Article Text


Human ganglion cells express the alpha-2 adrenergic receptor: relevance to neuroprotection
  1. F B Kalapesi1,
  2. M T Coroneo2,
  3. M A Hill1
  1. 1Cell Biology Laboratory, Department of Anatomy, School of Medical Sciences, University of New South Wales, Sydney 2052, Australia
  2. 2Department of Ophthalmology, Prince of Wales Hospital, University of New South Wales, Sydney 2052, Australia
  1. Correspondence to: Dr Mark Hill Department of Anatomy, University of New South Wales, Sydney NSW 2052, Australia;


Background/aim: Alpha-2α adrenergic receptor (α2-AR) agonists are thought to be neuroprotective, preventing retinal ganglion cell death independent of pressure reduction. Previous studies have identified α2-ARs in rat retina. The authors aimed to demonstrate the presence and localisation of α2-ARs in human and rat retina and on the rat retinal ganglion cell line, RGC-5.

Methods: Seven postmortem human and three postmortem rat eyes were paraformaldehyde fixed and frozen. RGC-5 cells were also paraformaldehyde fixed. The expression of α2A-ARs was determined by antibody immunofluorescence.

Results: α2A-AR expression was identified in the human retina, on ganglion cells, and cells in the inner and outer nuclear layers (INL, ONL). Differential α2A-AR staining patterns in the INL and ONL suggest a further restriction to as yet unidentified neuronal subclasses. The RGC-5 cell line also expressed α2A-ARs in undifferentiated cells and an increased expression upon fully differentiated cells.

Conclusion: α2-AR agonists in addition to their pressure lowering effects in the eye, may act directly upon retinal neurons, including retinal ganglion cells. The presence of α2-ARs on the RGC-5 cell line allows future investigation of these possible direct effects using in vitro glaucoma model systems.

  • α2-AR, alpha-2 adrenergic receptor
  • CRALBP, cellular retinaldehyde binding protein
  • GCL, ganglion cell layer
  • INL, inner nuclear layer
  • ONL, outer nuclear layer
  • PBS, phosphate buffered saline
  • human ganglion cells
  • alpha-2 receptors
  • α2-AR, alpha-2 adrenergic receptor
  • CRALBP, cellular retinaldehyde binding protein
  • GCL, ganglion cell layer
  • INL, inner nuclear layer
  • ONL, outer nuclear layer
  • PBS, phosphate buffered saline
  • human ganglion cells
  • alpha-2 receptors

Statistics from

In vivo studies have shown that in addition to lowering intraocular pressure, alpha-2α adrenergic agonists, such as brimonidine, decrease retinal ganglion cell death subsequent to increases in intraocular pressure,1–4 retinal ischaemia,5–8 or optic nerve crush.9 Alpha-2 agonists, unlike other proposed neuroprotective agents—for example, β adrenergic antagonists, are thought to work directly via interaction with their receptor. This has been shown in studies where the protective effect, enabled in the presence of an α2 agonist, is reversed or diminished in the presence of an α2 selective antagonist, such as yohimbine6,10,11 or rauwolscine.5,6

Studies evaluating brimonidine or other α2 adrenergic receptor (α2-AR) agonist effectiveness as neuroprotective agents have used the drug applied either topically6–8,12,13 or systemically.1,3,5,7–9,14 These studies do not allow discrimination between whether the drug is acting directly to be neuroprotective, or indirectly by affecting surrounding mediators. These surrounding mediators include the vasculature systemically or locally, inflammatory mediators or other non-retinal ocular tissue such as in the anterior segment. An agent could be directly protective in glaucoma if it interferes with or blocks retinal ganglion cell apoptosis, the terminal outcome in glaucoma, and cause of irreversible loss of corresponding visual fields. A target for a neuroprotective agent in glaucoma would be the retinal ganglion cell, being both the output neuron for the retina and the cell that dies via apoptosis in glaucoma.15–19 Such a drug may thus have a direct effect on the cells involved in glaucoma pathogenesis.

