TNF-α and TNF-α Receptor-1 in the Retina of Normal and Glaucomatous Eyes

Abstract

purpose. To determine the expression and localization of tumor necrosis factor (TNF)-α and TNF-α receptor-1 in the retina of normal and glaucomatous eyes.

methods. Using immunohistochemistry and in situ hybridization, retinal expression and localization of TNF-α and TNF-α receptor-1 were studied in retina sections from 20 eyes of donors with glaucoma, and 20 eyes of age-matched normal donors.

results. According to immunohistochemistry, the intensity of the immunostaining and the number of labeled cells for TNF-α or its receptor were greater in retina sections of glaucomatous eyes than in control eyes of age-matched normal donors. In situ hybridization showed that mRNA signals for TNF-α or TNF-α receptor-1 were similarly more intense in glaucomatous eyes than in age-matched control eyes. Both protein and mRNA of TNF-α or TNF-α receptor-1 were predominantly localized to the inner retinal layers. Double-immunofluorescence labeling demonstrated that retinal immunostaining for TNF-α was predominantly positive in the glial cells, whereas immunostaining for TNF-α receptor-1 was mainly positive in the retinal ganglion cells.

conclusions. Upregulation of TNF-α and its receptor-1 in glaucomatous retina suggest that TNF-α–mediated cell death is involved in the neurodegeneration process of glaucoma.

Tumor necrosis factor (TNF)-α is a potent immunomediator and proinflammatory cytokine that is rapidly upregulated in the brain after injury. ^{1 2} The dramatic increase in TNF-α production after ischemic and excitotoxic brain injury suggests an important role for this cytokine in modifying the neurodegenerative process, and therefore it has been implicated in the pathogenesis of several diseases of the central nervous system, such as multiple sclerosis and autoimmune encephalomyelitis. ^{3 4} Its excessive synthesis after trauma has been correlated with poor outcome, ⁵ and its inhibition is accompanied by reduced brain damage. ⁶ In addition, TNF-α has been thought to account for axonal degeneration and glial changes observed in the optic nerves of patients with AIDS.⁷ It is an inducer of apoptotic cell death through TNF-α receptor-1 (p55) occupancy in a caspase-mediated pathway. ⁸ In addition, TNF-α is a potent activator of neurotoxic substances such as nitric oxide and excitotoxins. ^{9 10} Furthermore, a picogram concentration of TNF-α that is known to be noncytotoxic induces neuronal cell death through the silencing of survival signals. ¹¹  

In addition to our in vitro studies demonstrating activation of retinal caspase-8 in response to glaucomatous stressors,^12,13 our preliminary in vivo studies using a rat model of high-pressure glaucoma revealed caspase-8 activation during retinal cell death cascade in rat eyes following elevation of intraocular pressure (unpublished observation). Although both caspase-dependent and -independent components of mitochondrial cell death pathway are involved in this cascade, activation of caspase-8 that is a proximal effector protein is known to be a hallmark of TNF receptor family cell death pathway.⁸ Therefore, observation of retinal caspase-8 activation, in vitro and in vivo, created the first idea that TNF-α–mediated cell death may be involved in glaucomatous neurodegeneration. Subsequently, histopathologic studies in human donor eyes revealed there is increased immunostaining for TNF-α and TNF-α receptor-1 in the glaucomatous optic nerve head compared to age-matched control eyes.^14,15 These observations thus provided additional evidence that TNF-α may have a role in tissue remodeling and/or neurodegeneration in glaucoma. Recently, in vitro studies using primary co-cultures of retinal ganglion cells and glial cells provided direct evidence that elevated pressure or ischemia, which are two prominent stress factors identified in the eyes of patients with glaucoma, can initiate the apoptotic cell death cascade in retinal ganglion cells, largely through TNF-α secreted by reactivated glial cells in response to these stressors. Furthermore, retinal ganglion cell death in these cultures can be attenuated approximately 66% by inhibition of the bioactivity of TNF-α.¹⁶  

