Clic6 Deficiency Triggers Aberrant Apical Microvilli in RPE

Abstract

Purpose: We sought to demonstrate the impact of Clic6 deficiency in the retina and RPE of C57BL/6N mice that are homozygous for the Crb1^rd8 mutation.

Methods: Six Clic6^−/− and six age-matched wild-type mice were used. Ophthalmic examination, ERG, and retinal optical coherence tomography were performed for clinical phenotyping. The mice were euthanized, and the eyes were processed for hematoxylin and eosin stain, immunohistochemistry, and transmission electron microscopy. Additionally, adult nonhuman primate and fetal human retinas were used for immunohistochemistry.

Results: Both Clic6^−/− and wild-type mice showed retinal lamination defects on optical coherence tomography and hematoxylin and eosin staining with a significantly greater number of lesions identified in Clic6^−/− mice. No significant differences in ERG parameters were identified between genotypes. The expression of ezrin/radixin/moesin seemed to be flat and blunted at the apical RPE surface in Clic6^−/− vs. wild-type mice. On transmission electron microscopy, Clic6^−/− mice showed broadened and stout RPE microvilli with a significantly decreased apical microvilli height and significantly increased width of each RPE microvillus. Although the height of the RPE cells and basal infoldings and pigment granule density did not differ significantly, phagosome density was significantly greater in Clic6^−/− mice. Both CLIC6 and ezrin/radixin/moesin were expressed in the RPE of adult nonhuman primate and fetal human tissues.

Conclusions: Clic6 deficiency induced aberrant RPE microvilli leading to a decreased intimacy between the RPE and photoreceptors and worsened the Crb1^rd8 mutation phenotype that was inherent in C57BL/6N mice. CLIC6 may have a potential role in RPE dysfunction across species, including humans.

The RPE is crucial for optimal retinal function and vision through a number of critical functions including formation of the outer blood-retinal barrier, transport of nutrients and waste products to and from the retina, providing mechanical support to photoreceptor (PR) cells and maintenance of PR via diurnal phagocytosis of shed PR outer segments.¹ In contrast with other epithelia where Na/K-ATPase is localized to the basolateral aspect of the cell, the RPE has a unique feature of polarity such that the Na/K-ATPase is expressed on the apical surface of the RPE cells where microvilli extend from and envelop the outer segments of the PRs.² 

Chloride intracellular channel (CLIC) 6 is in the six-member CLIC family of Cl⁻ channel proteins. Despite its name as an ion channel, CLIC family members exist in cytosolic soluble as well as membranous forms and can have nonchannel cellular functions. As soluble proteins in the cytoplasm, they localize to cell membranes and interact with the actin cytoskeleton, playing a role in diverse pathological conditions. For example, absence of Clic5 in mice with jitterbug (jbg) mutation compromises the integrity of the hair bundles in cochlear and vestibular hair cells by disrupting interactions with cytoskeletal filaments, resulting in hearing loss.³ CLICs are expressed ubiquitously in all cells and jitterbug mutant mice lacking Clic5 also develop abnormally broadened foot processes in podocytes of renal glomeruli and caused microalbuminuria.^4,5 In the heart, CLIC4 and CLIC5 present in mitochondria modulate reactive oxygen species and are associated with pulmonary hypertension and ischemia–reperfusion injury.^6,7 

A recent report demonstrated that RPE-specific Clic4 knockout mice exhibited drusen-like lipid deposits, a hallmark feature of AMD, via dysregulation of the autophagy pathway.⁸ In human retinas, CLIC6 expression was greatest among all ion channel proteins in the choroid–RPE region of the fovea in comparison to the periphery, highlighting the importance of CLIC6 in central vision.⁹ More specifically, both CLIC1 and CLIC6 were expressed in primary and human embryonic stem cell-derived RPE cells with CLIC6 being higher than CLIC1.¹⁰ However, the impact of CLIC6 in the RPE dysfunction has yet to be determined in vivo. Herein, we report the clinical, cellular, and ultrastructural characteristics of the RPE in Clic6-deficient mice. 

