treated group. On the contrary, LBP was expressed in the BT-treated group, while absent in the control group. Detection of more proteins associated with growth may be hampered by the resolution of the method and by the low abundance of ganglion cells within the retina, constituting ∼1% of all retinal cells. Plotted proteins of the rat retina are marked in Figure 8 and listed by name in Table 4. Through subsequent immunohistochemistry, the striking presence of the aforementioned proteins, GFAP, GRP58, and PAF, and the α2-ARs was demonstrated and illustrated in cultured retinal samples (Fig. 9). 

Figure 8.

 

2D-PAGE showing striking differences in the pattern of protein expression of normal (P16 rats) retinal samples after 96 hours in vitro between the control group (A) with higher magnifications of specific differentially regulated areas (A1, A2, A3) and the BT-treated (0.001 mg/mL) group (B) including higher magnifications of corresponding differentially regulated areas (B1, B2, B3).The protein spots, numbered and marked (P), are listed in Table 2. For differentially regulated spots, higher magnification is provided, showing GFAP and GRP58 (A1, B1), LBP (A2, B2), and PAF (A3, B3), with the spots highlighted by a black circle and the corresponding area in (A) and (B) by a black frame. GFAP, GRP58, and PAF were distinctive in the control group but absent in the BT-treated group, whereas LBP was expressed only in the BT group.

Table 4.

 

Retinal Proteins Identified by 2-D Gel Electrophoresis and Subsequent MALDI-MS

Abbr	Protein	Molecular Mass (kDa)
P1	Glucose regulated protein 78	78
P2	Myosin	29
P3	Calbindin 2	27
P4	Recoverin	23
P5	Phosphatidylethanolamin binding protein	22
P6	Thioredoxin peroxidase	22
P7	β-Synuclein	17
P8	Cellular retinol binding protein	16
P9	ATP synthase delta subunit	15
P10	ATP synthase D subunit	19
P11	ATP synthase	21
P12	Fertility protein 22	23
P13	1-Cys Peroxiredoxin	25
P14	Prohibitin	30
P15	Platelet-activating factor	29
P16	Proteasome	30
P17	Bovine serum albumin	31
P18	Pyruvat dehydrogenase	36
P19	Transducin β chain A	37
P20	Crystallin	38
P21	Alpha enolase	47
P22	Glucose regulated protein 58	58
P23	Glial fibrillary acidic protein	50
P24	Heat shock protein 60	60
P25	Heat shock protein 70	70
P26	Laminin-binding protein	34

 The abbreviations in the left column refer to Figure 8. Altered proteins are in bold.

Figure 9.

 

Expression of GFAP (red) in P16 rats, shown by immunohistochemical staining of slices of untreated (A) or 0.001 mg/mL BT-treated (B) samples and axonal growth in retinal explants of untreated (C) and 0.001 mg/mL BT-treated (D) samples. Growing axons are marked green with neurofilament, whereas red GFAP staining highlights glial cells. PAF was expressed weakly in untreated (E) or 0.001 mg/mL BT-treated (F) retinal tissue. The expression pattern of STAT 3 (costaining for GRP58) in (G) untreated and (H) treated samples. Each image of a retinal cross section includes a small vertical slice of DAPI staining on the left. The negative control is shown with (I) rhodamine and (J) Cy-2 as the secondary antibodies. Scale bar: (A, B, E–J) 12 μm; (C, D) 75 μm.

Discussion

There were three principal findings in our study: (1) α2-ARs are expressed within the retina of rats and are activated after optic nerve injury; (2) the α2-AR agonist BT promotes the growth of RGC axons from normal, crushed, or glaucomatous rat retinas in vitro; and (3) the proteome profile of cultured retinas is modified in retinal explants treated with BT, with visible changes in expression being observed for such core proteins as GFAP, PAF, and GRP58. The presented data suggest that ligands binding to α2-ARs modulate retinal metabolism to promote axonal growth. 

We found that α2-ARs are expressed within the retina and that this expression is upregulated on injury. The death of the related RGCs, mainly by apoptosis, ⁴³ is a common end point of either traumatic or glaucomatous injuries, although their etiologies are different. Although the mechanism underlying a putative contribution of α2-ARs to this cell death remains to be shown, some of the functions of α2-ARs are known. The receptor is involved primarily in vascular autoregulation ^7,8 by inhibiting adenylate cyclase activity that in turn inhibits the calcium channels and opening the potassium channels. ¹⁰ Furthermore, α2-ARs are involved in the coupling of K⁺ activation that leads to the hyperpolarization of neurons, the reduction of excitability, and the release of neurotransmitters during ischemic injury. ¹¹  

