Transfer of the Experimental Autoimmune Glaucoma Model from Rats to Mice—New Options to Study Glaucoma Disease

Studies have suggested an involvement of the immune system in glaucoma. Hence, a rat experimental autoimmune glaucoma model (EAG) was developed to investigate the role of the immune response. Here, we transferred this model into mice. Either 0.8 mg/mL of the optic nerve antigen homogenate (ONA; ONA 0.8) or 1.0 mg/mL ONA (ONA 1.0) were injected in 129/Sv mice. Controls received sodium chloride. Before and 6 weeks after immunization, the intraocular pressure (IOP) was measured. At 6 weeks, retinal neurons, glia cells, and synapses were analyzed via immunohistology and quantitative real-time PCR (RT-qPCR). Additionally, optic nerves were examined. The IOP stayed in the normal physiological range throughout the study (p > 0.05). A significant reduction of retinal ganglion cells (RGCs) was noted in both immunized groups (p < 0.001). Remodeling of glutamatergic and GABAergic synapses was seen in ONA 1.0 retinas. Furthermore, both ONA groups revealed optic nerve degeneration and macrogliosis (all: p < 0.001). An increase of activated microglia was noted in ONA retinas and optic nerves (p < 0.05). Both ONA concentrations led to RGC loss and optic nerve degeneration. Therefore, the EAG model was successfully transferred from rats to mice. In further studies, transgenic knockout mice can be used to investigate the pathomechanisms of glaucoma more precisely.


Introduction
Glaucoma is a progressive neuropathy with changes in the optic nerve head, gradual retinal ganglion cell (RGC) death, and visual field loss [1]. Although an elevated intraocular pressure (IOP) is the main risk factor, IOP-unrelated pathomechanisms also occur. Since this disease is multifactorial, appropriate models mimicking possible pathological pathways are needed. In the past few years, a rat experimental autoimmune glaucoma model (EAG) was used to identify mechanisms related to immunological alterations in IOP-independent glaucoma. Here, rats were immunized with ocular antigens, such as heat shock proteins (HSPs) or an optic nerve antigen homogenate (ONA). This led to a loss of RGCs and optic nerve degeneration [2][3][4]. Additionally, an enhanced activation of glia cells and complement system proteins could be observed [4][5][6][7][8]. Furthermore, a remodeling of extracellular

Activated Microglia in the Retina
In the retina, the whole population of microglia was labeled with anti-Iba1 and activated microglia were visualized by combining this marker with anti-F4/80. Astrocytes were labeled with anti-GFAP ( Figure 4A). Additionally, RT-qPCR analyses were performed for Iba1, Cd68, and Gfap ( Figure 4C,E,G).

Optic Nerve Degeneration
Longitudinal optic nerve sections were stained with H&E to evaluate the degree of cellular infiltration and with luxol fast blue (LFB) to analyze the extent of demyelination. Additionally, optic nerves were labeled with anti-SMI-32 to detect possible changes in the neurofilament ( Figure 5A).  The evaluation of the H&E staining showed no alterations in the ONA 0.8 optic nerves (mean score: 0.98 ± 0.09; p = 0.2). However, significantly more cellular infiltrations were observed in the ONA 1.0 group (1.63 ± 0.12; p < 0.001) in comparison to controls (0.67 ± 0.18; Figure 5B).

Glia Activation in the Optic Nerves
The whole amount of microglia was labeled with anti-Iba1 and activated microglia were additionally marked with anti-F4/80. Macroglia in the optic nerves were visualized with anti-GFAP ( Figure 6A). The evaluation of the H&E staining showed no alterations in the ONA 0.8 optic nerves (mean score: 0.98 ± 0.09; p = 0.2). However, significantly more cellular infiltrations were observed in the ONA 1.0 group (1.63 ± 0.12; p < 0.001) in comparison to controls (0.67 ± 0.18; Figure 5B).