Materials and Methods

Animals and Human Tissues

All mice knockout lines generated by the International Mouse Phenotyping Consortium were on the C57BL/6N background, which is homozygous for the Crb1^rd8 mutation as described elsewhere.¹¹ A cohort of six Clic6^−/− mice and six age-matched wild-type mice at 14 to 16 weeks of age were used for this study. Furthermore, globes from the 18-year-old rhesus macaque (Macaca mulatta) were obtained from the California National Primate Research Center, which were scheduled to be humanely euthanized for reasons unrelated to ophthalmic diseases. The eyes from a 22-week-old human fetus were obtained for this study as well. 

This research complies with the tenets of the declaration of Helsinki. The use of discarded deidentified human fetal eyes was approved by the UC Davis Stem Cell Research Oversight Committee. The patients from whom the tissues were derived were informed and freely agreed for them to be used for research purposes, which was documented in the medical chart. Patients were not compensated. 

This study was also approved by the Institutional Animal Care and Use Committee of the University of California—Davis. All experiments were performed in compliance with the ARVO statement for the Use of Animals in Ophthalmic and Vision Research and the ARRIVE guidelines 2.0. 

Ophthalmic Examination

Complete ophthalmic examinations were performed at 14 weeks of age. Pupillary light reflexes were evaluated, and the ocular adnexa and anterior ocular segment were examined with a portable slit lamp (Kowa SL-15, Kowa, Tokyo, Japan) with magnification set at 16×. The pupil was then pharmacologically dilated with a solution of 1:7 10% phenylephrine HCl (Akorn Inc., Lake Forest, IL, USA): 1% tropicamide (Bausch & Lomb Inc., Tampa, FL, USA). Anterior ocular segment exams were repeated followed by posterior ocular segment fundus examinations performed via indirect ophthalmoscopy using a 60-diopter double aspheric handheld lens (Volk Optical Inc, Mentor, OH, USA) and a portable indirect headset (Keeler AllPupil II LED Vantage Plus Wireless Headset, Keeler Instruments Inc., Broomall, PA, USA). 

Electroretinography and Retinal Optical Coherence Tomography (OCT) Imaging

Mice underwent 12 hours of dark adaptation, and the pupil was dilated with a single drop of both tropicamide 1.0% and phenylephrine 2.5% ophthalmic solutions. A cocktail of ketamine and dexmedetomidine were injected intraperitoneally to induce anesthesia. Standard full-field scotopic and photopic ERG was performed on both eyes with the RETevet instrument (LKC Technologies, Gaithersburg, MD, USA) coupled to an ERG jet electrode. For data quantification, data from only one eye were used based on the quality of the ERG recording. We excluded data from eyes with noise owing to respiratory and other movements. 

Fundus photographs were obtained with the Micron III (Phoenix Research Laboratories, Pleasanton, CA, USA). The number of retinal lesions such as white to tan flecks¹² was counted on fundus photographs taken from both eyes of each animal, and the measurements were averaged for each mouse. Retinal OCT imaging was performed with an Envisu R2200 SD-OCT (spectral domain OCT; Bioptigen-Leica, Wetzlar, Germany). On OCT B-scan cross-sectional images, the thickness of the total retina was defined as the width from the retinal nerve fiber layer/ganglion cell layer (RNFL/GCL) to the RPE layer, including both layers and the thickness of retinal sublayers were also measured including RNFL/GCL, inner plexiform layer, inner nuclear layer, outer plexiform layer, outer nuclear layer (ONL), combined inner and outer PR segments, RPE, and choriocapillaris. These OCT retinal thickness parameters were manually measured at a distance of approximately 0.2 mm from superior, inferior, nasal, and temporal sides of the optic nerve head in ImageJ software as reported elsewhere.¹³ Measurements from all four quadrants of the optic nerve head and between the left and right eyes were averaged for each mouse. Left-sided fundus photography and OCT imaging from one wild-type mouse were not performed owing to anesthetic concerns and the parameters described elsewhere in this article were measured from the right-sided images only. 

