Retinal Ganglion Cells Recognized by Serum Autoantibody against γ-Enolase Found in Glaucoma Patients

Abstract

purpose. To study pathologic roles of the presence of serum autoantibodies against retinal ganglion cells in patients with glaucoma.

methods. Serum autoantibody reactions were detected by Western blot analysis using retinal soluble fractions in 79 patients with glaucoma (normal-tension glaucoma [NTG], 23 cases; primary open-angle glaucoma[ POAG], 56 cases) and 60 age-matched healthy subjects. Clinical characteristics including visual acuity, visual field, intraocular pressure (IOP), and optic disc features were compared between the serum autoantibody–positive and –negative patients. The retinal autoantigen recognized by patients’ sera was identified by a combination of in-gel digestion and Edman sequencing.

results. Western blot analysis revealed that serum autoantibody against retinal 50-kDa antigen was recognized in 20 out of 79 glaucoma patients (25.3%; 14 POAG and 6 NTG patients) and 60 age-matched control subjects (11.7%), respectively. Immunocytochemistry revealed that labeling of the ganglion cell layer (GCL) by IgG from glaucoma patients (POAG: 13/56, 23.2%; NTG: 6/23, 26%) existed at a significantly higher rate than that by IgG from control subjects (2/60, 3.3%; P < 0.05). In POAG, maximum IOP in the serum antibody positive–patients was significantly lower than that in the antibody-negative patients (P < 0.05). However, no statistical differences were observed in visual field loss, disc cupping, and other clinical factors between the antibody-positive and -negative groups in POAG and NTG. In-gel digestion of the 50-kDa band in two-dimensional polyacrylamide gels and Edman sequence analysis of the high-performance liquid chromatography–purified peptides identified the 50-kDa protein as γ-enolase. Injection of the 50-kDa IgG from glaucoma patients or anti-γ-enolase serum into the vitreous cavity of Lewis rats caused reduction of the b-wave of the electroretinogram and TdT-dUTP terminal nick-end labeling (TUNEL)–positive staining within the GCL.

conclusions. In the current study, serum autoantibody against 50-kDa protein identified as γ-enolase in 25% of glaucoma patients.

Glaucoma is known to be a major cause of optic neuropathy, eventually leading to loss of vision. Glaucomatous optic neuropathy is characterized by loss of retinal ganglion cells and their axons, excavated appearance of the optic nerve head, and progressive loss of visual field sensitivities. Clinically, two major forms of the glaucomatous optic neuropathy are known: primary open-angle glaucoma (POAG) and normal-tension glaucoma (NTG), which are associated with elevated and normal intraocular pressure (IOP), respectively. ¹ In terms of the pathologic course of the glaucomatous optic neuropathy, retinal ganglion cell death by apoptosis has been identified in postmortem studies of human eyes with POAG ^{2 3} and in experimental glaucoma models with elevated IOP. ^{4 5} Although the molecular mechanism triggering the apoptosis has not been identified, deprivation of neurotrophic factors, ⁴ ischemia, ⁶ chronic elevation of glutamate, ⁷ and disorganized nitric oxide (NO) metabolism ⁸ have been suspected to be possible mechanisms. In addition, it was recently suggested that autoantibodies directed toward retinal antigens, such as rhodopsin, ⁹ 60-kDa heat shock protein (hsp 60), ¹⁰ 27-kDa heat shock protein (hsp 27), and α-crystallin ¹¹ may be involved in facilitating apoptotic cell death in some glaucoma patients, particularly in NTG. 

To study the contribution of autoimmune factors in glaucomatous optic neuropathy, we examined sera from 79 patients with NTG or POAG and 60 age-matched control subjects to detect specific serum autoantibodies in glaucoma. 

Materials and Methods

The studies were performed in accordance with our institution’s guidelines and the Declaration of Helsinki on Biomedical Research Involving Human Subjects. Protocols were approved by the institution’s Committee for the Protection on Human Subjects. All experimental procedures were designed to conform to both the ARVO Statements for Use of Animals in Ophthalmic and Vision Research and our own institution’s guidelines. 