Because retinal expression and localization of TNF-α and TNF-α receptor-1 have not been described in either normal or glaucomatous eyes, by using immunohistochemistry and in situ hybridization, we studied their protein and gene expression and localization in the retina of human donor eyes with glaucoma in comparison with age-matched normal donor eyes. Our observations revealed increased protein and gene expression of TNF-α and TNF-α receptor-1 in the retina of glaucomatous eyes, which suggest that TNF-α–mediated cell death is involved in glaucomatous neurodegeneration. Whereas localization of TNF-α was prominent in glial cells, TNF-α receptor-1 was mainly localized to retinal ganglion cells. This observation provides evidence that retinal ganglion cells are sensitive targets for the cytotoxic effects of TNF-α that is produced by glial cells in glaucomatous retina. The predominant localization of TNF-α receptor-1 to retinal ganglion cells may partly explain their increased selective sensitivity to primary and/or secondary degeneration in glaucoma. 

Materials and Methods

Eyes

Twenty human eyes from 14 donors with a diagnosis of glaucoma (ages, 56–94 years), and 20 eyes from 10 age-matched donors with normal eyes (ages, 55–96 years) were obtained from the Glaucoma Research Foundation (San Francisco, CA), the Mid-America Eye Bank (St. Louis, MO), and Martin B. Wax, MD (Washington University, St. Louis, MO). Clinical findings in the patients were well documented and included optic disc assessments and visual field tests (Table 1) . Although 11 of the glaucomatous eyes were recorded to have primary open-angle glaucoma and 9 to have normal-pressure glaucoma, the criteria for differential diagnosis of primary open-angle glaucoma and normal-pressure glaucoma may not have been uniform among donors, because data were retrospectively collected from different sources. In addition, primary open-angle glaucoma and normal-pressure glaucoma may not be entirely different diseases, and they share a common feature of retinal ganglion cell death. Therefore, we did not evaluate our results for primary open-angle glaucoma and normal-pressure glaucoma separately, rather we studied them in one group representing open-angle glaucoma. Normal donors had no history of eye disease. There was no diabetes, collagen vascular disease, infection, or sepsis in any of the donors. The cause of death in all the donors was acute myocardial infarction or cardiopulmonary failure. 

The eyes were enucleated within 2 to 4 hours after death and were processed and fixed within 6 to 12 hours in either 10% buffered formaldehyde or 4% paraformaldehyde. The posterior poles were dissected from the surrounding tissues, washed extensively in 0.2% glycine in phosphate-buffered saline at pH 7.4, embedded in paraffin, and oriented sagittally to obtain 6-μm sections. Slides for in situ hybridization were handled using sterile techniques to avoid RNase contamination of the sections. To control variations in the immunolabeling or detection of mRNA signals, serial sections from the glaucomatous eyes and normal donor eyes were simultaneously subjected to immunohistochemistry or in situ hybridization. All the histologic slides were masked for the identity and diagnosis of the donors. In addition, the protein examined (TNF-α or its receptor) was not indicated on the slides; rather, all the slides were numbered by a technician who was not familiar with the histopathology of retina. The intensity of immunostaining or mRNA signals in different layers and different regions of the retina was then qualitatively evaluated (negative, faint, or increased) by an experienced observer (GT) in a masked fashion. At least five histologic sections from each donor eye were examined for each protein and mRNA. The relationship between the intensity of immunostaining and the level of glaucomatous damage was not evaluated because of technical difficulties, such as inadequacy of optic nerve tissue for axon count, insufficient information about retinal orientation in blocks to correlate with visual fields, and the regional variability of retinal ganglion cell counts in histologic slides. 