Antibodies

Rabbit monoclonal antibody to phospho-ezrin (Thr567)/radixin (Thr564)/moesin (Thr558) (41A3) (catalog no. 3149; anti-phospho-ezrin/radixin/moesin [ERM]) was used for mice and rabbit monoclonal antibody to phospho-ezrin (Thr567)/radixin (Thr564)/moesin (Thr558) (48G2) (catalog no. 3726; anti-phospho-ERM used for nonhuman primate [NHP] and human) were purchased from Cell Signaling Technology (Danvers, MA, USA). Mouse monoclonal antibody to Clic6 (catalog no. sc-365303; anti-Clic6) was purchased from Santa Cruz Biotechnology. Secondary antibodies that were used include goat anti-rabbit IgG Alexa Fluor Plus 488 (ThermoFisher, catalog no. A32731) and donkey anti-mouse IgG Alexa Fluor Plus 568 (Thermo Fisher Scientific, Waltham, MA, USA; catalog no. A10037). 

Light Microscopy, Immunohistochemistry (IHC), and Electron Microscopy

Mice were euthanized at 16 weeks of age and both eyes were enucleated. Eyes were fixed in 4% paraformaldehyde for hematoxylin and eosin staining and IHC and in 2.5% glutaraldehyde and 2% PFA in 0.1 M sodium cacodylate buffer for transmission electron microscopy (TEM). 

Hematoxylin and eosin staining was performed on 12-µm sections of paraffin-embedded mouse globes. The tissue sections were deparaffinized in xylene for 5 minutes three times and rehydrated in a serial dilution of ethanol (twice in 100% ethanol and once in 95%, 75%, and 50% ethanol), 5 minutes each time. Then, the slides were stained with hematoxylin for 10 minutes and immersed in acid alcohol for 2 seconds and ammonia water for 20 seconds. The slides were rinsed in tap water for 5 minutes between each of those steps. The slides were then stained with eosin for 10 minutes and rinsed in a series of ethanol (twice in 75% ethanol, once in 95% ethanol, and twice in 100% ethanol), 2 minutes each time. The slides were finally rinsed in xylene twice, 2 minutes each time and mounted with Permount mounting medium. Slides were then imaged with OLYMPUS BX43 light microscope. 

The mouse, NHP, and human tissues were prepared for IHC as follows. Mouse globes were fixed in 4% PFA for 1 hour at room temperature, washed in PBS for 15 minutes three times, taken through 10%, 20%, and 30% sucrose gradients, and frozen embedded in optimal cutting temperature compound. Optimal cutting temperature–embedded mouse globes were sectioned at 10 µm thickness and collected on Superfrost plus slides. For NHP and human globes, the eye cups were prepared by gentle removal of anterior ocular segments, including the lens and vitreous. The eye cups were fixed in 4% PFA for 1 hour at room temperature followed by four PBS washes 15 minutes each and were cryoprotected with 30% sucrose at 4° until they sank. Then, four nonadjoining radial cuts were created with scissors and the macular region including choroid was dissected from the rest of the tissue in a rectangular shape by a razor blade, which was then frozen embedded in OCT compound. Optimal cutting temperature–embedded NHP and human tissues were sectioned at 16 and 12 µm thicknesses, respectively, and collected on Superfrost plus slides. 