Glaucomatous Damage in EAG Mice
The European Glaucoma Society defines glaucoma as a chronic, progressive neuropathy with morphological changes at the optic nerve head and retinal nerve fiber layer, associated with RGC death and visual field loss [1]. We therefore analyzed both tissues, retina and optic nerve, in this study to confirm glaucomatous damage in mice. Six weeks after immunization, a loss of RGCs was noted in both ONA groups. Additionally, optic nerve degeneration was observed. In rats, studies showed that ONA treatment leads to RGC death, starting 22 days after immunization [3,4,6]. To ensure degenerative effects in our mice, investigations were carried out 6 weeks following immunization. We conclude that the transfer of the EAG model from rats to mice was successful. In future, the 129/Sv mouse strain can be used for knockout studies in combination with the autoimmune glaucoma model. This could help to test hypotheses about glaucoma pathologies affected by genetic alterations.
To investigate the mechanisms occurring during glaucoma more precisely, we additionally analyzed a possible remodeling of synapses as well as glia cell alterations.

Synaptic Alterations after Immunization
Previously, it was noted that a disruption of the axonal transport in RGCs might represent an early event in glaucoma disease [10]. In the rat EAG model, we could detect a loss of the presynaptic active zone protein bassoon and the postsynaptic protein PSD-95 [11]. These alterations were found 4 weeks after immunizing with S100B in combination with HSP27. In the study presented here, mice immunized with ONA did not display any alterations regarding PSD-95 expression 6 weeks after immunization. Since PSD-95 labels post-synapses of the photoreceptors, it is possible that these are not affected by ONA immunization at this point in time. However, remodeling of GABAergic and glutamatergic synapses was noted. In Morbus Alzheimer, synapse and dendritic spine loss appear in proximity to amyloid beta plaques [12,13]. For example, an increased amyloid pathology in brains with Morbus Alzheimer is correlated with a diminished Vglut1 protein localization, a glutamatergic transporter [14]. A decrease of Vglut1 could also be observed in retinas of an animal model where the pathological effects of apolipoprotein 4 were investigated [15]. A downregulation of Slc17a7 mRNA levels was noted in the higher concentrated ONA group in our study, assuming that it might also play a crucial role in glaucoma neurodegeneration.
Gephyrin is a postsynaptic anchor protein, which tethers glycine and GABA A receptors to the cytoskeleton [16], and glycine subtypes are present in the brain and spinal cord [17,18]. Gephyrin immunoreactivity was found in co-localization with amyloid plaques in post-mortem tissue of Morbus Alzheimer patients [17,19]. Besides, gephyrin dysfunctions are also linked to other neurological diseases, like stiff-person syndrome, hyperekplexia, molybdenum cofactor deficiency, schizophrenia, and autism [20][21][22][23]. In the retina, clusters of glycine and GABA A receptors are expressed by RGCs [24]. In our study, we could demonstrate a downregulation of gephyrin in the ONA 1.0 group. Possibly, the loss of RGCs is accompanied with a lower synaptic immunoreactivity of gephyrin.