Patients

Seventy-nine patients with glaucoma (NTG, 23 cases; POAG, 56 cases), and 60 age-matched healthy subjects were used in the present study. The diagnostic criteria for NTG were as follows: the presence of open iridocorneal angles, no evidence of IOP higher than 21 mm Hg, glaucomatous changes in visual fields, optic nerve cupping, and the absence of alternative causes of optic neuropathy. Criteria for diagnosis of POAG were identical with that of NTG, except the IOP had to be higher than 21 mm Hg. IOP was determined by Goldmann applanation tonometer (Haag Streit, Bern, Switzerland). Visual field was examined by Goldmann perimeter (Haag Streit). Peripheral venous blood samples were immediately subjected to serum separation and stored at− 80°C before use. For immunocytochemistry, the serum samples were subjected to IgG purification using protein G column chromatography as described by Ohguro et al. ¹²  

Western Blot Analysis

Western blot analysis was performed using bovine or rat retinal soluble protein fractions, as described previously. ¹² The sample containing approximately 20 μg protein was analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) using a 12.5% polyacrylamide gel. Electrotransferred polyvinylidene fluoride (PVDF) membranes were successively probed by sera (1:400 dilution) and horseradish peroxidase–labeled anti-human IgG (1:3000 dilution) after nonspecific binding was blocked with 5% skim milk in phosphate-buffered saline (PBS). Specific antigen and antibody binding was visualized by chemiluminescence (ECL; Amersham Pharmacia Biotech, Buckinghamshire, UK). 

Identification of the Retinal Antigens by the In-Gel Digestion Method

For identification of the retinal antigens reacted with glaucoma patients’ sera, in-gel digestion of 2D-PAGE of the rat retinal soluble fraction (approximately 100 μg of protein) by endoproteinase Lys C was performed by a method described previously. ¹² The reversed-phase high-performance liquid chromatography (HPLC)–purified peptides were analyzed by Edman degradation sequencing method. ¹²  

Vitreous Injection of Antibodies to Lewis Rats

Six-week-old Lewis rats (approximately 180 g) reared in cyclic light conditions (12 hours on–12 hours off) were used. For anesthesia induction, rats inhaled diethylether. Once unconscious, the animals were injected intramuscularly with a mixture of ketamine (80–125 mg/kg) and xylazine (9–12 mg/kg). Adequacy of the anesthesia was tested by tail clamping, and supplemental doses of the mixture were administrated intramuscularly if needed. To rats under anesthesia, a total of 5 μl of patients’ IgG (2 mg/ml), control IgG (2 mg/ml), anti-γ-enolase serum, or anti-α-crystallin serum was administrated into the vitreous cavity of a Lewis rat eye. The injection was performed with a 26-gauge Hamilton microneedle syringe (Wilmad, Reno, NV) through the sclera at a point 1 mm from the limbus to avoid puncturing the lens. Animals showing apparent traumatic changes after vitreous injection, such as cataract, were excluded from the study. After the surgery, a drop of 0.5% ofloxacin was administrated to avoid infection. Anti-human γ-enolase serum and anti-human α-crystallin serum were purchased from UltraClone (Wellow, UK) and StressGen Biotech (Victoria, British Columbia, Canada), respectively. Both sera were free of any known retinal toxic substances, such as high concentrations of glutamate. ⁷ The specificity and titers of these antibodies were examined by Western blot, by using bovine retinal soluble fractions, as described in our previous study, ¹² before the antibody penetration experiment was performed as described. To exclude systemic effects between two eyes, one of the antibodies, randomly chosen, was administrated to the left eye and a different antibody to the right eye in each rat. 

Electroretinogram

The anesthetized animals were kept in dark adaptation for at least 1 hour in an electrically shielded room. The pupils were dilated with drops of 0.5% tropicamide. The scotopic electroretinogram (ERG) response was recorded with a contact electrode equipped with a suction apparatus to fit on the cornea (Kyoto Contact Lens, Kyoto, Japan). A grounding electrode was placed on the ear. Responses evoked by white flashes (3.5 × 10² lux, 200-msec duration) were recorded by a clinically used ERG recording instrument (Neuropack MES-3102; Nihon Kohden, Tokyo, Japan). 