For IHC staining, slides were thawed and rinsed in PBS for 3 minutes three times. Blocking and permeabilization were performed in PBS with 2% BSA and 0.2% Triton X-100 for an hour at room temperature. Then the tissue sections were incubated with primary antibodies (anti-phospho-ERM [1:100 dilution for mice and 1:50 for NHP and human] and anti-Clic6 [1:200 dilution for mice and 1:100 for NHP and human]) at 4°C overnight. Slides were then washed in PBS for 15 minutes three times before incubation with secondary antibodies (1:500 dilution for mice, NHP and human) for 1 hour at room temperature. Slides were then covered in DAPI (1 µg/mL) for 10 minutes, washed in PBS for 15 minutes three times, and coverslipped with FluorSave Reagent (Millipore, Burlington, MA, USA; catalog # 345789). Confocal fluorescence microscopy was performed on an Olympus FV3000. 

For TEM, small pieces of the fixed retina were processed as described before and imaged on a transmission electron microscope (Talos L120C with Ceta CMOS digital camera [4 K × 4 K], F.E.I. Company, Hillsboro, OR, USA).¹¹ The height of apical microvilli portion of RPE cell, the height of the remainder of RPE cell, the width of microvillus, the height of basal infoldings, pigment granule density in RPE, mitochondria density in RPE, and phagosome density in RPE were measured at an original magnification of ×4300 on ImageJ (Supplementary Fig. S1). Three different images from three different regions were obtained from each mouse and each parameter was measured in each image and averaged for each mouse. 

Statistical Analysis

Results are shown as mean ± SEM. A Student's t test was used to test the difference in the number of retinal lesions on fundus photographs, OCT and TEM measurements between Clic6^−/− mice and wild-type mice. All statistical analyses were done using GraphPad Prism (GraphPad Software Inc., La Jolla, CA, USA). 

Results

Potentiation of rd8 Phenotype in Clic6^−/− Mice

On slit-lamp examination, anterior subcapsular and nuclear cataracts and crystalline-like and pigmented vitreous opacities were found in all wild-type and Clic6^−/− mice, which was consistent with the previously reported background ocular lesions of C57BL/6N mice.^11,14 On fundus examination, all mice at 14 weeks of age showed multiple variable-sized white to tan retinal spots (Figs. 1A1 and 1B1) with significantly increased number of lesions in Clic6^−/− mice (Fig. 1C and Supplementary Fig. S2).^10,12 These retinal lesions correlated with focal distortion at the level of the outer plexiform layer and disorganization of deeper retinal layers extending into the ONL on OCT imaging (Figs. 1A2 and 1B2), consistent with the lamination defects of the rd8 phenotype inherent in C57BL/6N mice.^11,15,16 

Figure 1.

 

A more severe retinal phenotype was observed in Clic6^−/− mice. Multiple white to tan retinal spots were observed in all genotypes (A1 and B1). However, Clic6^−/− mice showed a significantly greater number of retinal lesions in comparison with the wild-type mice (P = 0.0144) (C). Representative OCT images at the peripapillary region from the same mice were shown (A2 and B2). These retinal lesions correlated with focal distortion at the level of the OPL (arrow in A2) and disorganization of deeper retinal layers extending into the ONL (bracket in B2) on OCT imaging, which is consistent with known retinal lamination defects of rd8 mice.

The ONL thickness was significantly thinner in Clic6^−/− mice vs. wild-type mice (54.2 ± 1.1 µm for Clic6^−/− mice and 57.4 ± 0.6 µm for wild-type mice; P < 0.05). Other OCT parameters, including total retinal thickness, thickness of RNFL/GCL, inner plexiform layer, inner nuclear layer, outer plexiform layer, combined inner and outer PR segments, RPE, and choriocapillaris, did not differ significantly between the genotypes (Fig. 2 and Table). There were no statistically significant differences in scotopic and photopic a- and b-waves of the ERG between genotypes (Supplementary Fig. S3). 

Figure 2.

 

The ONL thickness was significantly thinner in Clic6^−/− mice vs. wild-type mice. There was a significant difference in ONL thickness between wild-type and Clic6^−/− mice (P < 0.05). The other OCT parameters did not differ significantly between the two genotypes. An unpaired t test was performed, and the line and error bars indicate mean values and SEM. CC, choriocapillaris; INL, inner nuclear layer IPL, inner plexiform layer; OPL, outer plexiform layer; PRS, combined inner and outer PR segments.

