In Vivo Protection against Retinal Neurodegeneration by Sigma Receptor 1 Ligand (+)-Pentazocine

Abstract

Purpose

To evaluate the neuroprotective properties of the sigma receptor 1 (σR1) ligand, (+)-pentazocine in an in vivo model of retinal neurodegeneration.

Methods

Spontaneously diabetic Ins2^Akita/+ and wild-type mice received intraperitoneal injections of (+)-pentazocine for 22 weeks beginning at diabetes onset. Retinal mRNA and protein were analyzed by RT-PCR and Western blot analysis. Retinal histologic sections were measured to determine total retinal thickness, thicknesses of inner-outer nuclear and plexiform layers (INL, ONL, IPL, INL), and the number of cell bodies in the ganglion cell layer (GCL). Immunolabeling experiments were performed using antibodies specific for 4-hydroxynonenal and nitrotyrosine, markers of lipid peroxidation, and reactive nitrogen species, respectively, and an antibody specific for vimentin to view radial Müller fibers.

Results

σR1 mRNA and protein levels in the Ins2^Akita/+ retina were comparable to those in the wild-type, indicating that σR1 is an available target during the disease process. Histologic evaluation of eyes of Ins2^Akita/+ mice showed disruption of retinal architecture. By 17 to 25 weeks after birth, Ins2^Akita/+ mice demonstrated ∼30% and 25% decreases in IPL and INL thicknesses, respectively, and a 30% reduction in ganglion cells. In the (+)-pentazocine-treated group, retinas of Ins2^Akita/+ mice showed remarkable preservation of retinal architecture; IPL and INL thicknesses of (+)-pentazocinetreated Ins2^Akita/+ mouse retinas were within normal limits. The number of ganglion cells was 15.6 ± 1.5 versus 10.4 ± 1.2 cells/100 μm retinal length in (+)-pentazocine-treated versus nontreated mutant mice. Levels of nitrotyrosine and 4-hydroxynonenal increased in Ins2^Akita/+ retinas, but were reduced in (+)-pentazocine-treated mice. Retinas of Ins2^Akita/+ mice showed loss of the uniform organization of radial Müller fibers. Retinas of (+)-pentazocine-treated mice maintained the radial organization of glial processes.

Conclusion

Sustained (+)-pentazocine treatment in an in vivo model of retinal degeneration conferred significant neuroprotection, reduced evidence of oxidative stress, and preserved retinal architecture, suggesting that σR1 ligands are promising therapeutic agents for intervention in neurodegenerative diseases of the retina.

Sigma 1 receptors (σR1) are Ca²⁺-sensitive, ligand-operated receptor chaperones at the mitochondrion-associated endoplasmic reticulum (ER) membrane.¹ Originally, they were thought to be a subtype of opiate receptor, but it is clear now that σR1 is pharmacologically distinct from other known receptors. Two types of sigma receptors (σR1 and σR2) have been identified²; σR1, the better characterized of the two, has been cloned from several species.^3-7 σR1 cDNA predicts a protein of 223 amino acids (M_r 25-28 kDa).³

σR1 has been implicated in neuroprotection.^8-14 In the brain, where σR1 is expressed abundantly, ligands for the receptor have been useful in attenuating neuronal loss in in vitro^9,11 and in vivo models of acute neurodegeneration.^10,13,14 Over the past several years, σR1 expression has been analyzed in ocular tissue,^15-21 including the retina.^18-21 RT-PCR analysis amplified σR1 in neural retina and the RPE-choroid complex. In situ hybridization studies revealed abundant expression of σR1 in the ganglion cell layer, (GCL) inner nuclear layer (INL), inner segments of photoreceptor cells, and RPE cells.¹⁸ Immunohistochemical analysis confirmed these observations. Recent studies in which primary cultures of mouse Müller cells were used localized σR1 to the ER and nuclear membranes.²¹ These cells and other retinal cell types demonstrate robust σR1 binding activity with an apparent K_d of ∼25 nM.²¹

The promising evidence of neuroprotection by σR1 ligands in studies of brain tissue coupled with the finding that σR1 is expressed in retina led us to explore the usefulness of ligands for σR1 in retinal degeneration. Our initial studies were performed in vitro using a ganglion cell line, RGC-5,²² and subsequently using primary ganglion cells isolated from neonatal mouse retina.⁸ These studies demonstrated that treatment of cells with (+)-pentazocine, a highly-specific σR1 ligand, led to marked attenuation of cell death induced by the excitotoxins glutamate and homocysteine. These findings suggested that in vivo neurodegenerative diseases of the retina, such as glaucoma and diabetic retinopathy, which result in progressive loss of retinal neurons, may be amenable to treatment with σR1 ligands. Earlier studies, showing that σR1 continues to be expressed in neural retina under hyperglycemic conditions and during diabetic retinopathy,²⁰ prompted analysis of neuroprotection in this disease.