Immunohistochemistry

For immunoperoxidase staining, sections from normal and glaucomatous eyes were deparaffinized, rehydrated, and pretreated with 0.3% hydrogen peroxide in phosphate-buffered saline solution to decrease endogenous peroxidase activity. Monoclonal antibodies against TNF-α or TNF-α receptor-1 (2 μg/ml; R&D Systems, Minneapolis, MN) were localized by immunoperoxidase, with reagents purchased from Vector Laboratories (Burlingame, CA). The biotinylated secondary antibody was incubated with the sections for 30 minutes, washed with phosphate-buffered saline solution containing 0.1% bovine serum albumin, and reacted with streptavidin-horseradish peroxidase conjugated for 30 minutes. After several washes, color was developed by incubation with 3,3′-diaminobenzidine tetrahydrochloride (Sigma, St. Louis, MO) as a cosubstrate, for 5 to 7 minutes. Sections were counterstained with hematoxylin and mounted (Permount; Fischer, Pittsburgh, PA). For a negative control, nonimmune rabbit and mouse sera (Sigma) were used to replace the primary antibodies. Slides were examined in a microscope (Nikon, Tokyo, Japan), and images were recorded by digital photomicrography (Optronics, Goleta, CA). 

To study localization of TNF-α or TNF-α receptor-1 in the retina, we performed a double-immunofluorescence procedure using antibodies against specific cell markers. We used monoclonal antibody against glial fibrillary acidic protein as a marker of glial cells. To identify retinal ganglion cells, we used a monoclonal antibody to Brn-3a (Chemicon International, Inc., Temecula, CA) that is a member of the POU-domain genes and is known to be expressed by most ganglion cells across a variety of mammalian species. ^{17 18}  

For double-immunofluorescence labeling, sections were incubated with a mixture of mouse and rabbit antibodies at 1:100 dilution for 30 minutes. The sections were then incubated with a mixture of rhodamine-red and Oregon-green–labeled secondary antibodies (Molecular Probes, Eugene, OR) for another 30 minutes. Negative controls were performed by replacing the primary antibody with nonimmune serum or by incubating sections with each primary antibody followed by the inappropriate secondary antibody, to determine that each secondary antibody was specific to the species it was raised against. Slides were examined in a fluorescence microscope (Nikon) and images were recorded by digital photomicrography (Optronics). 

In Situ Hybridization

cDNAs encoding the full sequence of human TNF-α (American Type Culture Collection, Manassas, VA) or human TNF-α receptor-1 (Genentech, Inc., South San Francisco, CA) were subcloned into a plasmid transcription vector (pBluescript; Stratagene, La Jolla, CA). Plasmid cDNAs were purified, and the confirmation of recombinant plasmids was made by restriction enzyme analysis and DNA sequencing. Digoxigenin (DIG)-labeled single-stranded sense and antisense RNA probes were generated by in vitro transcription of linearized recombinant plasmids containing TNF-α and TNF-α receptor-1 in the presence of DIG-uridine triphosphate (UTP), with a kit (Roche Molecular Biochemicals; Indianapolis, IN). Antisense RNA probes for TNF-α and its receptor were transcribed by T3 RNA polymerase from recombinant plasmid linearized with EcoRI and BamHI, respectively, and sense RNA probes were transcribed by T7 RNA polymerase from recombinant plasmids linearized with HindIII. DIG-labeled probes were then used for in situ hybridization. Probe specificity to TNF-α and TNF-α receptor-1 mRNA was assessed by Northern hybridization.  

For in situ hybridization, tissue sections were deparaffinized and rehydrated in a graded series of ethanol solutions. To preserve the mRNA, the sections were fixed with 4% paraformaldehyde for 20 minutes. After washing in TBS (50 mM Tris-HCl [pH 7.5] and 150 mM NaCl), the sections were treated with proteinase K solution for 20 minutes and digestion was stopped by incubation with TBS solution. After a washing with TBS, the sections were treated with 200 mM HCl solution for 10 minutes to denature the proteins. The sections were then rinsed with TBS and incubated in 0.5% acetic anhydride solution containing 100 mM Tris (pH. 8.0) for 10 minutes to reduce nonspecific background. After dehydration in a graded series of ethanol solutions, the sections were incubated at 55°C for 30 minutes before hybridization. Hybridization was performed in a buffer containing 2× SCC, 10% dextran sulfate, 0.01% sheared salmon sperm DNA, 0.02% SDS, and 50% formamide. The hybridization mixture (50 μl per section), containing 10 ng of labeled RNA probe was applied to the sections. To increase the signal from RNA/RNA hybrids the slides were placed on a hot plate at 95°C for 4 minutes and then incubated in a humid chamber for 4 to 6 hours at 55°C to 75°C. After hybridization, the sections were incubated in 2× SSC overnight and washed for 3 × 20 minutes at 55°C in buffer containing 50% formamide, 1× SSC, followed by two 15-minutes washes with 1× SSC at room temperature and rinses with TBS. After blocking in 10% fetal calf serum for 15 minutes, the sections were incubated with alkaline-phosphatase–conjugated anti-DIG antibody (150 mU/ml; Roche Molecular Biochemicals) for 60 minutes. After a rinse with TBS, the sections were incubated with nitroblue tetrazolium chloride/5-bromo-4-chloro-3-indolyl-phosphate (NBT-BCIP) color reagent in a refrigerator. Controls were performed by eliminating the RNA probes from the hybridization buffer or replacing the antisense probe with sense probe. The slides were examined in a microscope and images were recorded as for immunohistochemistry. 