Light Microscopy

Anesthetized animals were transcardially perfused with 100 ml 82-mM sodium phosphate buffer (pH 7.2) containing 4% paraformaldehyde. Enucleated eyes were embedded in paraffin and sectioned at 3 μm thickness, mounted on subbed slides, and dried. The sections were processed with hematoxylin–eosin staining after deparaffinization with graded ethanol and xylene solutions. Apoptotic cells in the retinal sections were detected by TdT-dUTP terminal nick-end labeling (TUNEL) stain using a commercially available kit (Takara Shuzo, Shiga, Japan), according to the protocol described by the manufacturer. 

Immunofluorescence Microscopy

Unfixed freshly dissected rat retinas were infiltrated with 30% sucrose in PBS at 4°C, cryosectioned at 4 μm thickness, mounted on subbed slides, air dried, and stored at −80°C before use. The sections were treated with ice-cold acetone for 10 minutes and air dried, and plastic rings were mounted around the sections to form incubation walls. For immunostaining with patients’ or control IgG, the sections were incubated with IgG (1:100 dilution) for 1 hour at room temperature. The sections were then rinsed three times with PBS for 5 minutes, and incubated with goat anti-human IgG labeled with Cy3 (Jackson ImmunoResearch, West Grove, PA) at 1:400 in PBS with 0.3% Tween 20 at room temperature for 1 hour. After the sections were washed three times with PBS for 5 minutes, they were coverslipped in mounting medium for immunofluorescence (Vectashield; Vector, Burlingame, CA). The sections were photographed using a Cy3 filter set. For double staining by anti-Thy-1 and TUNEL, acetone-treated sections, as described, were incubated with mouse anti-rat Thy-1 (1:100 dilution; Accurate Chemical & Scientific, Hornby, Ontario, Canada) for 1 hour at room temperature. The sections were washed with PBS as above and then incubated with a mixture of goat anti-mouse IgG labeled with Cy3 (Jackson ImmunoResearch) at 1:800, and terminal dUTP transferase and fluorescein-isothiocyanate (FITC)–labeled dUTP in TdT buffer (30 mM Tris-HCl, [pH 7.2] containing 140 mM sodium cacodylate and 1 mM cobalt chloride; Takara Shuzo) at room temperature for 1 hour. After the sections were rinsed with PBS and coverslipped in mounting medium as described earlier, they were observed with a laser scanning confocal microscope in the transmitted-light mode. 

Statistic Analysis

The clinical data, including age, maximum and mean IOPs, and disc cupping, and the experimental data, including ERG amplitudes and apoptotic cell counts, are expressed as means ± SD. Significant differences between groups were found using the Mann–Whitney test with a significance level of P < 0.05. Positive rates of anti-50-kDa or GCL labeling between glaucoma and control groups were statistically analyzed by χ² test with a significance level of P < 0.05. 

Results

As shown in Figure 1 and Table 1 , Western blot analysis using bovine retinal soluble extracts revealed that a 50-kDa protein band was specifically probed by sera from 14 patients with POAG (25.0%) and 6 with NTG (26.1%). Among 60 age-matched healthy control subjects, sera of 7 subjects (11.7%) showed immunoreactivity toward the 50-kDa protein. The rates of presence of the anti-50-kDa antibody were relatively higher in the glaucoma groups than those in control subjects, but these differences were not significant (χ² test: POAG versus control, P = 0.062; NTG versus control, P = 0.105). To further characterize the serum autoantibody, immunolabeling of frozen rat sections was performed. As shown in Figure 2 , the labeling of the retina occurred in five distinct patterns: 1) ganglion cell layer (GCL) labeling (lane 1), 2) inner nuclear layer (INL) labeling (lane 2), 3) GCL + INL labeling (lane 3), 4) diffuse labeling (lane 4), and 5) no staining. Among the several staining patterns, GCL (patterns 1, 3, and 4) was recognized in sera from 13 and 6 of the 56 patients with POAG and 23 patients with NTG, respectively. In contrast, GCL labeling was observed in sera from 2 of 60 control subjects. These rates of positive GCL labeling in POAG or NTG were significantly higher than those in control subjects (χ² test: POAG versus control, P = 0.014; NTG versus control, P = 0.017). In 19 of 20 glaucoma patients with the 50-kDa antibody, the GCL labeling was recognized, whereas only 2 of 7 control subjects with the 50-kDa antibody showed GCL labeling. 

