Molecular Vision 2019; 25:890-901
Received 30 June 2018 | Accepted 30 December 2019 | Published 31 December 2019
1Cole Eye Institute, Cleveland Clinic, Cleveland, OH; 2Department of Biochemistry & Molecular Genetics, University of Louisville, Louisville, KY; 3Research Service, Louis Stokes Cleveland VA Medical Center, Cleveland, OH; 4Department of Ophthalmology, Cleveland Clinic Lerner College of Medicine of Case Western Reserve University, Cleveland, OH
Correspondence to: Junzo Kinoshita, 1-16-13, Kitakasai, Edogawa, Tokyo 134-8630, Japan, Phone: +81-3-3680-0151; FAX: +81-3-5696-8335; email: firstname.lastname@example.org
Purpose: The Grm6nob8 mouse carries a missense mutation in the Grm6 gene (p.Met66Leu), and exhibits a reduced b-wave of the electroretinogram (ERG), abnormal localization of metabotropic glutamate receptor 6 (mGluR6) to the depolarizing bipolar cell (DBC) soma, and a reduced level of mGluR6 at the DBC dendritic tips. Although the underlying mechanism remains unknown, one possible explanation is that DBCs cannot efficiently traffic the mutant mGluR6. In that scenario, reducing the total amount of mutant mGluR6 protein might normalize localization, and thus, improve the ERG phenotype as well. The second purpose of this study was to determine whether the abnormal cellular distribution of mutant mGluR6 in Grm6nob8 retinas might induce late onset DBC degeneration.
Methods: We crossed Grm6nob8 animals with Grm6nob3 mice, which carry a null mutation in Grm6, to generate Grm6nob3/nob8 compound heterozygotes. We used western blotting to measure the total mGluR6 content, and immunohistochemistry to document mGluR6 localization within DBCs. In addition, we examined outer retinal function with ERG and retinal architecture in vivo with spectral domain optical coherence tomography (SD-OCT).
Results: The retinal content of mGluR6 was reduced in the retinas of the Grm6nob3/nob8 compound heterozygotes compared to the Grm6nob8 homozygotes. The cellular distribution of mGluR6 in the Grm6nob3/nob8 compound heterozygotes matched that of the Grm6nob8 homozygotes, with extensive expression throughout the DBC cell body and limited expression at the DBC dendritic tips. The dark-adapted ERG b-waves of the Grm6nob3/nob8 mice were reduced in comparison to those of the Grm6nob8 homozygotes at postnatal day 21 and 28. The overall ERG waveforms obtained from 4- through 68-week old Grm6nob8 mice were in general agreement for dark- and light-adapted conditions. The maximum response and sensitivity of the dark-adapted ERG b-wave did not change statistically significantly with age. SD-OCT revealed the maintained laminar structure of the retina, including a clear inner nuclear layer (INL) at each age examined (from 11 to 57 weeks old), although the INL in the mice older than 39 weeks of age was somewhat thinner than that seen at 11 weeks.
Conclusions: Mislocalization of mutant mGluR6 is not normalized by reducing the total mGluR6. Mislocalized mutant mGluR6 does not trigger substantial loss of DBCs.
The signal transduction cascade of depolarizing bipolar cells (DBCs) is initiated when a light-induced reduction in glutamate release from photoreceptor terminals is sensed by the metabotropic glutamate receptor 6 (mGluR6) [1,2]. DBC activity underlies the b-wave component of the electroretinogram (ERG) [3-7], and the b-wave is missing in humans carrying mutations in GRM6 (OMIM 604096) [8,9], and in mouse models lacking mGluR6 due to null mutations in Grm6 [10-13].
We recently reported the Grm6nob8 mutant mouse model in which the ERG b-wave is reduced but is not eliminated . The mouse carries a missense mutation, p.Met66Leu. Immunohistochemistry documented abnormal distribution of mGluR6, with a substantial fraction retained within the DBC soma and abnormally low levels of mGluR6 at the DBC dendritic tips. In the Grm6nob8 model, mutant mGluR6 was abnormally glycosylated, a state that has been shown to result in abnormal trafficking of other G-protein coupled receptors [15-17].