Results

Immunohistochemistry

Immunohistochemical staining using specific antibodies to TNF-α or TNF-α receptor-1 to detect protein expression in normal donor eyes demonstrated constitutive expression of both TNF-α and its receptor in human retina. Regarding immunostaining for TNF-α, faint immunostaining was barely detectable in the control retina, which was confined to a few glial cells and their processes and blood vessels (Fig. 1A ). However, the intensity of the immunostaining and the number of stained cells were noticeably greater in retina sections from glaucomatous eyes (Fig. 1C) . Based on the morphologic assessment, the immunostaining for TNF-α in glaucomatous donor eyes was mostly associated with glial cells located in the nerve fiber and retinal ganglion cell layers (Fig. 1E) . For example, at the level of light microscopy, Müller cells are characterized by their radial orientation, and astrocytes are characterized by their darker, smaller, and irregular nucleus relative to that of ganglion cells, and by their close localization to the blood vessels of the inner retina. ¹⁹ Some immunostaining was also observed in other retinal layers, which was associated with either the cell bodies of the Müller cells located in the inner nuclear layer ^{20 21} and the processes of glial cells all through the retina or with blood vessels (Fig. 1) . In addition to differences between retinal layers, qualitative evaluation of the immunostaining in different retinal regions revealed that the immunostaining for TNF-α in retina sections from all glaucomatous eyes was more intense in retinal areas close to the optic nerve head and adjacent to the parapapillary chorioretinal atrophy compared with the more peripheral retina (Fig. 2) . 

Examination of retina sections from normal donor eyes using immunohistochemistry revealed barely detectable immunostaining for TNF-α receptor-1 that was limited to a few glial cells and their processes. In addition, there was positive immunostaining associated with the blood vessels, which was more prominent than the immunostaining for TNF-α (Fig. 1B) . In glaucomatous eyes, the intensity of the immunostaining and the number of stained cells for TNF-α receptor-1 were notably greater than that in normal eyes (Fig. 1D) . Positive immunostaining for TNF-α receptor-1 in retinal sections from donor glaucomatous eyes was detectable in the cytoplasm as well as on the cell surface. In addition to faint immunostaining observed in all retinal layers, which was probably associated with glial cell processes or blood vessels, immunostaining for TNF-α receptor-1 was predominant in most large cell bodies in the retinal ganglion cell layer (Fig. 1F) . Control sections in which the primary antibodies were omitted or replaced with nonimmune sera were all negative for specific immunostaining of either TNF-α or TNF-α receptor-1. 

