Adult Long-Evans (LE) rats regardless of sex used in this study were provided by the Animal Centre of the Third Military Medical University (Army Medical University). Rats were maintained in the Animal Center of Southwest Hospital under a regular 12 h light/dark cycle. LE rats were worn animal masks and anesthetized by inhalation mixed O${}_2$ and isoflurane. Oxybuprocaine hydrochloride eye drops were used for ocular surface anesthesia. 20 $\mu$L of 30 mg/mL magnetic microspheres suspension was injected into the anterior chamber from the corneal limbus with a 33 G needle. After injection, magnetic microspheres were distributed at the angle of the anterior chamber by a small magnet to block the trabecular meshwork [23].

IOP of all rats was measured in both eyes 1 day before and 1 day after microspheres injection, then weekly until 8 weeks post-injection (wpi) in a dark room from 9:00 AM to 10:00 AM. After rats were anesthetized by inhalation successfully, the Icare tonometer was used to quickly measure IOP in both eyes of glaucoma rats three times, and the average value was taken.

OECs were isolated and purified from nerve layers of olfactory bulbs from adult LE rats as described previously [16]. After being anesthetized deeply with pentobarbital sodium (10 mg/kg; Sigma-Aldrich, St. Louis, MO, USA). The nerve layer was isolated from the olfactory bulb, cut into 0.5 mm${}^3$ small pieces, and digested for 15 min with 0.1% trypsin (Gibco, Grand Island, NY, USA) under 37 °C and 5% CO${}_2$. Primary OECs were cultured in DMEM/F12 with 10% Fetal Bovine Serum (FBS) and 1% PS for 5 days and then changed culture medium every 3 days. When OECs reached 80%–90% confluence, the Nash differential adhesion method is used to purify OECs. Then, OECs that were purified 2~3 times were identified and used for subretinal transplantation and co-culture experiment.

OECs were digested into single-cell suspension in Phosphate-Buffered Saline (PBS) for transplantation, as previously described [15]. Rats were anesthetized by intramuscular injection of sodium pentobarbital (10 mg/kg). Oxybuprocaine hydrochloride eye drops were used for ocular surface anesthesia, and compound tropicamide was used for mydriasis. The bulbar conjunctiva of the corneoscleral limbus was cut with microscissors, and the bulbar conjunctiva and fascia were bluntly separated to expose the sclera. Anterior chamber puncture was performed at the limbus to flow out aqueous humor. The sclera was punctured at 2 mm behind the corneoscleral limbus; 10${}^5$ cells/3 $\mu$L of cell suspension was injected into subretinal space by a 33-gauge Hamilton needle. The right eyes of glaucoma rats were injected into OECs and the left eyes were injected into 3 $\mu$L PBS.

Rats were dark-adapted for 2 hours and then placed on the experimental platform for 1 min before the experiment [24]. The optokinetic test device consisted of upper and lower mirrors, four surrounding computer screens, and a platform located in the center. A camera suspended on the top of the test device was used for recording the behaviors of rats during the test. During the test, a total of 8 bars with different spatial frequency (0.05, 0.075, 0.1, 0.2, 0.3, 0.4, 0.5, and 0.6 cycles/degree from wide to narrow under 100% contrast were applied for detecting the visual acuity. Statistical analysis was performed after the highest spatial frequency with which rats could respond was converted into the corresponding visual acuity.