In the present report, we follow up on the observation in the mouse model mentioned above in two respects. To determine whether reducing levels of mutant mGluR6 might normalize localization within Grm6nob8 DBCs, we genetically reduced the amount of mGluR6 by crossing the Grm6nob8 mutant with a Grm6 null model (Grm6nob3). Although we are unaware of a comparable approach, we note that mislocalized mutant proteins are associated with neuronal degeneration in other systems [e.g., 18,19]. Given the relevance of Grm6 mutant mice to the human condition complete congenital stationary night blindness (cCSNB)  and the identification of human mutations near the nob8 locus , we examined the retinal structure and function in Grm6nob8 mice at ages up to 1 year old.
Mice were obtained from the Jackson Laboratory (Bar Harbor, ME). The lines used were CBA/CaJ that are homozygous for the Grm6nob8 allele (stock #000654; hereafter Grm6nob8), C57BL/6J (stock #000664), and B6.B10(D2)-Grm6nob3/BOC mice that are homozygous for a null mutation in Grm6 ; hereafter, Grm6nob3. In addition to these homozygous lines, we crossed Grm6nob8 and Grm6nob3 mice to generate Grm6nob3/nob8 compound heterozygotes. Mice were housed conventionally, in microisolator cages with free access to food and water. All procedures involving live animals were approved by the Cleveland Clinic Institutional Animal Care and Use Committee, and were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.
Mouse retinas were isolated and homogenized in lysis buffer (1% Nonidet P40, 2 mM EDTA, and 20 mM HEPES, pH 7.4, supplemented with protease inhibitor cocktail (P8340, Sigma-Aldrich, St. Louis, MO), and lysed further by rotating at 4 °C for 45 min. Homogenates were cleared by centrifugation at 17,000 ×g for 20 min at 4 °C. Protein samples were separated with sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes, and immunoblotting was performed as described previously . Protein bands were visualized by scanning the membranes in an Odyssey Infrared Imaging System (LI-COR, Lincoln, NE) using 700 nm and 800 nm channels. Band densities were quantified using Image Studio Version 5.0 software (LI-COR).
At 4 weeks, the mice were euthanized with carbon dioxide inhalation, and the eyes were enucleated. Dissected retinas were fixed for 20 min by immersion in 4% (w/v) paraformaldehyde in 0.1 M phosphate buffer (PB) with a pH 7.4, then washed in PB, cryoprotected through a graded sucrose series (5%,10%, 15%, and 20% in PB), and frozen in OCT (Sakura Finetek, Torrance, CA):20% sucrose (2:1) . A cryostat was used to cut 18-μm sections, which were then mounted on Super-Frost glass slides (Thermo Fisher Scientific, Waltham, MA), air-dried, and stored at −80 °C.
After acclimating to room temperature, the sections were washed for 5 min in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH 7.4), then another 5 min in PBX (PBS containing 0.5% (v/v) Triton X-100), and then incubated for 1 h in blocking solution (PBX containing 5% (v/v) normal donkey serum). The retinal sections were incubated in primary antibodies diluted in blocking solution, at room temperature overnight. The antibody against mouse mGluR6 amino acids 853–871 (KKTSTMAAPPKSENSEDAK) was generated in sheep . Anti-PKCα antibody was used at 1:1,000 dilution (Sigma, P4334). After incubation with the primary antibody, the sections were washed with three 10-min cycles of PBX and one 10-min cycle of PBS. The sections were then incubated at room temperature for 1 h with fluorescently labeled secondary antibodies (1:1,000 in blocking solution). The secondary antibodies were Alexa Fluor 488 donkey anti-sheep and Alexa Fluor 546 donkey anti-rabbit (Invitrogen, Carlsbad, CA). The slides were then washed three times in PBX and coverslipped with Vectashield Mounting Media (Vector Laboratories, Burlingame, CA). The sections were imaged on a confocal microscope (Olympus FV1000, Center Valley, PA), using a 60X oil objective (1.45 NA). Images shown here are maximum projections of confocal stacks, after adjustment for contrast and brightness with Fluoview software.