In Situ Hybridization

In situ hybridization with specific probes used to detect the mRNAs clearly demonstrated induction of TNF-α and TNF-α receptor-1 genes in the retina of glaucomatous eyes compared with the retina of age-matched normal eyes (Fig. 3) . In addition, in situ hybridization demonstrated that the localization of mRNAs of TNF-α and TNF-α receptor-1 was similar to the localization of their proteins, as detected by immunohistochemistry. Although faint mRNA signals for TNF-α or TNF-α receptor-1 were detectable in all retinal layers in association with glial cells or blood vessels, increased gene expression for TNF-α or TNF-α receptor-1 was predominantly localized in the inner retinal layers (Fig. 3) . As shown in Figures 3C and 3D , when a full-thickness retina was viewed at low power, the most intensely stained layer for TNF-α or TNF-α receptor-1 mRNAs in the glaucomatous retina was the ganglion cell layer. Based on morphologic assessment, astrocytes or retinal ganglion cells located in this layer were prominently stained for mRNAs of TNF-α or TNF-α receptor-1, respectively. In addition, mRNA signals for TNF-α were prominently increased in some of the cells located in the inner nuclear layer, which probably correspond to Müller cells. Control slides for in situ hybridization using sense RNA probes for TNF-α or TNF-α receptor-1 were all negative for specific staining (Figs. 3E 3F) . 

Double-Immunofluorescence Labeling

Double-immunofluorescence labeling to examine localization of TNF-α and TNF-α receptor-1 to retinal cell types demonstrated that immunostaining for TNF-α was associated with retinal glial cells, but predominant immunostaining for TNF-α receptor-1 was in the retinal ganglion cells. Although immunostaining for TNF-α was mostly localized to retinal glial cells, which were simultaneously stained with glial fibrillary acidic protein, immunostaining for TNF-α receptor-1 was colocalized with Brn-3a, which is a marker of retinal ganglion cells (Fig. 4) . 

Discussion

We observed increased protein and gene expression of TNF-α and TNF-α receptor-1 in the retina of glaucomatous eyes. Whereas immunostaining for TNF-α was predominantly localized to the processes of retinal glial cells as well as their cell bodies, immunostaining for TNF-α receptor-1 was identified mainly in the retinal ganglion cells. Positive immunostaining for TNF-α receptor-1 detected in the cytoplasm as well as on the cell surface is in accordance with the observation that after receptor binding, this receptor-ligand complex is internalized, which is critical for cell death signaling. ^{22 23 24}  

TNF-α is mostly produced by reactivated astrocytes ^{25 26} and microglia ²⁷ as well as macrophages. ²⁸ Reactivated glial cells at sites of central nervous system damage arising from a wide variety of disorders are implicated in tissue injury through release of TNF-α. ^{29 30 31} Increased TNF-α production by reactivated glial cells in several retinal diseases has similarly been implicated in the ensuing death of neuronal cells. ^{32 33 34} The localization of TNF-α, which was detected most prominently in the inner retinal layers, is in accordance with the distribution pattern of retinal glial cells, because astrocytes are mostly located in the retinal ganglion cell and nerve fiber layers and cell bodies of the Müller cells are located in the inner nuclear layer. ^{19 20 21} In addition, double-immunofluorescence labeling provided further verification that TNF-α is mostly produced by retinal glial cells. The upregulation of TNF-α in retinal glial cells in glaucomatous eyes agrees with previous observations that retinal glial cells undergo a reactivation process in glaucoma ³⁵ similar to that identified in the glaucomatous optic nerve head. ³⁶  

Elevated intraocular pressure and ischemia are common stress factors identified in glaucomatous eyes, which are thought to facilitate retinal ganglion cell death. ^{37 38} Previous evidence suggests that both elevated pressure and ischemia can induce expression of TNF-α in different cells, including retinal cells. ^{39 40} In addition, recent in vitro studies using primary cocultures of retinal ganglion cells and glial cells provided direct evidence that production of TNF-α is upregulated in retinal glial cells after exposure to elevated hydrostatic pressure or simulated ischemia. ¹⁶ Therefore, upregulation of TNF-α in retinal glial cells in glaucomatous eyes is not surprising. These findings thus support previous in vitro evidence that retinal glial cells are the source of increased production of TNF-α in glaucoma. Although glial reactivation accompanying neuronal damage in glaucoma may initially be a cellular attempt to limit the extent of injury and to promote tissue repair process, increased production of TNF-α, a neurotoxic substance, by reactivated glial cells suggests that these cells may have neurotoxic influences as well. 