Diabetic retinopathy, a leading cause of blindness, is characterized by loss of retinal neurons and disruption of vasculature.^23,24 Patients with diabetes lose color and contrast sensitivity within 2 years of onset.^25,26 Focal ERG (electroretinogram) analysis, which detects electrical responses of ganglion cells, reveals dysfunction of these cells early in diabetes.^27,28 Analyses of retinal tissue samples from diabetic patients provide further support of an involvement of retinal neurons in diabetic retinopathy, including detection of several markers of apoptosis.²⁹ The neuronal death observed in humans has also been reported in rodent models.³⁰ Using the TUNEL assay, Barber et al.³⁰ analyzed the retinas of diabetic patients as well as the streptozotocin-induced rat model and observed significantly more apoptotic neurons in retinas of diabetic subjects than in those of control subjects. Similar results were observed by Bek³¹ and Kerrigan et al.³²

Recently, two mouse models of diabetic retinopathy were described in which retinal neuronal loss was reported.^33,34 Spontaneously arising diabetes has been reported in the Ins2^Akita/+ mouse, which has a point mutation of the Insulin2 gene leading to hyperglycemia and hypoinsulinemia in heterozygous mice by ∼4 weeks.³⁵ In addition to increased retinal vascular permeability and an increase in acellular capillaries, Ins2^Akita/+ mice demonstrate ∼20% to 25% reduction in the thickness of the inner plexiform layer (IPL), ∼16% reduction in the thickness of the INL, and ∼25% reduction in the number of cell bodies in the retinal GCL.³³ Cells in the GCL are immunoreactive for active caspase-3 after 4 weeks of hyperglycemia, consistent with cell death by apoptosis. In an induced model of diabetes, mice made diabetic by multiple injections of the pancreatic toxin streptozotocin, but not maintained on insulin, demonstrated 20% to 25% fewer cells in the GCL by 14 weeks after the onset of diabetes.³⁴ Data obtained using the TUNEL assay and analysis of active caspase-3 suggested that these neurons die by apoptosis. Viability is compromised in these animals, and most animals die by 16 to 20 weeks after onset of diabetes. Given that retinal neurons die over a relatively short period in these two mouse models, they were ideal for assessing neuroprotection by the σR1 ligand (+)-pentazocine. A pilot study to test (+)-pentazocine dosage was conducted in the streptozotocin-induced model and the comprehensive analysis of neuroprotection was performed in the Ins2^Akita/+ model.

MATERIALS AND METHODS

Animals

These studies adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. For the pilot study to determine an effective (+)-pentazocine dosage, 3-week old C57Bl/6 female mice, purchased from Harlan Sprague-Dawley (Indianapolis, IN), were made diabetic with streptozotocin, according to our published protocol.³⁴ C57BL/6J Ins2^Akita/+ mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Ins2^Akita/+ female mice were bred to wild-type male mice in the animal facilities of the Medical College of Georgia, in accordance with institutional policy for appropriate care and use of animals. Genotyping of Ins2^Akita/+ mice was performed per the protocol recommended by The Jackson Laboratory. Retinas of male and female mice were examined; however, only those of the males were used in the morphometric retinal analyses, because disease progression in females is slower and less uniform.³⁵ Diabetes onset was verified by measuring urine glucose with a urine strip test (Clinistix; Fisher Scientific, Pittsburgh, PA) and blood glucose with a blood glucose meter (Prestige Smart System; Home Diagnostics, Inc., Ft. Lauderdale, FL). Fasting blood glucose levels >250 mg/dL were considered diabetic. Diabetic animals were not maintained on insulin. Age-matched, nondiabetic wild-type mice were used as the control. All mice received food and water ad libitum and were maintained on a 12-hour lightdark cycle. Body weight and blood glucose were measured at the time of death (Table 1).

Table 1.