To identify the 50-kDa antigen, soluble extracts from rat retinas were subjected to a 2-D PAGE followed by staining with Coomassie blue and immunostaining with patient’s serum (Fig. 3) . The corresponding band in gels were collected, cut out, and subjected to in-gel digestion with endoproteinase Lys C. The resultant peptides were purified on a reversed-phase HPLC C18 column, by using a linear gradient of acetonitrile from 0% to 80% during 60 minutes (Fig. 4 , top). Edman sequence analysis of the major peak fractions revealed the 50-kDa antigen to be γ-enolase (neuron-specific enolase; Fig. 4 , bottom). 

To elucidate clinical significance of the presence of these serum autoantibodies, clinical factors such as IOP, visual field loss, and disc cupping were compared between serum autoantibody–negative and– positive groups (Table 2) . In each group, one eye was randomly selected from each patient. In POAG (n = 56 eyes), maximum IOP and mean IOP during the follow-up period were statistically (P < 0.05) and relatively lower, respectively, in the positive group (n = 14 eyes) than in the negative group (n = 42 eyes). However, other clinical observations, including follow-up periods, age, sex, visual field defects, disc cupping, and treatments with eye drops, were comparable between the two groups. These observations suggested that the autoantibody-positive group showed degrees of the visual fields loss similar to those in the negative group, even though the maximum IOP levels of the former were lower than those in the latter in POAG. No significant difference in the clinical factors was observed between the autoantibody-positive (n = 6 patients, 12 eyes) and -negative groups (n = 17 patients, 34 eyes) in NTG (n= 23 patients, 46 eyes). 

To study the pathogenic effects of autoantibody against γ-enolase on retinal cells, we injected IgG from patients with POAG with the 50-kDa antibody or anti-human γ-enolase serum into the vitreous cavity of Lewis rats (n = 12 rats, 12 eyes at each condition). As a control, IgG from normal subjects without the 50-kDa antibody was injected (n = 12 rats, 12 eyes). In addition, these effects were compared by intravitreal administration of anti-α-crystallin (n = 12 rats, 12 eyes), which has been recently identified as a retinal autoantigen in patients with NTG. In advance, the specificity and the titers of these antibodies were determined by a Western blot analysis using retinal soluble homogenates. The specific labeling by IgG from glaucoma patients (lane 1), anti-γ-enolase serum (lane 2), and anti-α-crystallin serum (lane 3) were obtained by up to 1:3000, 1:6000, and 1:6000 dilutions, respectively (Fig. 5) . After the injection, evaluations of retinal function and morphology were performed by ERG and light microscopy examination of the retinal sections. Examinations by slit lamp and fundoscopy detected no significant changes, such as retinal detachment, vitreoretinal hemorrhage, uveitis, or cataract in any animals without trauma after the injection. 

In ERG, significant lower amplitudes of b-wave in ERG were observed in eyes injected with anti-50-kDa IgG or anti-γ-enolase serum compared with control IgG, 1 week after the injection (P < 0.05; Fig. 6A ). However, no significant changes were observed in eyes that received anti-α-crystallin serum (Fig. 6A) . Light microscopy of the retinal sections stained by hematoxylin–eosin showed no significant changes such as destruction of cell morphology and lymphocyte infiltration in the retina (Fig. 7A ). In TUNEL staining of the sections, apoptotic cells within the GCL were observed in the affected retinas with patients’ IgG (Fig. 7B) or anti-γ-enolase serum (Fig. 7C) , and TUNEL-positive nuclei were significantly higher in number than in control or anti-α-crystallin antibody–treated retinas (Fig. 6B , P < 0.01). In addition, few TUNEL-positive cells were identified in the INL and ONL in these retinas. However, the numbers of the positive nuclei were not different among these four groups (data not shown). 

To determine whether the TUNEL-positive cells were ganglion cells, double staining by TUNEL and anti-Thy-1 antibody, the specific maker of the ganglion cells, ¹³ was performed. As shown in Figure 8 , positive-stained cells with both anti-Thy-1 antibody and TUNEL were observed in the anti-γ-enolase serum–treated retinas. Taken together, the above observations suggested that serum autoantibody against γ-enolase may induce apoptosis of retinal ganglion cells. 