The cellular distribution patterns of TNF-α and TNF-α receptor-1 in the retina are similar to previous observations in the glaucomatous optic nerve head using immunohistochemistry. ^{14 15} In both the optic nerve head and retina, TNF-α was mostly expressed by the glial cells; however, the expression of TNF-α receptor-1 was prominent in neuronal tissue and was increased in the glaucomatous eyes. The presence of TNF-α receptor-1 in neuronal tissue, specifically in the retinal ganglion cells and their axons indicates that these cells are sensitive to the effects of TNF-α produced by glial cells in glaucoma. This is supported by previous observations of Madigan et al. who demonstrated TNF-α can produce axonal degeneration in rabbit optic nerves following intravitreal injection.⁴¹  

Previous observations indicate a selective vulnerability of retinal ganglion cells to damage in glaucoma. ^{42 43 44 45} Retinal ganglion cell death in glaucoma is commonly thought to be associated with the injury of their axons at the level of the optic nerve head. For example, the blockade of axoplasmic flow at the lamina cribrosa in the optic nerve head and the resultant blockade of neurotrophin transport to the retinal ganglion cells has been suggested to be a mechanism that contributes to retinal ganglion cell death in glaucoma. ^{37 46 47 48} Nitric oxide damage has also been implicated in the glaucomatous injury of retinal ganglion cell axons. ⁴⁹ Although axonal damage at the level of the optic nerve head may explain selective loss of ganglion cell bodies by retrograde degeneration, there are regional ^{50 51 52} and cellular ^{43 53} differences in the susceptibility of individual retinal ganglion cells to glaucomatous damage that are not well understood. Evidence suggests that intraretinal events including chronic retinal ischemia, ^{38 54 55} excitotoxicity, ⁵⁶ and an autoimmune mechanism, ^{13 57} may facilitate primary and/or secondary degeneration of retinal ganglion cells in glaucoma as well. TNF-α–mediated cell death appears to be an important component of these noxious events triggered by elevated intraocular pressure and/or ischemia in glaucomatous eyes.¹⁶  

Based on the findings presented herein, it is tempting to propose that relatively selective expression of TNF-α receptor-1 in retinal ganglion cells may partly explain the increased vulnerability of retinal ganglion cells to apoptosis during the process of glaucomatous optic nerve degeneration, in which TNF-α is an important mediator of cell death. What remains unclear, however, is that in different cell types, or even within the same cell type, responses to TNF-α may result in either cell death or survival and proliferation. In most cells, TNF-α receptor-1 occupancy by TNF-α induces apoptosis by activating the apoptotic caspase cascade. However, under certain conditions it may provide protection by induction of survival genes, including nuclear factor (NF)-κB and heat shock proteins. ^{58 59 60 61 62 63} It is apparent that the balance between positive and negative regulators modulated by selective signaling pathways initiated by TNF-α binding to its specific receptor effects the survival or demise of cells. Therefore, better understanding of the signaling cascades, including that initiated by TNF-α receptor-1 occupancy, should provide further information about the molecular mechanisms, which account for the selective vulnerability of retinal ganglion cells to glaucomatous damage. 

Another line of evidence suggesting a potential role of TNF-α–mediated cell death in retinal ganglion cells in glaucoma is provided by previous observations on retinal heat shock protein expression. Induction of heat shock proteins in the central nervous system and in peripheral nerves in response to several environmental stresses, including ischemia, has been suggested to be an early response against stress that facilitates restoration of damaged areas after injury. ^{64 65 66} It has been reported that heat shock proteins, including hsp27 and hsp60, are upregulated in the retinal ganglion cells in glaucoma. ⁶⁷ This suggests that these proteins play a role as a native defense mechanism of stressed or injured neurons in glaucoma. One of the protective mechanisms attributed to heat shock proteins, particularly to hsp27, is that they counteract TNF-α–mediated disruption of actin architecture and enhance cellular resistance to TNF-α–mediated oxidative stress and apoptotic cell death. ^{68 69 70 71 72} A concurrent increase in the immunostaining of hsp27 ⁶⁷ and TNF-α in glaucomatous eyes, predominantly in the retinal ganglion cell layer, may therefore imply that hsp27 is a key component of the native defense mechanisms that provide preferential protection against TNF-α–mediated cell death in retinal ganglion cells. 