Average Mouse Weights and Blood Glucose Levels

Treatment Groups	n	Mean Weight ± SEM (g)	Blood Glucose ± SEM (mg/dL)	Age of Mice/Duration of Diabetes (wk)
Pilot Study Testing Two Dosages of (+)-Pentazocine in the STZ-Induced Diabetic Mouse
Control	4	24.2 ± 1.5	119.5 ± 28.6	14 wk/nondiabetic
STZ-DB	4	19.0 ± 1.8	346.8 ± 21.5	14 wk/10 wk diabetic
STZ-DB/1.0 mg · kg-1 (+)-PTZ	2	18.9 ± 1.1	342.0 ± 35.3	14 wk/10 wk diabetic
STZ-DB/0.5 mg · kg-1 (+)-PTZ	2	22.5 ± 3.1	471.5 ± 61.5	14 wk/10 wk diabetic
Analysis of R1 Gene and Protein Expression in the Ins2Akita/+Mouse
Wild-type	8	21.8 ± 2.2	159.0 ± 39.5	8 wk/nondiabetic
Wild-type (0.5 mg · kg-1 (+)-PTZ)	6	23.8 ± 1.3	168.5 ± 33.2	8 wk/nondiabetic
Ins2Akita/+	10	21.3 ± 0.9	461.5 ± 57.4	8 wk/4 wk diabetic
Ins2Akita/+ (0.5 mg · kg-1 (+)-PTZ)	7	22.3 ± 0.9	414.0 ± 35.8	8 wk/4 wk diabetic
Analysis of Effects of 0.5 mg kg-1 (+)-PTZ in the Ins2Akita/+Mouse
Wild-type (no (+)-PTZ)	6	30.2 ± 1.7	104.5 ± 16.3	17-25 wk/nondiabetic
	6	25.2 ± 0.9	153.1 ± 25.3	10 wk/nondiabetic
	4	22.5 ± 1.5	160.5 ± 29.0	7 wk/nondiabetic
Wild-type (0.5 mg · kg-1 (+)-PTZ)	9	31.9 ± 0.9	119.8 ± 41.2	25 wk/nondiabetic
Ins2Akita/+ (no (+)-PTZ)	9	27.6 ± 2.1	597.3 ± 6.0	17-25 wk/13-21 wk diabetic
	3	23.2 ± 0.4	584.5 ± 21.9	10 wk/6 wk diabetic
	4	21.4 ± 1.1	553.7 ± 44.6	7 wk/3 wk diabetic
Ins2Akita/+ (0.5 mg · kg-1 (+)-PTZ)	8	25.9 ± 1.6	515.8 ± 44.7	25 wk/21 wk diabetic
Histological Analysis of Fixed Plastic-Embedded Retinal Specimens
Wild-type	3	27.9 ± 2.5	169.0 ± 22.5	12 wk/nondiabetic
Ins2Akita/+	3	27.1 ± 1.2	596.5 ± 11.4	12 wk/8 wk diabetic
Ins2Akita/+(0.5 mg · kg-1(+)-PTZ)	3	24.7 ± 0.8	571.0 ± 16.4	12 wk/8 wk diabetic

(+)-PTZ, (+)-pentazocine.

Administration of (+)-Pentazocine

A total of 101 mice were used in these experiments: 12 were used in the pilot study assessing (+)-pentazocine dosage in the streptozotocininduced diabetic model, and 89 were used to assess (+)-pentazocine in the Ins2^Akita model. Mice were distributed over four groups: diabetic plus (+)-pentazocine treatment, diabetic with no treatment, nondiabetic plus (+)-pentazocine treatment, and nondiabetic with no treatment (Table 1). At the time of diabetes onset (or comparable age in nondiabetics), the mice were injected intraperitoneally with (+)-pentazocine (Sigma-Aldrich, St. Louis, MO) that had been dissolved initially in DMSO and diluted with 0.01 M PBS. For the streptozotocin-induced diabetic mice, (+)-pentazocine was administered via intraperitoneal injection at two dosages: 0.5 mg · kg^-1 or 1.0 mg · kg^-1 three times per week for 10 weeks. Based on the outcome of these pilot studies, Ins2^Akita/+ and wild-type mice received intraperitoneal injection of (+)-pentazocine (0.5 mg · kg^-1) two times per week for 22 weeks.

RT-PCR and Immunoblot Analysis to Detect σR1

The neural retina was dissected from the remainder of the eye cup of Ins2^Akita/+and wild-type mice. Three to four eyes per group per experiment were pooled for analysis. For gene expression, total RNA was isolated (TRIzol reagent; Invitrogen, Carlsbad, CA). RT-PCR was performed with primer pairs specific for mouse σR1⁶: 5′-CTC GCT GTC TGA GTA CGT G-3′ (sense) and 5′-AAG AAA GTG TCG GCT GCT AGT GCA A-3′ (antisense) (nucleotide position 315-333 and 572-593; expected PCR product size, 279 bp). 18S RNA was the internal standard. RT-PCR was performed at 35 cycles, with a denaturing phase of 1 minute at 94°C, an annealing phase of 1 minute at 59°C, and an extension of 2 minutes at 72°C. A 20-μL portion of the PCR products were gel electrophoresed and stained with ethidium bromide. For protein identification, we performed immunoblot analysis to detect σR1 protein in neural retinas according to our published method.⁸ Protein samples were subjected to SDS-PAGE and transferred to nitrocellulose membranes, which were incubated with anti-σR1 antibody (1:500) followed by incubation with HRP-conjugated goat anti-rabbit IgG antibody (1:3000). The proteins were visualized by using a chemiluminescence Western blot detection system. The membranes were reprobed with mouse monoclonal anti-β-actin antibody (1:5000) as a loading control. The films were analyzed densitometrically, as described.⁸