There are at least three distinct α2-AR subtypes.20 The α2A-AR is the predominant subtype in the central nervous system and the receptor subtype for which brimonidine has highest affinity.21 Despite the knowledge that α2 adrenergic agents exert their neuroprotective effect by receptor binding, little is known about the α2 adrenergic binding sites within the retina. Table 1 reviews the current published data on localisation of α2-ARs to the eye, including knowledge of localisation to the retina. Previous evidence of the intraretinal localisation of α2-ARs comes from radioligand binding or autoradiography studies, which have localised α2-AR to the mammalian (in particular to the human) retina without intraretinal receptor subtype localisation. Recently, Wheeler et al3 demonstrated that α2-ARs were present on rat ganglion cells as well as cells in the inner nuclear layer. Owing to the known difference in tissue expression of α2-ARs among different animal species,22–24 and the fact that the intraretinal α2-AR expression has not yet been shown in humans, we aimed to define the intraretinal localisation of α2-AR in the human retina. We also aimed to confirm previous rat intraretinal α2-AR localisation.3

Table 1

 Previous studies localising α2 adrenergic receptors to the eye

In addition to human and rat intraretinal localisation, we looked for the presence of specific ganglion cell localisation using a novel retinal ganglion cell line, RGC-5. The RGC-5 cell line is a virally transformed rat retinal ganglion cell line and is currently the only retinal ganglion cell line.25 It expresses Thy1-1 and Brn-3c, ganglion cell markers and also various neurotrophin receptors. The RGC-5 line morphologically differentiates with serum deprivation and succinyl concavalin A treatment, and also undergoes excitotoxic apoptosis with glutamate.25


Preparation of retinas for immunohistochemistry

Seven human postmortem donor eyes (ethics approval HREC 03225) from five different donors were obtained from the Lions Eye bank (Sydney, Australia). Patient age ranged from 26 to 84 years with no significant known ocular conditions besides cataract extractions. Human eyes were pre-fixed in 2% paraformaldehyde overnight. Retinal trephines were equilibrated in 30% sucrose, blocked, mounted, frozen, and sectioned onto sialinated coated slides.

A total of three freshly enucleated adult rat eyes (ethics approval ACEC 1997/101) were used for this study and were processed as per human retinas.

Cell culture and preparation of RGC-5 cells

The transformed rat retinal ganglion cell line, RGC-5, (gift from N Agarwal) was grown as previously described40 on poly-l-lysine coated glass coverslips. Adherent cells were differentiated as previously described for 3, 5, and 7 days.25 Coverslips were fixed in 4% paraformaldehyde.


A chicken polyclonal antibody generated to recognise the α2A-AR41 was generously supplied by J Regan. Primary antibody and control normal chicken IgY (R&D Systems, AB-101-C) were used at a dilution of 1:150. Secondary antibody goat anti-chicken IgG conjugated to alexa 546 (molecular probes, A-11040) was used at a dilution of 1:200. Cell nuclei were identified counterstaining with Hoechst 33258 (molecular probes, H1398).


Immunohistochemistry was performed using the indirect antibody method. Briefly, specimens were blocked with 5% normal goat serum, incubated overnight with primary antibody diluted into 0.4% saponin for permeabilisation. Following extensive washes with phosphate buffered saline (PBS), the immunoreactivities were revealed by secondary antibody goat anti-chicken conjugated to fluorochrome alexa 546.

Controls included omission of the primary antibody, normal chicken IgY substituted in place of the primary antibody, and staining of human cornea, a known positive control for this α2AR antibody.29 Preliminary co-localisation studies were performed by dual incubation with the α2A-AR antibody along with vimentin (Dako, M0725), cellular retinaldehyde binding protein (CRALBP) (gift from M Maddigan) and appropriate FITC secondary antibodies (Zymed 04–6111 and molecular probes A12373) (data not shown).


Fluorescence was visualised and photographed using a confocal laser microscope (Leica TCS SP) equipped with Leica confocal software (version 2.5) attached to a two photon imaging system (Leica TCS MP). Images were taken with 20× (HC PL fluotar 20× 0.5 dry, Leica) and 63× water objectives (HC PL apo 63× 1.2Wcorr, Leica). Images collected with the 543 nm excitation laser were pseudocoloured red while those taken with the 790 nm excitation laser were pseudocoloured blue.

A total of 82 images of human retina stained with α2A-AR antibody, 165 human retina negative controls (115 secondary only and 50 IgY negative controls), 73 human cornea, 25 of rat retina stained with α2A-AR antibody, 34 rat negative controls (29 secondary only and four IgY negative controls). Total number of images stained with α2A-AR antibody for the RGC-5 cells were 27 undifferentiated, 9 day three, 6 day five, and 28 day seven differentiated RGC-5 cells. Thirty six negative control images were taken (27 secondary only and 9 IgY negative controls). Intensity of the human tissue staining was graded with − representing no staining; + representing low or weak staining; ++ representing moderate or intermediate staining; and +++ representing high or strong levels of staining.