We observed that the immunostaining for TNF-α in glaucomatous eyes was more intense in retinal areas close to the optic nerve head than in the more peripheral retina. Serum protein has been shown to infiltrate into brain parenchyma after blood–brain barrier disruption that results in neuronal damage by activating glial cells to release neurotoxic substances such as TNF-α. It has been suggested that even though the size and duration of primary disruption of the blood–brain barrier is small, the disruption may allow some serum to leak from the circulation into the brain parenchyma in several neurologic diseases. The pathologically reactivated glial cells exposed to a very low concentration (0.1%) of the serum can eventually be activated to produce TNF-α in large quantities. ⁷³ We wonder whether a similar mechanism may be operative in glaucomatous eyes, because the blood–retina barrier may be defective within retinal areas close to the optic nerve head in these eyes. One of the indications of defective blood–retina barrier within this region is parapapillary chorioretinal atrophy, ^{74 75} which is a common finding in glaucomatous eyes and is associated with disease progression. ^{52 76 77 78} In addition, serum leakage into the retina is possible through nerve fiber hemorrhages that are commonly observed within this region in glaucomatous eyes and is similarly associated with disease progression. ^{76 79 80} Therefore, the increased immunostaining for TNF-α within retinal areas adjacent to the optic nerve head that we observed may suggest that increased participation of TNF-α–mediated cell death may contribute, in part, to increased susceptibility of neuronal tissues to glaucomatous damage within this region. ^{52 77 78}  

In conclusion, findings of the present study indicate that both protein and gene expression of TNF-α and TNF-α receptor-1 are upregulated in the retina of glaucomatous eyes. The presence of TNF-α receptor-1 in the retinal ganglion cells indicates that they are sensitive to the cytotoxic effects of TNF-α. Increased production of TNF-α by glial cells in glaucoma may therefore participate in the death of retinal ganglion cells through direct activation of the apoptotic cell death cascade. Improved understanding of molecular mechanisms of cell death and protection events in retinal ganglion cells, including events initiated by TNF-α receptor-1 binding to its native ligand TNF-α, may provide specific targets for pharmacologic interventions to modulate neuronal cell survival in glaucoma. 

 Supported in part by the Glaucoma Foundation, New York, New York (GT); Glaucoma Research Foundation (GT); Grant EY012314 (MBW) from the National Eye Institute; and an unrestricted grant to Washington University School of Medicine, Department of Ophthalmology and Visual Sciences from Research to Prevent Blindness Inc.

 Submitted for publication January 12, 2001; revised March 22, 2001; accepted April 6, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Clinical Data for Glaucomatous Donor Eyes

	Age (y)	Gender	Diagnosis	C/D	VF Damage
1	76	F	POAG	0.9	Advanced
2	94	F	POAG	0.7	Moderate
3	69	F	POAG	0.5	Moderate
4	69	F	POAG	0.5	Moderate
5	56	F	POAG	0.9	Advanced
6	82	M	POAG	N/A	N/A
7	78	M	POAG	N/A	N/A
8	74	F	POAG	0.7	Moderate
9	82	F	POAG	0.8	Moderate
10	91	F	POAG	0.9	Advanced
11	91	F	POAG	0.8	Moderate
12	84	F	NPG	0.95	Advanced
13	84	F	NPG	0.95	Advanced
14	68	F	NPG	0.8	Moderate
15	82	F	NPG	0.8	Moderate
16	82	F	NPG	0.8	Moderate
17	74	F	NPG	0.8	Moderate
18	74	F	NPG	0.9	Advanced
19	75	F	NPG	0.85	Advanced
20	75	F	NPG	0.8	Moderate

 C/D, cup-to-disc ratio; VF, visual field; POAG, primary open-angle glaucoma; NPG, normal pressure-glaucoma.

Figure 1.