Tissue Processing and Morphometric Analysis

The mice were killed by CO₂ asphyxiation followed by cervical dislocation. The eyes were removed immediately and oriented in OCT (Tissue-Tek; Miles Laboratories, Elkhart, IN), such that the plane of section included the temporal and nasal ora serrata. Left and right eyes were frozen slowly by immersion in liquid nitrogen (without fixation); 10-μm-thick cryosections were prepared as described.³⁶ Cryosections were stained with hematoxylin and eosin (H&E) and used for systematic morphometric analysis with a fluorescence microscope (Axioplan2 equipped with HRM camera and the Axiovision 4.6.3 program; Carl Zeiss Meditec, Dublin, CA). Measurements included the thickness of the total retina, IPL, INL, outer plexiform layer (OPL), and the photoreceptor inner/outer segments. The number of rows of photoreceptor cell nuclei in the outer nuclear layer (ONL) was counted. The number of cell bodies in the GCL was quantified by counting cells from the temporal ora serrata to the nasal ora serrata and expressing the data as number of cells per 100-μm length of retina. Sections analyzed were from the center of the eyeball, thus the optic nerve was visible. Three measurements were made on each side (temporal and nasal) of the optic nerve at 200- to 300-μm intervals resulting in six measurements obtained per eye. Both eyes were analyzed in each mouse. In addition, the eyes of nine mice (three wild-type, three Ins2^Akita/+, and three (+)-pentazocine-treated Ins2^Akita/+, age: 12 weeks) were enucleated and fixed for 1 hour at room temperature in 2% paraformaldehyde and 2% glutaraldehyde in 0.1 M cacodylate buffer in sucrose and postfixed for 1 hour with osmium tetroxide. Processing and embedding of tissue in resin (EMbed-812; Electron Microscopy Sciences, Hatfield, PA) was performed according to our published protocol.³⁷

TUNEL Assay and Immunodetection of Vimentin, 4-Hydroxynonenal, and Nitrotyrosine

Retinal cryosections were subjected to a TUNEL assay (ApopTAG Fluorescein In Situ Apoptosis Detection Kit; Chemicon, Temecula, CA), according to the manufacturer’s protocol. They were counterstained with DAPI dye, to label all nuclei. For immunohistochemical detection of vimentin, 4-hydroxynonenal, and nitrotyrosine, retinal cryosections were fixed for 10 minutes in 4% paraformaldehyde prepared in PBS, washed thrice in PBS for 5 minutes, and blocked for 1 hour (PowerBlock; Biogenx, San Ramon, CA). All procedures were performed at 4°C. Sections were incubated in a humidified chamber overnight at 4°C with either mouse anti-human vimentin monoclonal antibody (1:25; Chemicon), rabbit polyclonal anti-4-hydroxy-2-nonenal antibody (1:3000; Alpha Diagnostic, San Antonio, TX) or rabbit antinitrotyrosine polyclonal antibody (1:250; Chemicon). In control experiments, some sections were incubated with normal donkey serum instead of the primary antibody. For detection of the primary antibodies, the sections were washed three times in PBS for 5 minutes and incubated with either Oregon Green Alexa Fluor 488-conjugated goat anti-mouse IgG (1:1000) to detect vimentin, or Alexa Fluor 555-conjugated donkey anti-rabbit IgG (1:1000) to detect 4-hydroxynonenal and nitrotyrosine (Invitrogen, Carlsbad, CA). Tissues were viewed by epifluorescence with the fluorescence microscope equipped as described earlier and with filters to detect FITC and DAPI. TUNELpositive (green fluorescing) cells in the GCL, INL, and ONL were counted along the full length of the retina from the nasal to the temporal ora serrata. In each experimental group, retinas of five mice were analyzed.

Statistical Analysis

Analysis of variance was used to determine whether there were significant differences in morphologic measurements and in the number of TUNEL-positive cells in diabetic versus nondiabetic, age-matched control mice, as well as in diabetic or age-matched, nondiabetic mice with no treatment versus diabetic mice treated with (+)-pentazocine. The Tukey paired comparison test was the post hoc statistical test. Data were analyzed with commercial software (SPSS, ver. 14; SPSS, Chicago, IL). P < 0.05 was considered significant.

RESULTS

σR1 Gene and Protein Expression in Retinas of Ins2^Akita/+ Mice

Before testing neuroprotective effects of (+)-pentazocine, σR1 mRNA and protein expression were analyzed in Ins2^Akita/+ mouse retinas (Fig. 1). The expression levels were comparable to wild-type, suggesting that σR1 is an available therapeutic target in the diabetic retina. These findings were similar to our previously published data showing that σR1 is expressed in neural retina of the streptozotocin-induced diabetic mouse retina.²⁰

Figure 1.