Discussion

Regarding the association of the cause of glaucoma with autoimmune mechanisms, the presence of immunorelated diseases, increased prevalence of monoclonal gammopathy, ¹⁴ and presence of serum antibodies toward rhodopsin, ⁹ heat shock proteins including hsp 60, ¹⁰ hsp 27, andα -crystallin ¹¹ have been reported in glaucoma patients. These observations suggest that autoimmunity may be involved in the pathogenesis of glaucomatous optic neuropathy. In the present study, we found serum autoantibodies toward γ-enolase in approximately 20% of patients with glaucoma (POAG and NTG). However, no immunoreactivities toward rhodopsin, α-crystalline, hsp 27, and hsp 60 were recognized. Although the serum antibody toward the 50-kDa antigen was also identified in approximately 10% of the healthy control subjects, immunofluorescence labeling revealed that most of the glaucoma patients’ sera specifically reacted with GCL, whereas sera from control subjects recognized other or no retinal layers. In addition, maximum and mean IOP were significantly and relatively lower in the antibody-positive patients than in the negative ones. Intravitreal injection of patients’ IgG or anti-γ-enolase serum caused lowering of b-wave in ERG- and TUNEL-positive cells within the GCL of Lewis rats. From this evidence, we conclude that serum autoantibody toward GCL-specific γ-enolase may cause apoptotic cell death within GCL. 

Enolase (2-phospho-d-glycerate hydroxylase) is the glycolytic enzyme, that occurs in three homologous but distinct forms:α , found in many tissues; β, predominant in muscle; and γ (neuron-specific enolase), found only in neurons and neuroendocrine tissue. ¹⁵ Usually enolases are present as homo- or hetero-oligomers. Recently, α-enolase was identified as one of the autoantigens of cancer-associated retinopathy (CAR). ¹⁶ CAR is a visual paraneoplastic syndrome that has been identified in small cell carcinoma of the lung and other malignant tumors. ¹⁷ CAR is clinically characterized by photopsia, progressive visual loss with a ring scotoma, attenuated retinal arterioles, and abnormalities of the a- and b-waves of the ERG. It has been suggested that photoreceptor cell death (apoptosis) may be caused by an autoimmune reaction against enolase and other retinal antigens, including recoverin, ^{12 17 18} heat shock cognate protein (hsc) 70 ¹² and neurofilaments. ¹⁹ We do not know why autoimmune reaction toward enolase cause apoptosis of two different retinal cell layers: the photoreceptor layer in CAR and GCL in glaucoma. This may be ascribed to the difference in the immunoreactivities of serum autoantibody against γ-enolase toward retinal cells as shown in Figure 3 . In CAR, Adamus et al. ¹⁶ reported a rate of presence of autoantibody almost identical with that of α-enolase in normal subjects (10 positive out of 110 people), and different immunolabeling patterns by the serum antibody, similar to our results. We speculated that there were several possible reasons for the difference in immunoreactivities of the enolase antibodies between glaucoma patients, CAR patients, and control subjects as follows: γ-enolase was the autoantigen in glaucoma patients, but other enolases (α- and β-) may function as the autoantigen in CAR patients and control subjects; and IgG of glaucoma patients, CAR patients, and control subjects may differ in affinity, avidity, specificity, or subclass, even though autoantibodies of all the groups were anti-γ-enolase. In addition, our experiment of intravitreal injection of the patients’ IgG into Lewis rat eyes demonstrated TUNEL-positive cells within GCL but not in the other layers, even though the IgG recognized not only GCL but also INL and/or ONL in immunocytochemistry. This difference between TUNEL staining and immunostaining may be ascribed to the differences in viability among retinal neuronal cells. In fact, this different viability has been identified in several animal models—for example, apoptosis of GCL after intravitreal injection of N-methyl-d aspartate (NMDA) in rats ²¹ and apoptotic cell death, predominantly in INL in ischemia–reperfusion of rats. ²¹ As another possibility, additional factors may be required for apoptosis in the other retinal layers (e.g., the presence of anti-hsc 70 antibody facilitates anti-recoverin antibody induced apoptosis of photoreceptors in CAR ^{12 22} ). 