 

Immunoperoxidase staining for TNF-α and TNF-α receptor-1 in the human retina. (A, C, and E) Immunostaining for TNF-α. (B, D, and F) Immunostaining for TNF-α receptor-1. There was faint immunostaining for both TNF-α and TNF-α receptor-1 in a few glial cells and their processes and in blood vessels (v) in the control retina (A, B). However, the intensity of the immunostaining and the number of stained cells were greater in retina sections from glaucomatous eyes. Although some immunostaining was detectable in all retinal layers, mostly associated with glial cells or blood vessels, immunostaining for TNF-α and TNF-α receptor-1 was most prominent in inner retinal layers, especially in the retinal nerve fiber (nfl) and ganglion cell (gc) layers (C, D). Black arrows: positive-labeled retinal ganglion cells; white arrows: negative-labeled retinal ganglion cells; black arrowhead: positive-labeled astrocytes; white arrowhead: negative-labeled astrocytes. Higher magnification images of the inner retina indicate that although immunostaining for TNF-α was mostly associated with glial cells lining the internal limiting membrane and surrounding retinal ganglion cells (E), immunostaining for TNF-α receptor-1 was predominant in the retinal ganglion cells (F). in, inner nuclear layer; on, outer nuclear layer. Chromagen, DAB; nuclear counterstain, Mayer’s hematoxylin. Magnification, (A– D) ×90; (E, F) ×375.

Figure 2.

 

Immunoperoxidase staining for TNF-α in different retinal areas in a glaucomatous eye. (A, arrows) Parapapillary chorioretinal atrophy area in a glaucomatous eye. (B) Higher magnification of the boxed area of inner retina in (A). (C) A similar inner area in more peripheral retina of the same eye. As seen in (B) and (C), immunostaining for TNF-α was more prominent in the retinal area close to the optic nerve head and adjacent to the area of parapapillary atrophy than in the more peripheral retina. v, blood vessel; ONH, optic nerve head. Chromagen, DAB; nuclear counterstain, Mayer’s hematoxylin. Magnification, (A) ×45; (B, C)× 150.

Figure 3.

 

In situ hybridization for TNF-α and TNF-α receptor-1 in the human retina. (A, C, and E) mRNA signals for TNF-α; (B, D, and F) mRNA signals for TNF-α receptor-1. There was some signal for mRNAs of TNF-α and TNF-α receptor-1 in the retina sections of normal donor eyes (A, B). However, the signals were greater in the glaucomatous retinas, especially in the inner retinal layers (C, D). Note induction of TNF-α mRNA in glial cells, especially in the retinal nerve fiber and ganglion cell layers, as well as in glial cells located close to a retinal blood vessel (v). However, TNF-α receptor-1 mRNA was predominantly localized to cells with large somas in the retinal ganglion cell layer. Control slides for in situ hybridization using sense RNA probes were negative for specific staining for either TNF-α or TNF-α receptor-1 (E, F). Chromagen, NBT/BCIP. Magnification, ×90.

Figure 4.

 

Double immunofluorescence labeling in retina sections from glaucomatous donor eyes. (A) Immunostaining for TNF-α. (B) Immunostaining for glial fibrillary acidic protein, a cell marker of glial cells. (C) Colocalization of TNF-α and glial fibrillary acidic protein. (D) Immunostaining for TNF-α receptor-1. (E) Immunostaining for Brn-3a, a cell marker of retinal ganglion cells. (F) Colocalization of TNF-α receptor-1 and Brn-3a. (G) Immunostaining for TNF-α. (H) Immunostaining for glial fibrillary acidic protein. (I) Immunostaining for TNF-α receptor-1. (J) Immunostaining for Brn-3a. (A–F) Retinal ganglion cell layer in which colocalization of TNF-α with glial fibrillary acidic protein and colocalization of TNF-α receptor-1 with Brn-3a are shown. (G–J) Lower magnification showing that this colocalization pattern was characteristic in the inner retinal layers, in which immunostaining for TNF-α was mostly detectable in cells positive for glial fibrillary acidic protein, and immunostaining for TNF-α receptor-1 was mostly localized to cells positive for Brn-3a. nfl, nerve fiber layer; gc, ganglion cell layer; in, inner nuclear layer; on, outer nuclear layer; v, blood vessel. Magnification, (A–F) ×200; (G–J)× 90.

The authors thank Belinda McMahan for excellent technical assistance.