σR1 mRNA and protein levels. (A) Total RNA was isolated from neural retinas of wild-type and Ins2^Akita/+ mice (age, 8 weeks; 4 weeks diabetic), which had been maintained in the absence or presence of 0.5 mg · kg^-1 (+)-pentazocine (two times/wk IP; 4 weeks). RNA was subjected to semiquantitative RT-PCR with primer pairs specific for mouse σR1 mRNA (279 bp). 18S RNA (489 bp) was the internal control. RT-PCR products were run on a gel and stained with ethidium bromide. DNA standards (100 bp to 12.2 kbp) were run in parallel. (B) Band intensity, quantified by densitometry, of σR1 cDNA relative to 18S cDNA in control and Ins2^Akita/+ retinas. (C) Proteins were extracted from retinas of Ins2^Akita/+ mice (age, 8 weeks; 4 weeks diabetic) and agematched wild-type mice, which had been maintained in the absence or presence of 0.5 mg · kg^-1 (+)-pentazocine for 4 weeks. Proteins were subjected to SDS-PAGE, followed by immunoblot analysis with an affinity-purified antibody against σR1 (M_r ∼ 27 kDa) or β-actin (M_r ∼ 50 kDa) (internal loading control). (D) Band density, quantified by densitometry, indicated that σR1 protein levels were similar between wild-type and Ins2^Akita/+ mice, and the levels were not altered by a 4-week treatment with (+)-pentazocine. The ratio of the σR1/β-actin bands of the wild-type (+/+) mice was assigned an arbitrary unit of 1.

Effects of Pentazocine on Neuronal Death in Diabetic Retinopathy

The dosage of (+)-pentazocine used in the study with Ins2^Akita/+ was selected based on data from the pilot study of streptozotocin-induced diabetic mice. Histologic sections of retinas from streptozotocin-induced diabetic mice are shown in comparison to control and (+)-pentazocine-treated mice (Fig. 2). The diabetic mice had fewer cells in the GCL, but the retinal architecture was similar to that of the control. Streptozotocin-induced diabetic mice were administered various dosages of (+)-pentazocine (Table 1). At the 0.5-mg · kg^-1 dosage, (+)-pentazocine reduced cell loss in the GCL (Fig. 2B) and reduced apoptotic neuron death (Fig. 2C). Thus, this dosage was used in the comprehensive analysis of the effects of (+)pentazocine on retinopathy in the Ins2^Akita/+ mice.

Figure 2.

Effects of (+)-pentazocine on retinas of STZ-induced diabetic mice. (A) Photomicrographs of H&E-stained retinas of control (C57BL/6) mice, STZ-induced diabetic mice that had been diabetic 10 weeks, or STZinduced diabetic mice treated with 0.5 mg · kg^-1 (+)-pentazocine over the 10-week period. Note the well-organized morphology of the control and the (+)-pentazocine-treated retinas with dense distribution of ganglion cells compared to the untreated diabetic retina showing fewer cell bodies in the GCL. ipl, inner plexiform layer; inl, inner nuclear layer; onl, outer nuclear layer. (B) Number of cell bodies in the GCL of retinal cross-sections from control (nondiabetic) mice, untreated diabetic mice (UN), and diabetic mice treated with 0.5 mg · kg^-1 (+)-pentazocine (PTZ). (C) Number of TUNEL-positive cells in the GCL. *Significantly different from control and PTZ-treated diabetic mice (P < 0.01). Error bars, SE.