Pattern electroretinogram (PERG) was recorded as described previously [25]. Rats were dark-adapted over 8 hours and preparations were performed under dim red light before PERG recording. Rats were anesthetized by intramuscular injection of sodium pentobarbital (10 mg/kg) and placed on a heating plate to maintain the body temperature of rats at 37 °C to prevent hypothermia. Briefly, the ring recording electrode made of gold was put on the center of the cornea, the ground electrode was inserted subcutaneously in the tail, and the reference electrode was inserted into the scleral conjunctiva around the equator of the eyes. The generation, amplification of stimulus (0.1–500 Hz band-pass), and data acquisition were performed with a RETIscan system (Roland Consult, Brandenburg, Germany). White and black alternating reversing bars (contrast, 100%; mean luminance, 50 cd/m${}^2$; spatial frequency, 0.033 cycles per degree; temporal frequency, 1 Hz) were used to generate the stimulus. At a distance of 11 cm from the screen, the projection of the pupil and visual axis were aligned. The average of superimposed 250 sweeps was recorded as a PERG reading. Measurements were made between a peak and adjacent trough of the waveform for the PERG amplitudes.

For cultured cells, cell climbing slices were fixed in 4% paraformaldehyde (PFA) for 15 min. for retinal tissues, after anesthesia with an intramuscular injection of sodium pentobarbital (10 mg/kg), the rats were perfused with 0.9% normal saline through the left ventricle continuously. After the blood was washed out, rats were perfused with 4% PFA solution for internal fixation. Cut the bulbar conjunctiva around the corneal limbus, separated fascia and muscle, and the optic nerve was cut near the orbital apex. The cornea, iris, lens, and vitreous body were cut off, and only the 0.5 mm sclera and optic nerve close to the optic disc were kept. The optic nerve was placed in 4% PFA, fixed at 4 °C for 0.5 hours, gradient dehydrated in 10%, 20%, and 30% sucrose/PB solution, and then put into 30% sucrose/PB solution overnight. Then, the sucrose solution was exhausted, and the optic nerve was embedded in the optimum cutting temperature compound and quickly frozen in liquid nitrogen. A cold microtome was used to cut the tissues into 10-$\mu$m-thick sections.

Frozen sections were washed three times in PBS, permeabilized with 0.3% Triton X-100 for 10 min, and blocked by 1% goat serum for 1 hour at 37 °C. Frozen sections were incubated with primary antibody at 4 °C overnight. After washing in PBS three times, sections were incubated in secondary antibody at 37 °C for 30 min. After staining of DAPI, the sections were scanned and photographed by confocal microscopy (Zeiss LSM 880, Oberkochen, Germany) The antibodies and dilutions were used in present study: mouse anti-Tuj1 (ab78078 1:200, Abcam, Cambridge, UK), Rabbit anti-GFAP (ab7260 1:200, Abcam, Cambridge, UK), Rabbit anti-EAAT2 (ab41621 1:200, Abcam, Cambridge, UK), Rabbit anti-GS (ab64613 1:100, Abcam, Cambridge, UK), mouse anti-CRALBP (ab15051 1:50, Abcam, Cambridge, UK), mouse anti-Caspase-3 (ab13847 1:100, Abcam, Cambridge, UK).

Muscle and connective tissue around the eyeball were isolated and removed to expose the optic nerve. The optic nerve was minced and digested in 0.25% trypsin at 37 °C for 40 min, and D/F12 containing 15% FBS and 1% PS was used to terminate digestion. The cell suspension was centrifuged for 5 min, seeded in T75, and cultured at 37 °C. Once cells reached confluence, T75 was shaken on a shaker at 250 rpm for 18 h, the supernatant was removed and fresh medium was added to T75.

P4 AC (astrocytes) was seeded on a cell climbing slice that was placed in a 24-well plate at 10${}^5$ cells/well. AC was treated with 50 mM, 100 mM, and 200 mM glutamate for 6 and 24 hours to observe the states of the cells.

The eyeball of LE rats at postnatal 8 days was taken out and incubated in dulbecco’s modified eagle medium (DMEM) without serum at 4 °C overnight. Then, the retina was separated from the eyeball that was digested in 1 mg/mL trypsin 1 mg/mL type IV collagenase at 37 °C for 40 min, then was mechanically dissociated into a single cell and cultured in DMEM containing 10% FBS for 7 days. Upon reaching complete confluence, Müller glia was trypsinized for purification and identification.