In agreement with previous studies, ^12,13 we found that α2_A-ARs are expressed within the retina, and in particular within the ganglion cell and inner plexiform layers, as shown by using the techniques of Wheeler et al. ⁴⁴ and Zarbin et al. ⁴⁵ Three different α2-AR subtypes are expressed in the rat retina the α2A-, α2B- and α2C-AR. ^14,15,45 α2_A-ARs are mainly located in the GCL, while being present as well in the INL, specifically in the amacrine and horizontal cells. ¹⁵ In the human retina α2_A-AR expression was identified on ganglion cells and on cells in the inner and outer nuclear layers. ⁴⁶ α2_B-ARs are located in all retinal layers, whereas the presence of α2_C-ARs is restricted to photoreceptors. ¹⁵ Most evidently, α2-AR agonists mainly exert their neuroprotective effect via the α2_A-ARs. ¹⁶ The presence of α2_A-ARs in retinal cells provides a site for the neuroprotective activity of α2 agonists in different models of retinal injury. ^{17 –21}  

In addition to the localization, we examined the kinetics of saturation using radioactive clonidine and found that a significant increase in α2-AR expression occurs in the optic-nerve–injured retina. There have been limited and conflicting data concerning the regulation of α2-AR expression by retinal or brain damage. Of note, α2-AR activation after severe hypoxia appears to significantly restrict CNS damage. ⁴⁷ However, it has been shown that α2-AR expression is enhanced after initiation of abortive brain damage in the hilus and striatum ⁴⁸ and that extracellular noradrenalin levels increase after injury, helping to protect the brain from injury. ⁴⁹ The β-adrenergic receptor subtype is upregulated on injury in glial cells, and this change in expression is due to both the hypertrophy and proliferation of glial cells that form glial scar. ⁵⁰  

The specific saturation of α2-ARs by the agonist clonidine may explain why agonists of the receptors such as dexmedetomidine ²² and clonidine ²³ reduce the negative effects of brain ischemia and protect neurons from death. Activation of the α2-AR pathway by another α2-AR agonist, brimonidine, enhances survival of RGC after oxidative stress, chronic ocular hypertension, and acute ischemia of the optic nerve. ¹² Experimental evidence has demonstrated that brimonidine is a potential neuroprotective agent that meets the first three criteria for neuroprotection ⁵¹ —namely, expression of receptors on its target tissues, adequate penetration into the vitreous and retina at pharmacologic levels, and induction of intracellular changes that enhance neuronal resistance to insults or interrupt apoptosis in animal models. However, clinical trials remain outstanding. 

Based on this, we then explored the α2-AR agonist BT, which is used clinically to reduce IOP in glaucoma ²⁴ and to control postoperative IOP, ²⁵ and found that it interacts with the ability of the retina to regrow axons in organ culture. This feature is most likely independent of the ability of BT to reduce the production of aqueous humor and increase uveoscleral outflow. ²⁶ In contrast to topically applied BT, which penetrates the ocular tissue, ²⁷ the in vitro assays used in the present study allowed the evaluation of precise concentrations of BT. Even though we used cultured retinas, we were able to demonstrate the beneficial effects on total amount of axonal outgrowth, whereas the axonal growth rate remained unchanged and confirm the findings of previous in vivo studies in which BT was applied either topically or systemically. ^11,28,32 In addition to former experiments we showed that RGCs regain their function to regrow axons on BT treatment. 

It is assumed that brimonidine is mediated by the direct activation of α2-ARs on RGCs. ⁵² Our data showed a dose and time-dependent effect of BT supplementation on axonal growth. BT is a very potent selective α2-AR agonist activating α2-ARs at 2 nM, ⁵³ appropriate for the 0.0007 mg/mL in our study, which is achieved as well after topical application of 0.2% BT. ²⁷ Activation of the potentially toxic α 1-AR requires 2000 nM BT, ⁵³ explaining the adverse effects of higher concentrations in our study. Our findings concerning dose and time dependency concur with results in other studies. ^18,28 The growth-promoting activity of BT was enhanced significantly by early application, whereas delayed substitution did not show significant beneficial effects on axonal outgrowth. The beneficial therapeutic window of brimonidine is small. Therefore, the brimonidine-mediated neuroprotective mode of action is not likely to be a secondary effect of preventing cells from dying and subsequently disgorging their intracellular contents. ¹⁸  

The mechanisms proposed to be involved in BT-induced neuroprotection in vivo include inhibition of glutamate and aspartate accumulation; upregulation of antiapoptotic molecules such as bcl-2 and bcl-xl; production of neurotrophins such as fibroblast growth factor, ^10,29 causing a prolonged increase of several growth factors as well as receptors in the retina and superior colliculus ¹³ ; and α2-AR-induced N-methyl-d-aspartate receptor mediation, which reduces intracellular calcium levels. Our data obtained with normal, hypertensive, or acutely injured retinal tissue show that in addition to neuroprotective activities on RGC as proven in former studies, ²⁸ BT supports axonal growth in vitro. However, the data do not aid in distinguishing between differences in axonal growth or the amount of surviving RGCs that preserve their function and regrow. The data concur with a previous line of evidence showing that BT directly effects retinal neurons through interactions with the α2-ARs, which can be blocked with the α2-AR antagonists. ¹⁸ Still, the exact mode of action remains unclear. 