Ins2^Akita/+ mice were injected intraperitoneally at diabetes onset (∼4-weeks of age) with 0.5 mg · kg^-1 (+)-pentazocine twice weekly for 22 weeks. The progression of changes in Ins2^Akita/+ retinas compared with wild-type over this time period is shown in Figure 3. Wild-type mice had uniform thickness of layers throughout the central and midperipheral retina (Fig. 3A). Ins2^Akita/+ mice had modest INL thinning at 7 weeks (Fig. 3B) and more dramatic INL cell dropout at 10 weeks (Fig. 3C). By 17 to 25 weeks, there was marked INL and GCL cell loss in Ins2^Akita/+ mice (Figs. 3D-F). The IPL, which is composed of synaptic processes of cells in the INL and GCL, was also thinner. The cell loss and misalignment of inner retinal layers resulted in a somewhat wavy appearance in some of the retinas of 17- to 25-week-old Ins2^Akita/+ mice. We found that (+)-pentazocine treatment of Ins2^Akita/+ mice led to marked preservation of retinal architecture. The data shown (Figs. 3G-I) are from retinas of three different (+)-pentazocinetreated Ins2^Akita/+ mice, representative of the excellent retinal structure observed in the eyes of all diabetic mice treated with (+)-pentazocine (n = 8 mice, 16 eyes). Morphometric analysis indicated a significant decrease in the thickness of Ins2^Akita/+ mouse retinas, whereas (+)-pentazocine-treated Ins2^Akita/+ mice were comparable to wild-type mice (Fig. 3J). The IPL and INL in Ins2^Akita/+ mice measured 30.3 ± 6.4 and 19.68 ± 2.72 μm, respectively. In (+)-pentazocine-treated Ins2^Akita/+ mice, the values for the thicknesses of the IPL and INL (51.2 ± 4.9 and 31.3 ± 1.3 μm, respectively) were comparable to those in wild-type mice (51.1 ± 4.6 and 31.9 ± 2.4 μm, respectively; Figs. 3K, 3L). There were 30% fewer cell bodies in the GCL of Ins2^Akita/+ mice compared with wild-type mice (10.4 ± 1.2 vs. 15.4 ± 1.2 cells/100 μm retinal length, respectively) whereas the values for (+)-pentazocine-treated Ins2^Akita/+ mice (15.6 ± 1.5 cells/100 μm) were similar to those in wild-type (Fig. 3M). Of note, (+)-pentazocine-treated Ins2^Akita/+ mice remained hyperglycemic throughout treatment. Blood glucose levels were ∼500 mg/dL (similar to untreated Ins2^Akita/+ mice) and were significantly higher than those in wild-type mice (104 -160 mg/dL; Table 1), suggesting that hyperglycemia per se may not be sufficient to trigger neuronal loss in diabetes.

Figure 3.

Preservation of retinal structure in Ins2^Akita/+ mice administered (+)-pentazocine. Representative H&E-stained retinal cryosections of (A) wild-type mice: GCL cells were distributed evenly, and nuclear layers were uniformly thick. (B-F) Ins2^Akita/+ mice: INL became disrupted with age, and GCL density decreased. (G-I) Ins2^Akita/+mice treated with (+)-pentazocine (0.5 mg/kg, two times/wk IP; 22 weeks) showed marked preservation of retinal layers. PTZ, pentazocine; gcl, ganglion cell layer; ipl, inner plexiform layer; inl, inner nuclear layer; onl, outer nuclear layer. Retinal sections were subjected to morphometric analysis: (J) total retinal thickness, (K) IPL thickness, (L) INL thickness, and (M) number of cell bodies in GCL per 100-μm length of retina. Data are the mean ± SE of measurements from retinas of six wild-type (12 eyes), nine Ins2^Akita/+ (18 eyes), and eight Ins2^Akita/+ treated with (+)-pentazocine (16 eyes). *Significantly different from wild-type and (+)-pentazocinetreated mice (P < 0.001). Scale bar, 50 μm.

Assessment of Neuronal Apoptosis and Cellular Stress

Neuronal apoptosis has been reported in Ins2^Akita/+ mouse retinas.³³ We determined the number of TUNEL-positive cells in retinas of (+)-pentazocine-treated versus untreated Ins2^Akita/+ mice (Figs. 4A-C). There were significantly more TUNEL-positive cells in the GCL and INL in retinas of untreated mice than in those of treated and wild-type mice (Figs. 4D, 4E), demonstrating robust neuroprotective effects of (+)-pentazocine in vivo. To assess the effects of diabetes on Müller cells,we visualized the radial fibers by immunolabeling of the intermediate filament protein vimentin, a glial cell marker. Ins2^Akita/+ mouse retinas displayed disrupted, punctate labeling of fibers (Fig. 5B) whereas (+)-pentazocine-treated Ins2^Akita/+ mouse retinas had well-organized fibers comparable to wild-type retinas (Figs. 5A, 5C). Given the major role of Müller cells in maintaining retinal structural integrity, the disruption of fibers may contribute to the progressive damage observed in diabetic retinas; treatment of Ins2^Akita/+ mice with (+)-pentazocine preserved the radial organization. Reactive oxygen and nitrogen species have been implicated in the pathogenesis of diabetic retinopathy.^38-40 We analyzed the levels of 4-hydroxynonenal and protein-nitrotyrosine, markers of lipid peroxidation and reactive nitrogen species, respectively, in retinas of Ins2^Akita/+ mice compared with wild-type and (+)-pentazocine-treated Ins2^Akita/+ mice. There was increased immunodetection of these markers in Ins2^Akita/+ mice (Figs. 5E, 5H). Treatment with (+)-pentazocine (Figs. 5F, 5I) decreased the levels and were comparable to wild-type (Figs. 5D, 5G), suggesting that (+)-pentazocine reduces oxidative and nitrosative stress in diabetic retina.