Table.

 

Comparison of Retinal OCT Measurements Between Clic6^— Mice and wild-type Mice

 

Hypomorphic Apical RPE Surface in Clic6^−/− Mice

On hematoxylin and eosin staining, lamination defects were more commonly observed with greater severity in Clic6^−/− compared with wild-type mice (Fig. 3). Clic6 immunofluorescent expression was absent in Clic6^−/− mice, but detected in both apical and basal surfaces of RPE cells in wild-type mice (Figs. 4A and 4D). On the apical RPE surface in wild-type mice, Clic6 expression was observed in the long, extended microvilli that were interdigitated into PR layer (Fig. 4A, inset). ERM, including ezrin, a marker of RPE microvilli, was colocalized with Clic6 at the apical RPE surface in mice (Figs. 4B and 4C). ERM expression was still detected in Clic6^−/− mice (Fig. 4E). However, it seemed to be flat and blunted, and the long and thin projections from the RPE apical surface, a typical appearance of microvilli, were not identified in Clic6^−/− vs. wild-type mice (Fig. 4E, inset). 

Figure 3.

 

The lamination defects were more commonly observed with greater severity in Clic6^−/− mice. The pseudorosettes (arrows) caused by Crb1^rd8 mutation are more severe in the outer retina of Clic6^−/− mice (D–F) compared with the wild type (A–C).

Figure 4.

 

ERM expression was altered in RPE apical microvilli of Clic6^−/− mice. ERM expression was found along the apical RPE surface and microvilli in wild-type mice (A–C). In Clic6^−/− mice, the appearance of ERM expression appeared blunted and flat and the long and thin projections from RPE apical surface, a typical appearance of microvilli, was not observed (D–F).

Short and Broadened RPE Apical Microvilli in Clic6^−/− Mice

On TEM, the apical portion of the RPE differed in Clic6^−/− vs. wild-type mice (Fig. 5). More specifically, the height of the apical microvilli portion was significantly shorter in Clic6^−/− mice (1.40 ± 0.17 µm) vs. wild-type mice (2.51 ± 0.11 µm; P = 0.0050). The width of each RPE microvillus appeared greater in Clic6^−/− mice (0.27 ± 0.08 µm) vs. wild-type mice (0.08 ± 0.01 µm; P = 0.0781), but it did not achieve statistical significance. These differences led to an abnormally broadened and stout appearance of RPE microvilli in Clic6^−/− mice (Fig. 5). In the cytoplasm, the phagosome density was significantly greater in Clic6^−/− mice (5.44 ± 1.06 counts/EM field) vs. wild-type mice (11.33 ± 0.77 counts/EM field; P = 0.0109). Additionally, electron-dense lipid droplets were occasionally seen in the RPE cytoplasm of Clic6^−/− mice (Fig. 5). 

Figure 5.

 

RPE microvilli appeared broadened or absent in Clic6^−/− mice on TEM. In comparison with wild-type mice (A), the apical microvilli of RPE appeared broadened (B) and almost obliterated in some regions of Clic6^−/− mice (C). The arrows indicate phagosomes, the M indicate broadened RPE microvilli, and the asterisks indicate lipid droplets. POS, PR outer segments.

The other TEM parameters, including the height of the remainder of the RPE cell (6.85 ± 0.57 µm and 6.13 ± 0.72 µm for Clic6^−/− and wild-type mice, respectively), the height of basal infoldings (1.13 ± 0.11 µm and 1.12 ± 0.05 µm for Clic6^−/− and wild-type mice, respectively), mitochondria density (16.33 ± 2.91 counts/EM field and 19.89 ± 1.68 counts/EM field for Clic6^−/− and wild-type mice, respectively), and pigment granule density (41.67 ± 8.29 counts/EM field and 42.33 ± 2.70 counts/EM field for Clic6^−/− and wild-type mice, respectively) did not differ significantly between genotypes (Fig. 6). The morphology of intracellular organelles including mitochondria and pigment granules appeared unremarkable and Bruch's membrane was intact in all genotypes. 