P4 AC was cultured in 24-well plates at a density of 1 $\times$ 10${}^5$ cells/well and was co-cultured with OECs or AC. OECs were seeded in the transwell chambers at a density of 1 $\times$ 10${}^5$/well in the experimental group and AC were seeded in the transwell chambers at the same density in the control group. Then, cells were cultured in a medium containing 200 mM glutamate for 24 hours for photographing and immunofluorescence staining.

Glutamate concentration was measured by a Glutamate detection kit (ab83389, Abcam, Cambridge, UK). Briefly, 0, 2, 4, 6, 8, and 10 $\mu$L glutamate solution (1 mM) were added to a 96-well plate and were made up to 50 $\mu$L to make a standard curve. 50 $\mu$L cell culture supernatant was added to a 96-well plate, then a mixture containing 90 $\mu$L buffer, 8 $\mu$L chromogenic reagent, and 2 $\mu$L glutamate enzyme was added to the 96-well plate. OD value measured by spectrophotometer at 450 nm wavelength after incubating for 30 min at 37 °C was used to Calculate the glutamate concentration according to the standard curve.

Total RNAs of cell and optic nerve tissue cells were extracted using RNAiso Plus (Cat. No. D9108A, Takara, Kyoto, Japan). RNA samples were reversed as cDNA by PrimeScriptTM RT reagent kit (Cat. No. #RR037A, Takara, Kyoto, Japan). RT-qPCR was used to measure the expression of various genes as previously described [13]. And the expressions were normalized to the level of GAPDH. All primers were listed in Table 1.

All data were analyzed statistically by using SPSS 13.0 (IBM Corp., Chicago, IL, USA). A T-test was used to compare the means of two groups and one-way or two-way analysis of variance (ANOVA) was used to compare means among multiple groups. Results are presented in the form of mean $\pm$ standard deviation ($\overline{x}$$\pm$ s). p values 0.05 were considered to be statistically significant.

We established the elevated IOP glaucoma model by injecting magnetic microspheres into the anterior chamber of LE rats. Magnetic microspheres were evenly distributed in the corner of the chamber under the attraction of the magnet, destroying the normal function of the trabecular meshwork (Fig. 1A). Before injection of microspheres, IOP was measured (10.75 $\pm$ 0.96 mmHg). Following injection, IOP raised continuously and reached the peak at 1 week post-injection (55.00 $\pm$ 1.41 mmHg), then fell over gradually and approached a normal level at 7 wpi (11.25 $\pm$ 1.89 mmHg) (Fig. 1B). Then, an optokinetic response test was used to evaluate the visual acuity of normal rats and elevated IOP rats from 2 wpi to 8 wpi (Fig. 1C). The visual acuity gradually declined over time and decreased to 0.30 $\pm$ 0.07 c/d at 8 wpi from 0.6 c/d in normal rats (Fig. 1D, p 0.01). Subsequently, the functions of retinal ganglion cells (RGCs) were further assessed by PERG. The amplitudes of P1 and N2 waves declined progressively. The amplitude of the P1 wave measured in normal rats was 11.42 $\pm$ 2.23 $\mu$v, declined to 1.38 $\pm$ 0.79 $\mu$v (p 0.01) and the N2 wave amplitude in normal rats was 14.61 $\pm$ 1.37 $\mu$v, declined to 1$\pm$ 0.71 $\mu$v (Fig. 1E–I, p 0.01). Therefore, the injection of magnetic microspheres establishes an elevated IOP glaucoma model in which visual function is damaged.

Fig. 1.

The establishment and characteristics of the glaucoma model. (A) The distribution of magnetic microspheres in glaucoma model rats. (B) Line chart showing the changes of IOP after injection in glaucoma model rats (n = 5). (C) Diagram of optokinetic response test. (D) Visual acuity through optokinetic response test from 2 weeks to 8 weeks post-injection (n = 5). (E–G) Representative PERG traces in P-wave at 2, 4, and 8 weeks post-injection of magnetic microspheres. (H,I) Diagram showing P1 and N2 wave amplitude in normal and glaucoma model rats (n = 4). NR, normal rat; GR, glaucoma rat. *p 0.05, **p 0.01. IOP, elevated intraocular pressure; PERG, Pattern electroretinogram.