To detect treatment-specific protein expression patterns within the retina, we performed 2D-PAGE and MALDI-MS, which is an excellent tool for analyzing treatment-associated or disease-associated complex processes. ^36,54 The proteomic analysis performed with retinal samples exposed to BT showed patterns very similar to those seen in nontreated tissue. However, comparison of some protein spots that appeared differentially regulated showed that retinal proteins such as the cytoskeletal proteins GFAP and GRP58 were regulated by BT. 

GFAP is abundantly expressed in astrocytes and Müller cells and serves as a sensitive indicator for their activation. Under physiological conditions, the level of GFAP is usually minimal in retinal astrocytes in the GCL and nerve fiber layer. It is upregulated in astrocytes and Müller cells by various degenerative retinal conditions. ⁵⁵ This generalized GFAP upregulation points to a gliotic response, which is regarded as inhibitory to axonal growth. ⁵⁶ Downregulation of GFAP may indicate a restricted activation of glial cells and thus a better incidence of axonal regrowth. Additional localization of α2-ARs in Müller cells supports the view that BT may reduce glutamate neurotoxicity by acting throughout glial cells. Müller cells react highly sensitively to retinal insult by, increasing the expression of GFAP and bcl-2, ⁵⁶ and it has been shown that Müller cells lack the tight regulation of toxic extracellular glutamate and its transporters on injury. ⁵⁶ Our findings show that BT reduces GFAP expression significantly, as shown before. ^57,58 It is noteworthy that restricted activation of glial cells is accompanied by upregulation of GRP, which is thought to stabilize the glial cell cytoskeleton. ⁵⁹  

We found reduced expression of PAF in BT-treated samples. PAF is a phospholipid synthesized in a variety of cells throughout the body. PAF has been identified in the central nervous system (CNS) and has diverse physiological and pathologic functions. It has been shown to be a modulator of many CNS processes, ranging from long-term potentiation to neuronal differentiation. Excessive levels of PAF appear to play an important role in neuronal cell injury, resulting from ischemia, inflammation, and meningitis. ⁶⁰ PAF is an important inflammatory lipid mediator affecting neural plasticity, synaptic efficacy through activation of protein kinases and autophosphorylation of calcium/calmodulin-dependent protein kinase II (CaMKII), phosphorylation of synapsin I, and myristoylated, alanine-rich protein kinase C substrate (MARCKS). These results suggest that PAF induces synaptic facilitation through activation of CaMKII, PKC, and ERK in the hippocampal CA1 region. ⁶¹ The beneficial effects of PAF receptor antagonists and substances that lower PAF levels give rise to possible therapeutic strategies for the effects of neuronal injury. 

In addition to glial proteins and PAF, axonal growth-associated proteins such as LBP was highly expressed in the BT-treated samples, but was absent in the sham groups. Laminin, a large extracellular protein, has extraordinarily potent effects on spreading, migration, differentiation, proliferation, and neurite outgrowth. Laminin acts by binding to cell surfaces and several different cell surface proteins have been described to bind laminin with various affinities. ^62,63 The activation of LBPs is therefore essential for axonal outgrowth and upregulation of such proteins may be a hint for axonal growth. 

In conclusion, the data reported are the first to show differences in α2-AR expression between normal and degenerated retinas and α2-AR activation plays a key role in tissue restoration after retinal injury. The α2-AR BT supports neuronal growth in cultured retinal pieces. According to its time- and dose-dependency the growth-promoting effect of BT on neurites is most evidently mediated through α2-AR activation. The growth-promoting effect is associated with downregulation of proteins such as GFAP, GRP, and PAF, known to limit neurite outgrowth, and the upregulation of growth-facilitating proteins, such as laminin-binding protein, and most likely of other proteins that are expressed below the level of resolution of our methods. ⁶⁴  

Footnotes

 Supported by Allergan Laboratories (Irvine, CA) which provided the brimonidine tartrate salt to S.T. and by Deutsche Forschungsgemeinschaft Grants DFG 386-16/1 and -16-2, the IZKF, and the Medical Faculty of the University of Münster.

Footnotes

 Disclosure: V. Prokosch, None; L. Panagis, None; G.F. Volk, None; C. Dermon, None; S. Thanos, None

The authors thank Mechthild Langkamp-Flock for providing skillful technical assistance.