Mice were dark-adapted overnight, anesthetized (ketamine: 80 mg/kg; xylazine: 16 mg/kg), and administered eye drops for pupil dilation (1% tropicamide; 2.5% phenylephrine HCl; 1% cyclopentolate HCl) and to anesthetize the corneal surface (1% proparacaine HCl). Needle electrodes were placed in the cheek (reference) and tail (ground). ERGs were recorded using a stainless steel electrode wetted with 1% carboxymethylcellulose.
Strobe flash stimuli were presented to the dark-adapted eye, and after a minimum of 5 min of light adaptation using background illumination at 20 cd/m2. Stimuli ranged from −3.6 to 2.1 log cd s/m2. Responses were amplified (0.03–1,000 Hz), averaged, and stored using an LKC (Gaithersburg, MD) UTAS E-3000 signal averaging system.
The amplitude of the dark-adapted ERG a-wave was measured from the prestimulus baseline to 8 ms after the flash onset. The amplitude of the b-wave was measured from the a-wave trough to the positive peak, or from the baseline to the positive peak for stimulus conditions that did not evoke an a-wave. The amplitude of the light-adapted ERG b-wave was measured from the initial negative trough to 40 ms after the flash onset.
The stimulus-response function of the dark-adapted b-wave was fitted with a hyperbolic equation of the following form :
In this equation, R is the b-wave amplitude, Rmax is the maximum b-wave amplitude, L is the flash luminance (log cd s/m2), n is a dimensionless slope parameter, and K is the flash energy that elicits an amplitude of half Rmax (half-saturation coefficient). This model yielded values for two parameters (i.e., Rmax and K). Grm6nob3/nob8 and Grm6nob8 mice were tested at postnatal day (P) 15, P21, and P28. The Grm6nob8 mice were also tested at 9, 28, 48, and 68 weeks of age.
Procedures for animal preparation, imaging, and recovery have been published [25,26]. Briefly, a Bioptigen (Durham, NC) model SDOIS SD-OCT system was used for spectral domain optical coherence tomography (SD-OCT) imaging. B-scans were made through the optic nerve head at 1,000 A-scans/B-scan × 30 frames. B-scan images had dimensions of 0.4 mm (depth) by approximately 1.6 mm (width), and were exported to ImageJ, coregistered, and averaged using StackReg/TurboReg/Plugins . Using SD-OCT images taken across the horizontal and vertical meridians from each mouse, the thickness of the inner nuclear layer (INL) was measured at a point midway between the optic nerve head and the edge of the image. The Grm6nob8 mice were imaged at three ages (11, 39, and 57 weeks old).
The Grm6nob8 retina is characterized by a selective reduction in the amplitude of the ERG b-wave and mislocalization of mGluR6 to the cell body of DBCs, as opposed to its normal restriction to the DBC dendritic tips . To determine whether either of these abnormalities might be ameliorated by reducing the total amount of mutant mGluR6 protein, we compared these features in Grm6nob8 homozygotes and Grm6nob3/nob8 compound heterozygotes. These mice differ with respect to the total content of mGluR6, with retinas of Grm6nob3/nob8 compound heterozygotes having approximately half compared to Grm6nob8 homozygotes (Figure 1A,B).
The mGluR6 expression was observed only in the DBC dendritic tips of the control mice (Figure 2A,C), with no overlap with PKCα labeling in the DBC soma (Figure 2B,C). In the Grm6nob8 homozygotes, mGluR6 expression was reduced (Figure 1D) and colocalized with PKCα in DBC soma (Figure 2E,F). The distribution of mGluR6 in the Grm6nob3/nob8 compound heterozygotes (Figure 2G–I) matched that of the Grm6nob8 homozygotes with extensive expression throughout the cell body and limited expression at the DBC dendritic tips. This result indicates that the mislocalization of mGluR6 was not corrected when only a single Grm6nob8 allele was expressed, resulting in a reduction in the total amount of mutant mGluR6 (Figure 1A,B).