During intraocular pressure elevation, the structure of neuronal axons stained by Tuj1 gradually was disordered, and the loss of axons was observed in some areas at 2 wpi, and only a small amount of axonal fibers remained at the 8 wpi (Fig. 2A). The gene expression of Neurofilament declined progressively in glaucoma (Fig. 2C, p 0.05). In addition, the changes of astrocytes in the optic nerve region were evaluated by the glial fibrillary acidic protein (GFAP) staining. The optic nerve in normal rats was wrapped by the clear textured GFAP. With the increase of IOP, the structure of GFAP in astrocytes was disorganized and the expression was upregulated, which is most obvious at 4 wpi (p 0.05), and the expression of GFAP decreased slightly at 8 wpi (Fig. 2B,D, p 0.01). These results demonstrate that the nerve fibers of RGCs are lost gradually, accompanied by the gliosis of astrocytes during the pathogenesis of glaucoma.

Fig. 2.

Neuronal and glial responses in glaucoma model rats. (A) Tuj1 staining showing the changes of optic nerve axon in glaucoma model rats. (B) GFAP staining showing the gliosis of astrocytes in glaucoma model rats. (C,D) RT-qPCR analysis showing the relative mRNA expression of Neurofilament and GFAP (n = 9). *p 0.05, **p 0.01. Scale bars: 50 $\mu$m.

OECs were transplanted into subretinal space 2weeks after magnetic microspheres injection to rescue the elevated IOP-induced optic nerve injury and PBS was used for control. IOP in OECs and PBS groups declined progressively over time and there was no statistical difference in IOP between the two groups from 1 week post-transplantation (wpt) to 7 wpt (Fig. 3A, p $>$ 0.05). The visual acuity of glaucoma model rats increased after OECs transplantation compared PBS group significantly from 2 wpt to 8 wpt (Fig. 3B, 2 w: p 0.01; 4 w: p 0.05; 8 w: p 0.05). Besides, the amplitudes of P1-wave in the OECs group were higher than in the PBS group at 2 and 6 wpt (Fig. 3C,D, 2 w and 6 w: p 0.05) and this is also the case for N2-wave (Fig. 3C,E, 2 w: p 0.05; 6 w: p 0.01). Tuj1 and GFAP staining was used to further clarify the protective effect of OECs transplantation on the optic nerve of glaucoma model rats. The axon of the optic nerve in the PBS group is lost gradually from the onset of glaucoma and only a few Tuj1${}^{+}$ regions can be observed at the late stage. In the OECs group, the most of axon was integral and clear at 2 wpt and the area of the Tuj1${}^{+}$ region was larger than the PBS group at 6 wpt. Meanwhile, the expression of GFAP was decreased in the OECs group compared with the PBS group (Fig. 3F). Consistent with the results of IF, the mRNA level of Neurofilament increased in OECs from 2 wpt to 6 wpt (2 w: p 0.05; 4 w: p 0.05; 6 w: p 0.01) and GFAP mRNA level in OECs was lower than PBS group (4 w: p 0.05; 6 w: p 0.05) (Fig. 3G,H). Thus, these suggest that OECs transplantation can preserve visual functions of glaucoma model rats, protect the axon of the optic nerve and alleviate the gliosis of astrocytes.

Fig. 3.