Another important question is how the serum anti-γ-enolase antibodies get to the target cells and cause the apoptosis processes. Regarding antibody internalization, several lines of experimental evidence have been reported in paraneoplastic disorders including CAR and autoimmune diseases. ^{23 24} If this is the case, we reasonably speculate that anti-γ-enolase antibodies in the peripheral blood circulation may cause apoptosis of retinal ganglion cells in the presence of additional unknown factors causing breakdown of the blood–retinal barrier to facilitate the antibody to access to the target antigens. 

In conclusion, in our experiments γ-enolase was recognized as an autoantigen in some glaucoma patients, and this may be related to the molecular pathogenesis of glaucoma. This is still speculative at present, however, and further investigation is needed. 

 Supported by grants from the Japanese Ministry of Health, Naito Memorial Foundation, Ciba–Geigy Foundation for the Promotion of Science, The Mochida Memorial Foundation for Medical and Pharmaceutical Research, Uehara Memorial Foundation, and JRPS Research Foundation.

 Submitted for publication September 24, 1999; revised December 28, 1999; accepted January 26, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

Western blot analysis of sera from 23 patients with NTG and 25 selected patients with POAG and 20 selected control subjects. Sera from 23 patients with NTG, 25 selected patients with POAG, and 20 control subjects (1:400 dilution) were tested with bovine retinal soluble extract. Retinal 50-kDa protein (arrow) was probed by patients’ sera (NTG: lanes 9, 12, 15, 18, 21, and 22; POAG: lanes 2, 4, 5, 8, 11, 12, and 15; control: lanes 1, 6, and 13).

Table 1.

 

Presence of Anti-50-kDa Antibody in Glaucoma Patients and Immunoreactivity of the Antibody in Rat Retinal Sections

	POAG (n = 56)	NTG (n = 23)	Control (n = 60)
Immunoblot analysis
Anti-50 kDa (+)	14/56	6/23	7/60
Immunocytochemistry
GCL staining (+) (patterns 1, 3, 4)	13/56* (13/14)	6/23* (6/6)	2/60 (2/7)
Staining patterns
1) GCL	12 (12)	5 (5)	0 (0)
2) INL	0 (0)	0 (0)	2 (2)
3) GCL+ INL	0 (0)	0 (0)	1 (1)
4) Diffuse	1 (1)	1 (1)	1 (1)
5) No staining	43 (1)	17 (0)	56 (3)

 Distinct patterns 1 through 4 of imunocytochemistry are shown in Fig. 2 . Data in parentheses are from the serum 50-kDa autoantibody–positive group.

*   P < 0.05, compared with control by χ² test.

Figure 2.

 

Representative micrograph of immunofluorescence labeling of the rat retinal section by anti-50-kDa autoantibodies. Immunocytochemistry identified four distinct patterns of retinal labeling by anti-50-kDa IgG (1:100 dilution); 1) GCL staining, 2) INL staining, 3) GCL + INL staining, 4) diffuse. GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; OPL, outer plexiform layer; ONL, outer nuclear layer; OS, outer segment. Scale bar, 50 μm.

Figure 3.

 

Isolation of 50-kDa antigen by 2-D gel electrophoresis. Rat retinal soluble proteins (approximately 100 μg) were separated with 2-D gel electrophoresis (left), electroblotted to PVDF membrane, and immunostained using glaucoma patient serum (1:400 dilution; right). The immunoreactive spot and the corresponding band in the gel are indicated by filled and open arrows, respectively.

Figure 4.

 

Identification of 50-kDa retinal antigen. Top: The immunoreactive 50-kDa protein in 2-D gels was treated with endoproteinase Lys C (1 μg). Digested peptides were separated from each other by reversed-phase HPLC, and major peaks, designated as 1 through 8, were subjected to Edman sequencing. Bottom: Comparison of amino acid sequences of the proteolytic peptides from the 50-kDa protein and rat γ-enolase sequence. ²⁵ Sequence ofα -enolase ²⁶ that differs from that of γ-enolase is indicated with parenthesis.

Table 2.