Reactive Gliosis in the Optic Nerves
It is known that in response to an injury in the central nervous system, astrocytes become reactive [25,26]. In the retina, Müller cells are specialized glia cells contacting retinal neuron somata and processes, providing stability to the neural tissue [27,28]. As in astrocytes, Müller cell proliferation is increased in pathologic eye conditions, such as retinal detachment, ischemia, diabetic retinopathy, or glaucoma [27,[29][30][31]. However, until now, it remains unclear whether gliosis is neurodegenerative or neurodestructive [32]. The evaluation of GFAP, expressed by astrocytes and activated Müller cells [29,33], revealed no changes in the retina in our study. In EAG rats on the other hand, 4 weeks after immunization with ONA, enhanced GFAP levels were observed via Western blot in the retina but not in immunohistological analyses [4]. However, a similar number of astrocytes was also noted in a mouse OHT study two weeks after lasering [34]. The authors postulated that in the OHT eyes, a reactive, non-proliferative gliotic response occurred. This was also reported in other mouse and rat studies relating to glaucoma [35][36][37]. A non-proliferative response seems to be the consequence of a slow degeneration, while rapid damages lead to macroglia proliferation [35,37,38]. Furthermore, it is known that astrocytes exhibit multiple phases of remodeling during neurodegeneration [39]. In contrast to the retina, a strong increase of GFAP was noted in the ONA optic nerves in our study. Studies demonstrated that astrocytes in the optic nerve head respond strongly after glaucomatous damage [40]. Furthermore, in humans with a chronic elevated IOP and a moderate or advanced glaucomatous axonal damage, an increased immunoreactivity of GFAP is observable [41,42]. It remains unknown, whether an astrogliosis proceeds to promote detrimental effects on neurons or whether they have a neuroprotective role [43]. It is assumed that in the initial phase of the disease, reactive astroglia have a beneficial role, while with disease progression, these astroglia become neurodestructive [40,44]. The present paradigm could be useful to explore the pharmacological profile of novel anti-glaucoma molecules with a potential protection effect on retina along with an effect on IOP such as sigma receptor ligands [45,46].
Microglia are the resident macrophages of the retina and play an important role in the defense mechanisms of the immune system. Under normal conditions, microglia monitor and remove cellular detritus and maintain cellular homeostasis [47][48][49]. In glaucoma, several studies noted an increase in microglia numbers after induced hypertension or optic nerve damage [50][51][52]. In early stages of the EAG glaucoma rat model, a strong microglia response could be noted at 14 days [4]. Additionally, a decrease of this activation was described between 3 and 12 weeks after optic nerve transection [50]. In the study presented here, we could detect more activated microglia in both retinas and optic nerves. Our results indicate a contribution of the microglia in the EAG model.

Animals
All procedures concerning animals adhered to the ARVO statement for the use of animals in ophthalmic and vision research. All experiments involving animals were approved by the animal care committee of North Rhine-Westphalia, Germany (approval code: 84-02.04.2013.A291; July 2013). 129/Sv (129S2/SvPasCrl) mice were kept under environmentally controlled conditions with free access to chow and water.

Immunization
The preparation and immunization of ONA was carried out as previously described [3,53]. 129/Sv mice received an intraperitoneal injection with either 0.8 mg/mL (ONA 0.8) or 1 mg/mL ONA (ONA 1.0). The antigen was mixed with incomplete Freund's adjuvant (50 µL; Sigma-Aldrich, St. Louis, MO, USA). The animals of the control group were injected with NaCl in Freund's adjuvant. Additionally, all mice received 1 µg pertussis toxin (Sigma-Aldrich) intraperitoneally on days 0 and 2 [54].

Measurement of Intraocular Pressure
IOP of both eyes in all animals was measured before and 6 weeks after immunization using a rebound tonometer (TonoLab, Icare, Vantaa, Finland) as described previously (n = 3-5/group) [4,7]. For this procedure, mice were anesthetized with a ketamine (Ratiopharm, Ulm, Germany)/xylazine (Bayer healthcare, Berlin, Germany) cocktail (120/16 mg/kg). All measurements were performed by one examiner at the same time of the day. For each analysis, ten measurements per eye were calculated, and the average of the both eyes was used.

Retina and Optic Nerve Histology
After 6 weeks, retinas and optic nerves were fixed in 4% paraformaldehyde for 1 (retina) or 2 h (optic nerves), dehydrated in sucrose, and embedded in Tissue Tek (Thermo Fisher, Waltham, CA, USA). Cross-sections of the retina (10 µm) and longitudinal optic nerve sections (4 µm) were cut with a Cryostat (Thermo Fisher) and mounted on Superfrost slides (Thermo Fisher).

Immunohistology
In order to identify different cell types, specific immunofluorescence antibodies were applied (n = 6-7/group; Table 1) [5]. Briefly, retinal cross-sections and longitudinal optic nerve sections were blocked with a solution containing 10-20% donkey and/or goat serum and 0.1% or 0.2% Triton-X in PBS. Primary antibodies were incubated at room temperature overnight. Incubation using corresponding secondary antibodies was performed the next day for 1 h. Nuclear staining with 4',6 diamidino-2-phenylindole (DAPI, Serva Electrophoresis, Heidelberg, Germany) was included to facilitate the orientation on the slides. Negative controls were performed by using only secondary antibodies.