Figure 4.

Neuronal apoptosis decreased in Ins2^Akita/+ mice administered (+)-pentazocine. TUNEL analysis of cryosections of retinas from (A) wild-type (25 weeks), (B) Ins2^Akita/+ (17-25 weeks after onset of diabetes), and (C) Ins2^Akita/+ (+)-pentazocine-treated mice (25 weeks). Green and blue fluorescent signals reflect TUNEL⁺ and DAPI staining of nuclei, respectively. Arrows: numerous TUNEL⁺ cells in Ins2^Akita/+ mice (B). Quantitation of TUNEL⁺ cells in GCL (D) and INL (E). *Significantly different from wild-type and (+)-pentazocine-treated mice (P < 0.001).

Figure 5.

Oxidative stress was reduced in Ins2^Akita/+ mice administered (+)-pentazocine. (A-C) Immunodetection of vimentin labeling (green fluorescence) of radial glial fibers (arrows). Vimentin labeling of Ins2^Akita/+ retinas (B) was disrupted compared with that in wild-type (A) and (+)-pentazocine-treated Ins2^Akita/+ mice (B). (D-I) Immunodetection of 4-hydroxynonenal (D-F) and nitrotyrosine (G-I), oxidative stress and nitrative stress markers (bright red fluorescence, arrows), respectively, in Ins2^Akita/+ were increased compared with wild-type and (+)-pentazocine-treated Ins2^Akita/+ mice. PTZ, pentazocine; gcl, ganglion cell layer; ipl, inner plexiform layer; inl, inner nuclear layer; onl; outer nuclear layer, is; inner segments of photoreceptor cells. Scale bar, 50 μm.

The cell loss quantified in the present study (Fig. 3) is consistent with that reported for Ins2^Akita/+ mice by Barber et al.³³ In that study, the eyes were fixed in 4% paraformaldehyde and were subjected to cryosection/tape-transfer methodology. The morphology of the retinas of wild-type and pentazocinetreated Ins2^Akita/+ mice in the present study was well-preserved; however, cryoembedding, especially without fixation, can have untoward effects on retinal morphology, particularly if the retina is degenerating. Cryoembedding retinal tissue retains antigenicity and thus is advantageous for detection of markers of apoptosis and cellular stress (as needed for data presented in Figs. 4, 5); however, it does not preserve retinal architecture structure as well as other histologic processing methods. Hence retinas of nine additional mice were fixed in a formaldehyde-glutaraldehyde solution and processed for embedding in resin (Embed 812; Electron Microscopy Sciences). Figure 6 shows representative images from these specimens. The retinas of wild-type mice were well organized, with the ganglion cells being evenly distributed and the INL of a uniform thickness (∼6 -7 rows, Fig. 6A). Cells in both of these layers had lightly stained cytoplasm, and most had a plump, round shape. The retinas of the pentazocine-treated Ins2^Akita/+ mice had excellent morphology, an appearance similar to wild-type (Fig. 6D). The retinas of the Ins2^Akita/+ mice showed somewhat improved morphology than was observed with cryosectioning (compare with Fig. 3); however, disruption in the INL-OPL interface continued to be evident (Fig. 6B, arrow). Cells of the GCL had a variable appearance. Some were plump and round adjacent to others that were shrunken and darkly stained, indicative of cells that are dying. The INL of the Ins2^Akita/+ mice had approximately four to five rows of cells. There were noticeable gaps between many of the cells that were not observed in the wild-type mice or in the Ins2^Akita/+ mice treated with (+)-pentazocine. These data of fixed retinal specimens are consistent with data obtained from the unfixed specimens and reflect cellular loss in the retinas of the Ins2^Akita/+ mice, which can be prevented by administration of (+)-pentazocine.

Figure 6.

Eyes harvested from wildtype, Ins2^Akita/+ and (+)-pentazocinetreated Ins2^Akita/+ mice (age 12 weeks) fixed in paraformaldehyde/glutaraldehyde and embedded in resin. Representative retinas are shown from (A) wild-type, (B, C) Ins2^Akita/+, and (D) (+)-pentazocine-treated Ins2^Akita/+ mice. PTZ, pentazocine, gcl, ganglion cell layer, ipl, inner plexiform layer, inl, inner nuclear layer; onl, outer nuclear layer. Scale bar, 25 μm.