When the Grm6nob3/nob8 compound heterozygotes were examined with ERG, the results complemented those obtained with immunohistochemistry. In Figure 3A, the ERG phenotypes of representative Grm6nob8 and Grm6nob3/nob8 mice are compared at P28, along with a reference control mouse at the same age. The dark-adapted a-waves in the Grm6nob3/nob8 mice showed amplitudes and kinetics that were comparable to those of the Grm6nob8 mice (Figure 3A,B; p>0.05). In contrast, the dark-adapted b-waves of the Grm6nob3/nob8 mice were statistically significantly smaller than those in the Grm6nob8 mice (Figure 3A,B; p<0.01), consistent with these mice expressing a single Grm6nob8 allele. The small b-wave and oscillatory potentials present in the dark-adapted ERGs of the Grm6nob3/nob8 mice are, however, clearly larger than those reported for the Grm6nob3 homozygotes [12,13]. There was no statistically significant difference in the amplitude of the light-adapted b-waves obtained from the Grm6nob8 and Grm6nob3/nob8 mice (Figure 3C,D; p>0.05).
We also used ERG to examine mice at younger ages. Figure 4 presents representative ERGs obtained from Grm6nob8 homozygotes (black) and Grm6nob3/nob8 compound heterozygotes (red) at P15 and P21. At P21, a clear albeit reduced b-wave was present in all animals, with a greater reduction observed in the Grm6nob3/nob8 compound heterozygotes compared to the Grm6nob8 homozygotes. We did not detect a b-wave in either the Grm6nob3/nob8 compound heterozygotes or the Grm6nob8 homozygotes at P15, an age when the retina is still developing, and the ERG has not taken on an adult configuration .
To examine whether the Grm6nob8 phenotype might change with age, we examined mice at different ages. The overall waveforms of the ERGs obtained from 4- and 68-week old Grm6nob8 mice were in good general agreement for dark- and light-adapted conditions (Figure 5A,C). Figures 5B,D present summary luminance-response functions for ERG components obtained from mice at 4, 9, 28, 48, and 68 weeks of age. These data are not indicative of a progressive age-related change in the ERG amplitude. We further analyzed the dark-adapted b-wave luminance-response functions using the Naka-Rushton equation, which derives two parameters: the maximum response (Rmax) and the sensitivity (K). The Rmax values appeared to decrease slightly with age, which is in good agreement with a previous report showing a slight decrease in the ERG amplitude with age in wild-type mice . When the values seen at the youngest age (4 weeks old) were compared to the later time points, the difference was not statistically significant (Figure 5E; p>0.05). The sensitivity parameter K trended upward with age (Figure 5F). When the K values of the 4-week-old mice were compared to those obtained from the two oldest ages, there was no statistically significant difference (p>0.05). Together, these results indicate that retinal dysfunction in Grm6nob8 mice is not progressive.
In Figure 6A, the SD-OCT images taken across the horizontal meridian from Grm6nob8 mice at 11, 39, and 57 weeks of age are compared. At each age examined, a clear INL, housing bipolar cell nuclei, was seen in the Grm6nob8 mice. Although the INL thickness in the mice at 39 weeks or older was thinner than that at 11 weeks old (Figure 6B), the laminar structure of the retina was maintained until the oldest age was tested.
Unlike other Grm6 mouse models, where the retina lacks mGluR6 protein due to a null mutation [10-13], mGluR6 is detected in the Grm6nob8 retina . The distribution of the mutant protein differs from that of the wild type. First, mGluR6 is clearly seen in the DBC soma. Second, the amount of protein that is successfully trafficked to the DBC dendritic tips is reduced. In agreement with mislocalized mGluR6, the ERG b-wave is reduced in Grm6nob8 mice, and retinal ganglion cell function is also impaired . The mislocalization of mutant mGluR6 could be related to abnormal protein folding and trafficking mechanisms, secondary to the abnormal glycosylation of mGluR6 noted in the Grm6nob8 retina .