Strong α2A-AR staining was shown in human cornea, a known positive control (fig 2A).29 In the rat retina, we confirmed the presence of α2-ARs on the somas of the ganglion cells and many of the cells in the inner nuclear layer. The immunopositive labelled cells all displayed uniform staining (fig 1A). In the human retina, α2-ARs were localised to the soma of ganglion cells, a subpopulation in the inner nuclear layer (INL) and a smaller subpopulation within the outer nuclear layer (ONL) (figs 2E and G). The intensity of receptor staining varied between layers in the same eye and between similar layers in different eyes, especially in both the INL and ONL as demonstrated by table 2 and figure 2. Within layers, there was variability in the staining levels seen between different eyes as demonstrated by figures 2E and G, images taken of different eyes. A sample quantitative count was carried out on two eyes using five images, with Hoechst counts defining total cell numbers. The percentage of positive cells staining in the ganglion cell layer was >80%, INL 40% and ONL only 20% positive. Of the positive staining cells in the ONL nearly all were weak + compared to the INL, which was mainly +++ (43%) and ++ (49%).

Table 2

 Summary of averaged intensity of immunostaining for the α2-AR (alpha-2 adrenergic receptor) in the human cornea and retina

Figure 1

 α2A-AR expression in the rat retina. Confocal images of rat retina labelled with α2A-AR antibody throughout retinal layers. (A) Shows positive immunostaining for the α2A-AR on retinal ganglion cells, cells in the inner nuclear layer (INL) and less so, but also on the membrane of outer nuclear layer (ONL) cells. (C) A retina control with secondary antibody alone showing no non-specific staining. (B) and (D) are equivalent to (A) and (C) retinal sections, respectively, showing Hoechst nuclear counterstaining. All images are taken with the ×63 objective, (C) shows 25 μm scale bar. GCL, ganglion cell layer.

Figure 2

 α2A-AR expression in the human cornea and retina. Confocal images of human eye labelled with α2A-AR antibody in cornea and retinal layers. (A), (E), and (G) Immunostained with α2A-AR antibody and (C) with chicken IgY alone. (B), (F), (H), and (D) Corresponding Hoechst nuclear counterstaining. (A) The positive control, is human corneal epithelium showing specific strong staining for the 2AR. (C) A retina control with chicken IgY alone showing no non-specific staining. (E) and (G) Staining of two different human retinas with positive retinal ganglion cells (arrows) and variable staining in inner (INL) and outer nuclear layers (ONL). Between different retinas, there was some variability in overall intensity but a similar pattern of staining. All images are taken with the ×63 objective. (G) Shows 25 μm scale bar. GCL, ganglion cell layer.

In the RGC-5 cell line, the α2A-AR staining was seen on both undifferentiated (fig 3A) and differentiated cells (fig 3B, C, and D); however, staining was more intense on the day 7 differentiated RGC-5 cells (see fig 3D). Differentiated cells (fig 3D) cease to proliferate and display neurites unlike undifferentiated cells (fig 3A).

Figure 3

 α2A-AR expression in the rat retinal ganglion cell line, RGC-5. Confocal images of rat retinal ganglion cell line RGC-5 labelled with α2A-AR antibody. (A) Undifferentiated RGC-5 cells have low, though positive, level of α2A-AR expression. (B-D) 3, 5, and 7 days, respectively, of RGC-5 differentiation, showing positive α2A-AR expression, with maximum expression at 7 days of differentiation. (E) (inset) RGC-5 cells with chicken IgY alone showing no non-specific staining. Identical staining and confocal imaging conditions used for all images. All images ×40 objective. (C) Shows 25 μm scale bar.

Preliminary tissue co-localisation studies did not reveal the positive alpha immunostaining to coincide with the staining for CRALBP (a Muller cell marker)42–44 or vimentin,42 glial cell markers (results not shown).

The antibody was checked for any non-specific binding by incubation of retinas and RGC-5 cells with secondary antibody alone, to test for any non-specific binding of the secondary and with host normal chicken IgY to test for any non-specific binding of our chicken primary antibody. Both negative controls were negative (see figs 1C, 2C, and 3E), indicating no non-specific binding of our antibody for the α2A-AR.


This is the first study immunochemically demonstrating the presence of α2-ARs on human retinal ganglion cells and other retinal neurons in both the inner and outer human retina. This study also demonstrated the presence of the α2-ARs on the RGC-5 cell line.

Our demonstration of α2-AR at the protein level in the retina is superior to both RNA and homogenate binding studies, offering increased anatomical resolution and sensitivity.45 Previous studies have identified receptor localisation on rat retinal ganglion cells and other rat retinal neurons.3 This does not necessarily mean a similar expression in human retina. Specific receptor antibodies allow confirmation of adrenergic receptor subtype distribution as well as the possibility to test various pharmacological agents that interact specifically with these receptors.