DISCUSSION

This study represents the first report of the neuroprotective effects of the σR1 ligand (+)-pentazocine in an in vivo model of spontaneously arising retinopathy. The model we chose for our detailed analysis of the effects of (+)-pentazocine, the Ins2^Akita/+ mouse, was an ideal animal in which to test neuroprotection because of the quantifiable loss of retinal neurons over a relatively short time (4-5 months).³³ The observation that the σR1 gene and protein continued to be expressed in the mutant mice indicated that the appropriate target for the (+)-pentazocine treatment was available. (+)-Pentazocine is widely regarded as a potent and highly specific σR1 ligand based on pharmacologic studies.⁴¹ It binds σR1 with a very high affinity and is presumed to be an agonist for the receptor; however, because the actual signaling events associated with σR1 activation have not been characterized, it is difficult to ascribe definitively an agonist or antagonist function to this ligand. It is for that reason that in the present study we did not test other compounds (i.e., presumed antagonists of σR1), to demonstrate specificity of the effect. The most convincing molecular evidence that (+)-pentazocine is specific for σR1 comes from analysis of the σR1 knockout mice (mσR1^-/- mice) developed in the Montoliu laboratory.⁴² Northern blot and Western blot analysis of multiple organs known to express σR1 (liver, kidney, brain, and heart) revealed no specific mRNAs for the mouse σR1 gene or protein in any of these organs in the homozygous (knockout mice), whereas expression was reduced by half in heterozygous compared with wild-type mice. When brain membranes from the mσR1^-/- mice were analyzed in binding assays with [³H](+)pentazocine used as the radioligand, no binding activity in brain membranes was observed, and binding activity was reduced by half in heterozygous compared with wild-type animals (as shown in Fig. 3 of Ref. 42). Therefore, we conclude that the effects of (+)-pentazocine observed in the present study are due to effects on σR1.

The biweekly exposure of the Ins2^Akita/+ mice to (+)-pentazocine led to marked preservation of retinal structure. The morphology of the treated mice was excellent. The number of cells in the GCL in the (+)-pentazocine-treated Ins2^Akita/+ mice was comparable to the number in the wildtype. The inner nuclear and plexiform layers were of a thickness comparable to those in the wild-type.

Of interest, the preservation of retinal architecture and reduction of neuronal apoptosis did not appear to be a direct function of decreased hyperglycemia, because the (+)-pentazocine-injected Ins2^Akita/+ mice demonstrate marked elevation of blood glucose through the time of death. This finding is important and raises the question of whether retinal neuronal loss associated with diabetes involves hyperglycemia directly or whether it is due to complications secondary to hyperglycemia, such as oxidative stress. Our analysis of markers of cellular stress (4-hydroxynonenal, nitrotyrosine) showed a marked increase in expression in the Ins2^Akita/+ retinas, but a sharp quenching of these stress indicators in the presence of (+)-pentazocine. These findings are interesting in light of reports that ligands for σR1 can reduce NO production^13,43,44 and oxidative stress.⁴⁵ It is noteworthy that increased σR1 binding activity was observed in vitro when retinal Müller cells were treated with NO donors and donors of reactive oxygen species.²¹

The variable appearance of the mutant Ins2^Akita/+ retinas was intriguing and is reminiscent of the histologic features of retinas reported in studies of patients with diabetes^29,38 in which ganglion cell loss is extensive and INL disruption and cell loss are substantial. The alteration in retinal organization suggested that the scaffolding of the retina may be compromised. The cells known to be involved in such organization are the Müller cells and disruptions of Müller glial cells have been reported in diabetic retinopathy.^39,40 Our immunolabeling studies of vimentin, a Müller cell marker, showed punctate labeling in the Ins2^Akita/+ mouse retinas, but well-organized radial fibers in the (+)-pentazocine-treated animals. It would be interesting in future studies to isolate Müller cells from the Ins2^Akita/+and assess their myriad functions (e.g., transport activities), in the presence and absence of (+)-pentazocine.

In summary, we found that the σR1 ligand (+)-pentazocine afforded robust protection against retinal degeneration observed in Ins2^Akita/+ mice, a mouse model of spontaneously arising diabetes. The cell loss, neuronal apoptosis, marked disruption of retinal layers, and increased oxidative stress were prevented by treatment with (+)-pentazocine. The results are encouraging, because ligands for σR1 are already used clinically in the treatment of other diseases, including neuropsychiatric diseases (schizophrenia, depression, and cognitive disorders).⁴⁶ Promising results from the Bucolo laboratory in work involving acute injury-induced models in rat retina and new forms of σR1 ligands suggest that σR1 is a very attractive target for therapeutic intervention of retinal disease.⁴⁷ It remains to be determined whether other models of spontaneously arising nonacute retinal degeneration—for example, models of glaucoma—will respond positively to (+)-pentazocine. In addition, it is not known whether the treatment with (+)-pentazocine must start at the onset of disease, as performed in the present study, or whether delayed treatment would prove similarly beneficial. Currently, in vitro studies are under way to determine how (+)-pentazocine regulates genes involved in retinal neuronal apoptosis and will set the stage for elucidating the mechanism of in vivo neuroprotection observed in the present work.