OECs transplantation improved the visual functions in glaucoma models. (A) OECs transplantation decreased the IOP (n = 5). (B) Optokinetic response test showing OECs improved visual acuity from 2 weeks to 8 weeks post-transplantation (n = 5). (C) Representative PERG traces in P1-wave of the PBS and OECs groups at 2 and 6 weeks post-transplantation. (D,E) OECs improved the amplitudes of P1 and N2 wave in glaucoma model (n = 4). (F) Representative images of Tuj1 and GFAP staining in PBS and OECs groups. (G,H) RT-qPCR analysis showing the relative mRNA expression of Neurofilament and GFAP in PBS and OECs groups (n = 9). *p 0.05, **p 0.01. Scale bars: 50 $\mu$m. IOP, elevated intraocular pressure; PERG, Pattern electroretinogram; OECs, olfactory ensheathing cells; PBS, Phosphate-Buffered Saline.

Due to the key role of glutamate in the pathogenesis of glaucoma and the fact that AC are important regulators of glutamate metabolism, we studied the effects of OECs on astrocytes in vitro. Firstly, astrocytes isolated from the optic nerve of neonatal rats were identified by glutamine synthetase (GS) staining. P4 astrocytes with large and flat cell bodies had many thick processes, and the positive rate of GS was more than 95% (Fig. 4A). To mimic the glutamate microenvironment of the optic nerve in glaucoma, astrocytes were treated with different concentrations of glutamates. Astrocytes survived well and no apoptosis was observed under 100 mM for 24 h. Under 200 mM glutamate for 6 h, the process of astrocytes retracted, the cell body was atrophied, and apparent apoptosis was observed (Fig. 4B), suggesting that 200 mM glutamate could induce astrocytes apoptosis effectively. Next, RNA was extracted from astrocytes treated with 200 mM for 24 h to detect the changes in gene expression. Apoptosis-related genes Caspase-3 and Bax were upregulated, while apoptosis inhibitory gene Bcl-2 was downregulated (Fig. 4C, Caspase-3: p 0.01; Bcl-2 and Bax: p 0.05). Meanwhile, the gene expression of GFAP and glutamate metabolism related to GS and excitatory amino acid transporter 2 (EAAT2) decreased significantly after astrocytes were treated with 200 mM glutamate (Fig. 4D, GS and GFAP: p 0.01; EAAT2: p 0.05).

Fig. 4.

Establishment of glutamate-induced astrocytes apoptosis model in vitro. (A) The isolation and identification of primary astrocytes from the optic nerve of neonatal rats. (B) Glutamate induced the apoptosis of astrocytes. (C,D) Relative mRNA expression of GS, GFAP, EAAT2, Caspase-3, Bcl-2, and Bax in normal and glutamate-treated astrocytes for 24 h (n = 3). *p 0.05, **p 0.01. Scale bars: 50 $\mu$m.

Then, co-culturing was used to detect the capacity of OECs to metabolize glutamate and the effects on astrocytes in the glutamate-induced AC apoptosis model. In the OECs group, the process of astrocytes retracted and cell body size increased, suggesting the activation of astrocytes. However, the morphology of astrocytes was more integral and only a few apoptotic cells were observed compared control group (Fig. 5A). Co-culturing with OECs upregulated the expressions of GFAP, GS, and EAAT2 (Fig. 5D, GS: p 0.05; GFAP and EAAT2: p 0.01). The expression of the anti-apoptosis gene Bcl-2 increased and apoptotic-related genes Caspase-3 and Bax decreased in the OECs group (Fig. 5E, p 0.05). Meanwhile, we found that OECs decreased the glutamate concentration (Fig. 5F, p 0.01). Given that Müller glia was in charge of glutamate metabolism in the retina under normal and glaucoma conditions, we compared the capacity to metabolize glutamate of OECs, Müller glia, and astrocytes. Among three cells, the expression of genes associated with glutamate metabolism in OECs are highest, Müller glia was second and astrocytes were lowest (Fig. 5B,C). In summary, we established a glutamate-induced AC apoptosis model in which AC apoptosis and dysfunction of glutamate metabolism occurred. OECs contributed to reducing AC apoptosis by metabolizing excessive glutamate and maintained the role of glutamate metabolism in vitro.