We hypothesized that reducing the total amount of mutant mGluR6 protein might improve the amount of mGluR6 that is successfully delivered to the DBC dendritic tips. Previous studies demonstrated that a single copy of WT Grm6 is sufficient to support normal DBC function , and therapeutic strategies (such as hammerhead ribozymes) exist to reduce gene expression in the retina [31-33]. This has relevance for patients with cCSNB, as some are known to carry mutations that lie near the nob8 locus , although we do not yet know if they also cause mislocalization. To test this hypothesis, we generated Grm6nob3/nob8 compound heterozygotes, which express a single Grm6nob8 allele; Grm6nob3 is a null mutation . In contrast to the possibility that mGluR6 protein and/or DBC function might be improved in the Grm6nob3/nob8 retina compared to the Grm6nob8 retina, we observed neither normalization of mGluR6 localization nor an improved ERG b-wave. Instead, the b-wave was reduced in compound heterozygotes. These results indicate that simply reducing mutant mGluR6 protein will not provide a therapeutic avenue, and instead, suggest that some other means of improving mGluR6 trafficking to the DBC dendritic tips is needed.
A large number of mouse mutants have been reported where the ERG b-wave is reduced or abolished . In general, the b-wave is abolished in mice with mutations in members of the DBC signaling cascade, including Nyx [35,36], Trpm1 [30,37-39], Gpr179 , Lrit3 [41-43], Gnb3 , Gnb5 , Gna01 , and Grm6 [10-12], or in mice that lack DBCs . In comparison, the b-wave is reduced but not abolished in mice with mutations in proteins that are involved in multiple other systems, including the control of glutamate release from photoreceptor terminals (Cacna1f [48-50], Cacna2d4 [51,52], and Cabp4 ), the development of retinal vasculature (Fzd4 , norrin [55,56], and Lrp5 ), or the dystroglycan complex (pikachurin , POMGnT1 , and Large [60-62]). Based on the premise that the b-wave is abolished in mice with mutations of the DBC signaling process, we anticipated that the nob8 mice would involve one of the processes that reduces but does not abolish DBC function , and were surprised when the mutation in Grm6 was identified . The Grm6nob8 mouse is the first mouse mutant to retain a b-wave in the presence of a mutation in a component of the DBC signaling cascade, and raises the possibility that similar variation may occur in human patients with CSNB, who are currently classified into complete and incomplete forms based in part on the absence of the b-wave in complete CSNB and on the presence of a reduced but not abolished b-wave in patients with incomplete CSNB . Human patients expressing mutations in the glutamate ligand binding domain of GRM6, near the Met66 residue, have been reported . It would be of interest to determine if they retained a measurable ERG b-wave compared to patients with CSNB with other Grm6 mutations.
Protein-misfolding diseases, such as Alzheimer disease and Huntington's disease, involve the accumulation of misfolded proteins with abnormal conformation in neurons . Furthermore, it has been reported that patients with complete CSNB harboring a GRM6 mutation showed reduced inner retinal thickness with OCT imaging . This motivated us to investigate whether long-term accumulation of mutant mGluR6 protein in DBC cell bodies might impact the inner retina of Grm6nob8 mice. ERG and SD-OCT analyses of Grm6nob8 mice at multiple ages indicate that rod DBC function is maintained at a reduced but stable level up to 68 weeks of age, and that a clear laminar structure of the retina is maintained up to 57 weeks old, with only a slight thinning of the INL. The absence of major age-related changes in the ERG b-wave or in the overall retinal structure indicated that DBCs tolerate the presence of mislocalized mutant p.Met66Leu mGluR6.
This work was supported by the Foundation Fighting Blindness, and grants from the NIH (R01EY012354; R01EY019943; P30EY025585) and Research to Prevent Blindness (to Cleveland Clinic Cole Eye Institute and University of Louisville Department of Ophthalmology & Visual Sciences). NSP is a VA Research Career Scientist.