Histological Examination
All photographs were taken with a fluorescence microscope (Axio Imager M1 or M2, Zeiss, Oberkochen, Germany). Two photos of the peripheral and two of the central part of each section were captured. The images were transferred to Corel Paint Shop Pro (V13, Corel Corporation, Ottawa, ON, Canada), and equal excerpts were cut out [11]. Afterwards, RGCs, bipolar cells, microglia, and photoreceptor cells were counted using ImageJ software (National Institute of Health, Bethesda, MD, USA). Data were transferred to Statistica software (V13, Dell, Round Rock, TX, USA) for further analysis. GFAP, synapses, and rhodopsin were evaluated through area analyses using an ImageJ macro [11,55]. Briefly, images were transformed into grayscale. To minimize interference with background labeling, a rolling ball radius was subtracted (Table 2). Then, for each picture, a suitable lower and upper threshold was set. The ideal threshold was obtained when the grayscale picture and the original one corresponded. Afterwards, the mean value of the lower threshold was calculated, and this number was used for the final analysis. The percentage of the labeled area was measured between these defined thresholds ( Table 2). Data were transferred to Statistica software for further analysis.

Histopathological Staining and Scoring
Retinal cross-sections were stained with hematoxylin and eosin (H&E, Merck, Burlington, MA, USA) and cresyl violet (Sigma-Aldrich) to be able to detect any signs of inflammation or changes in retinal structure [56]. To evaluate the extent of cellular infiltration, longitudinal cryo-sections of optic nerves were stained with H&E. The degree of demyelination was examined via LFB (RAL Diagnostics, Martillac Cedex, France) [57]. After staining, ethanol was used for dehydration of the sections, followed by incubation in xylene (Merck) and coating with Eukitt (VWR, Langenfeld, Germany).
Three images of each optic nerve section (anterior, medial, and posterior) were taken with an Axio Imager M1 microscope at a 400x magnification (six sections per animal).
To examine the extent of inflammatory cell infiltration, an established score was used [58,59]: 0 = no infiltration, 1 = mild cellular infiltration, 2 = moderate infiltration, 3 = severe infiltration, and 4 = massive infiltration with formation of cellular conglomerates. Regarding the degree of demyelination, LFB-stained sections were assessed as previously described [59]: 0 = no demyelination, 0.5 = small holes, 1 = moderate demyelination, 1.5 = bigger holes, and 2 = severe demyelination up to complete loss of structural integrity. Data were transferred to Statistica software for further analysis.

Quantitative Real-Time PCR
Both retinas of each animal (5 animals/group) were pooled for RNA preparation and cDNA synthesis as previously described [5,60]. The designed oligonucleotides for the quantitative real-time-PCR (RT-qPCR) are shown in Table 3. ß-actin and Cyclophilin (Ppid) served as reference genes for retinal analysis. The RT-qPCR was performed using DyNAmo Flash SYBR Green (Thermo Scientific) on the PikoReal RT-qPCR Cycler (Thermo Scientific) [61,62]. Values were transferred to REST© software (Qiagen, Hilden, Germany) for further analysis.

Statistics
Immunohistological data are presented as mean ± SEM. Cell counts and evaluated area fractions were compared by ANOVA followed by Dunnet's post-hoc. Here, controls were set to 100%. Regarding RT-qPCR, data are presented as median ± quartile + minimum/maximum and were assessed using REST© software [63]. p-values below 0.05 were considered statistically significant.

Conclusions
In the current study, the rat autoimmune glaucoma could successfully be transferred to mice. This offers many more possibilities to investigate the pathomechanisms occurring in glaucoma in future knockout studies. Furthermore, this study provides novel insights into synaptic degeneration in the autoimmune glaucoma model.