Fig. 5.

Co-culturing with OECs alleviated the glutamate-induced astrocytes apoptosis and dysfunction. (A) The morphology and Caspase-3 staining of astrocytes co-cultured with/without OECs for 24 h. (B,C) Relative mRNA level of GS, EAAT1, EAAT2, EAAT3 in primary OECs, Müller glia, and astrocytes (n = 3). (D,E) Relative mRNA expression of GS, GFAP, EAAT2, Caspase-3, Bcl-2, and Bax in astrocytes co-cultured with/without OECs for 24 h (n = 3). (F) Co-culturing with OECs for 24 h decreased relative glutamate concentration in the medium (n = 3). *p 0.05, **p 0.01. Scale bars: 50 $\mu$m. OECs, olfactory ensheathing cells.

To explore the effect of OECs transplantation on glutamate metabolism in the optic nerve, we first clarified the expression of a gene associated with glutamate metabolism in the optic nerve and retina of normal rats. Although GFAP expression in the optic nerve was higher than in the retina, the level of GS and three glutamate transporters EAAT1, EAAT2, and EAAT3 in the optic nerve were lower than retina (Fig. 6A,B, GFAP, GS, EAAT1, and EAAT2: p 0.01; EAAT3: p 0.05), indicating that apoptosis of axons before cell body during glaucoma may be related to the ability of retina and optic nerve to regulate glutamate excitotoxicity. By immunostaining, GS and EAAT2 were reduced significantly from 4 wpi and little GS was observed in the optic nerve at 8 wpi (Fig. 6C,D). Consistent with the results of IF, the gene expressions of GS and EAAT2 declined from 4 wpi to 8 wpi compared to normal rats (Fig. 6E, 4 w: p 0.05; 8 w: p 0.01). Conversely, the relative glutamate concentration increased continuously from 2 wpi to 8 wpi (Fig. 6F, 2 w: p 0.05; 4 w: p 0.01; 8 w: p 0.01). These results show that the functions of AC for glutamate transport and metabolism are reduced and the glutamate content in the optic nerve increase with the progression of glaucoma, which may aggravate the damage of axons.

Fig. 6.

Grafted OECs improved the capacity for glutamate metabolism in the glaucoma model. (A,B) Relative gene expression of GS, EAAT1, EAAT2, EAAT3 in the retina and optic nerve of normal rats (n = 9). (C,D) Representative images of GS and EAAT2 staining in normal and glaucoma model rats at different timepoints. (E) Relative gene expression of GS, EAAT2 in optic nerve of normal and glaucoma model rats at different timepoints (n = 9). (F) Relative glutamate concentration in the optic nerve of normal and glaucoma model rats at different timepoints (n = 9). (G–I) Relative gene expression of GS, EAAT1, EAAT2 in optic nerve of PBS and OECs groups at different timepoints (n = 9). (J) Relative glutamate concentration in the optic nerve of PBS and OECs groups at different timepoints (n = 9). *p 0.05, **p 0.01. Scale bars: 50 $\mu$m. OECs, olfactory ensheathing cells; PBS, Phosphate-Buffered Saline.

Following OECs transplantation, the expressions of GS, EAAT1, and EAAT2 in the optic nerve of glaucoma rats increased significantly compared with the PBS group at 2 wpt (GS: p 0.05; EAAT1: p 0.01; EAAT2: p 0.05). The levels of these genes decreased but were higher than that in the PBS group at 4 wpt and 6 wpt (Fig. 6G–I, p 0.05). In addition, OECs reduced the relative glutamate concentration from 2 wpt to 6 wpt (Fig. 6J, 2 w and 4 w: p 0.01; 6 w: p 0.01). Therefore, OECs transplantation contributed to enhancing the functions of glutamate metabolism, improving the glutamate microenvironment in the optic nerve area of glaucoma model rats, reducing RGCs apoptosis caused by excitotoxicity