Our current study confirmed previous immunochemical studies3 showing the presence of α2-ARs (specifically α2A-ARs) on rat retinal ganglion and inner nuclear layer cells (fig 1), with even strong staining on all immunopositive cells.

In the human retina we found the presence of α2A-ARs on human retinal ganglion cells and retinal cells in the inner and outer nuclear layers. In general, all eyes show a similar overall staining pattern, with some intensity variability between eyes (fig 2E and G), possibly related to in vivo pharmacological interventions, different ages of eyes, and different time from enucleation to fixation. Differential staining occurred within the inner nuclear layer and outer nuclear layer and could reflect specific neuronal subtypes as these nuclear layers are composed of varying neuronal cell types. Preliminary studies using known glial cell markers,42–44 vimentin, and CRALBP did not co-localise with α2A staining, suggesting that immunopositive cells in the inner and outer nuclear layers are neuronal and not glial. Previous studies in other species suggested that the subpopulation of cells staining positively in the inner nuclear layer are amacrine cells.3 Studies involving other retinal subclass markers could help to further discern the cell types staining positive for the α2A receptor. The rat retina displays a less apparent differential staining than the human retina without a subpopulation of cells staining positive in the outer nuclear layer, reflecting interspecies variability and confirming the need for this current immunochemical study.

We found markedly enhanced α2A-AR expression on day 7 differentiated RGC-5 cells compared to undifferentiated or early differentiated RGC-5 cells (fig 3). The changing receptor expression in the RGC-5 cell line with differentiation could reflect changes to in vivo expression levels with neuronal development. The demonstration of α2A-ARs on the RGC-5 cells substantiates positive staining found in the ganglion cell layers of human and rat retina, suggesting that positive immunoreactivity in this layer are on the ganglion cells rather than displaced amacrine cells. The presence of α2A-ARs on the RGC-5 cell line allows the possibility of studying the direct effect of α2-AR agonists on retinal ganglion cells, the target cells of glaucoma. This eliminates the need for studying mixed primary cultures and then using ganglion cell markers in order to identify between the various cell types present. Thus, the direct effect of α2-AR agonists can be studied on retinal ganglion cells in culture, while holding relevance to potential human interactions.

Alpha-2 receptor activation has been implicated in enhanced neuronal survival in glaucomatous in vivo models1–9 and hence alpha-2 agonists form one of the most studied neuroprotective agents. In contrast with some other proposed neuroprotective agents—for instance, betaxolol, alpha-2 agonists could exert their neuroprotective effect on retinal neurons via direct interaction with their receptor.6,10,11 Alpha-2-AR stimulation may protect retinal ganglion cells by inhibiting pro-apoptotic mitochondrial signalling, but yet the precise mechanism of neuroprotection is unknown.46 Our demonstration of the α2A-AR, the most common subtype of α2-ARs found in the central nervous system and the subtype for which brimonidine has highest affinity,21 on the human retinal ganglion cell as well as on other retinal neurons, importantly demonstrates a direct target for neuroprotection on and in the vicinity of the ganglion cells, which are dying by apoptosis in glaucoma.

While a direct relation, in neurons, between α2A-ARs and apoptosis has not yet been demonstrated, preliminary data from meetings (Wheeler et al, Invest Ophthalmol Vis Sci 2001;42:411) suggest that α2A-AR agonists decrease apoptosis through an unknown mechanism that increases Bcl-2 expression. In contrast, the developing embryo expresses α2A-ARs in regions of embryo apoptosis,47 suggesting exactly the opposite relation with apoptosis. Further studies using α2A-ARs agonists and retinal ganglion cells, potentially using the RGC-5 cell line, are required in order to ascertain the relevance of α2-ARs in apoptosis and neuroprotection.


We thank Professor J Regan for alpha-2A antibody (Department of Pharmacology, University of Arizona, AZ, USA); Associate Professor N Agarwal for the RGC-5 cell line (Department of Cell Biology and Genetics, University of North Texas Health Science Center, Fort Worth, TX, USA); Dr M Madigan (Save Sight Institute, Sydney, Australia) for technical advice regarding these experiments and for facilitating sufficient ocular tissue for these experiments; and Mr Paul Halasz for his assistance with the operation of the two photon imaging system.

FBK is a recipient of an NHMRC scholarship.


View Abstract


  • Competing interests: The authors have no competing or financial interests.

  • Ethics approval: University of New South Wales, Human Research Ethics Committee Approval 03225; University of New South Wales, Animal Care and Ethics Committee Approval 1997/101.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.