 

Clinical Characteristics of Anti-50 kDa–Positive and –Negative Glaucoma Patients

	POAG (n = 56, 100 eyes)*			NTG (n = 23, 46 eyes)
		Anti-50 kDa (+) (n = 14, 21 eyes)	Anti-50 kDa (−) (n = 42, 79 eyes)	Anti-50 kDa (+) (n = 6, 12 eyes)	Anti-50 kDa (−) (n = 17, 34 eyes)
Sex
Male/female	7/7	23/19	1/5	8/9
Age	62.8 ± 7.8† >	62.4 ± 12.4	65.5 ± 10.3	62.8 ± 10.6
Maximum IOP (mm Hg)	23.4 ± 2.2, †	25.4 ± 4.4, †	18.3 ± 1.6	17.6 ± 1.8
Mean IOP (mm Hg)	16.4 ± 4.0† >	18.3 ± 4.1	14.2 ± 1.2	14.3 ± 1.7
Visual field loss (%)
Nasal step	10 (47.6)	18 (22.8)	6 (50)	20 (58.8)
Paracentral scotoma	5 (23.8)	16 (20.3)	7 (58.3)	10 (29.4)
Arcuate scotoma	8 (38.1)	10 (12.7)	3 (25)	10 (29.4)
Central island	1 (4.8)	2 (2.5)	0	1 (2.9)
Temporal island	0	0	0	1 (2.9)
Eye drops
β-blocker	14 (66.7)	54 (68.4)	5 (41.7)	19 (55.9)
Miopic	8 (38.1)	21 (26.6)	0	2 (5.9)
PG derivative	9 (42.9)	40 (50.6)	7 (58.3)	18 (52.9)
Adrenergic	6 (28.6)	14 (17.7)	1 (8.3)	2 (5.9)
Disc cupping	0.79 ± 0.24	0.75 ± 0.19	0.8 ± 0.2	0.75 ± 0.22
Trabeculectomy	6 (28.6)	20 (25.3)	1 (8.3)	0

 Male and female data represent persons; data for visual field loss, eye drops and trabeculectomy represent numbers of eyes. Data are means ± SD, unless marked otherwise.

*  Eyes that had no episode of IOP of more than 21 mm Hg were excluded.

†   P < 0.05; Mann–Whitney test.

Figure 5.

 

Western blot analysis of antibodies for in vivo administration. Anti-50-kDa IgG from glaucoma patients (lane 1, 1:2000 dilution; 0.5 μg/ml), anti-human γ-enolase (lane 2, 1:3000 dilution), and anti-rabbit α-crystallin (lane 3, 1:3000 dilution) were tested with bovine retinal soluble extract.

Figure 6.

 

Comparisons of ERG amplitudes of the b-wave and TUNEL-positive nuclei of retinas intravitreously treated with patients’ IgG, control IgG, anti-γ-enolase serum, or anti-α-crystallin. One week after the injection of antibodies into Lewis rat eyes (12 rats, 12 eyes in each experimental condition) ERG measurement (A) and histopathologic study by TUNEL staining (B) were performed. Numbers of rats showing similar changes in ERG were 10 (anti-50-kDa), 10 (anti-γ-enolase serum), 12 (control IgG), and 11 (anti-α-crystallin), respectively. In TUNEL staining, positive nuclei were counted over a 1-mm horizontal length of the retinal sections in 10 different areas. Data are expressed as means ± SD.* P < 0.05, **P < 0.01 (Mann–Whitney test).

Figure 7.

 

Histopathologic changes in the retina of Lewis rats treated with anti-50-kDa IgG from glaucoma patients or anti-γ-enolase serum. Hematoxylin–eosin staining (A) or TUNEL staining (B, C) of retinal sections near the posterior pole in Lewis rat eyes, which were treated with patients’ IgG (A, B) or anti-γ-enolase serum (C). TUNEL-positive nuclei are indicated by an arrowhead. Abbreviations: see Figure 2 . Scale bar, 50 μm.

Figure 8.

 

Confocal microscopicimages of immunostaining by anti- Thy-1 and TUNEL of anti-γ-enolase–treated rat retina. (A) Immunostaining by anti-rat Thy 1 serum (red), (B) TUNEL staining (green), (C) digitally overlaid image of (A) and (B). A retinal cell stained by anti-Thy-1 and/or TUNEL is indicated by an arrow. Abbreviations: see Figure 2 . Scale bar, 50 μm.