Figure 6.

 

Clic6 ^−/− mice showed significantly shorter and more broadened RPE apical microvilli with increased number of phagosomes in the cytoplasm. On TEM, the height of the apical microvilli portion was significantly shorter in Clic6^−/− mice (P < 0.01) and the width of each microvillus seemed to be wider in Clic6^−/− mice, which is consistent with the short and broadened appearance of RPE microvilli in Clic6^−/− mice. Phagosomes were significantly more commonly observed in Clic6^−/− mice (P < 0.05). All parameters were measured on TEM images taken at an original magnification of ×4300. An unpaired t test was performed, and the line and error bars indicate mean values and SEM.

CLIC6 Expressions in NHP and Human Retinas

In the 22-week-old human fetal retina, CLIC6 expression was observed in the RPE, as well as in the apical microvilli of RPE, while ERM was localized in the RPE only (Figs. 7A–C). In the adult NHP retina, both CLIC6 and ERM expressions were identified in the RPE cells. In comparison to mice and human fetal retinas, however, CLIC6 was not expressed in the apical RPE microvilli of the adult NHP (Figs. 7D–F). 

Figure 7.

 

CLIC6 expression was observed in the RPE of NHP and human retinas. In the 22-week-old human fetal retina, CLIC6 expression was identified in the apical RPE microvilli as well as in the RPE (A–C). In the 18-year-old adult Rhesus macaque retina, CLIC6 expression was localized in the RPE only and ERM was only found in the apical microvilli (D–F).

Discussion

In the present study, we demonstrated that Clic6 deficiency has a substantial impact on RPE microvilli morphogenesis. Although the apical surface of RPE cells in age-matched wild-type littermates extended very long and thin microvilli that interdigitated with PR outer segments, Clic6^−/− mice showed shortening and broadening of apical microvilli and thereby a reduced intimacy between the RPE microvilli and the PR. We also noted that phagosomes containing partially digested PR outer segments and electron-dense lipid droplets were observed more commonly in the RPE cytoplasm of Clic6^−/− mice, which may suggest impaired processing of PR outer segments discs in the RPE of Clic6^−/− mice. 

The molecular mechanism of how Clic6 deficiency resulted in abnormal RPE microvilli has yet to be determined. Ezrin, a member of the ERM family, is a highly conserved protein in mammals including humans, tethering actin filaments to the plasma membrane.¹⁷ In the RPE, ezrin is exclusively localized in the apical surface and present throughout the entire length of the RPE microvilli.^18,19 It was demonstrated that ezrin is the major determinant for the number and length of the actin-rich apical microvilli of RPE,¹⁹ which explains the fewer, shorter, and broader appearance of RPE microvilli in Clic6^−/− mice. In addition to the RPE microvilli morphogenesis, ezrin directly binds to lysosomal-associated membrane protein 1 and regulates phagosome–lysosome fusion, which is involved in the phagocytic digestion of PR outer segments by RPE.^20,21 The downregulation of ezrin activity in mouse macrophage cell line RAW264.7 and an imbalance between ezrin and lysosomal-associated membrane protein 1 expressions in ARPE19 cells were reported to affect phagocytosis.^20,21 The increased phagosome density in Clic6^−/− mice may be a consequence of reduced ezrin or ERM expression in RPE. 

Notably, ezrin-deficient mice showed fewer and reduced basal infoldings in the RPE in conjunction with the aforementioned changes in microvilli.²² Clic4, another member of the CLIC family, may function in the RPE similar to ezrin. Clic4-silencing in rats induced the concomitant loss of both microvilli and basal infoldings and RPE-specific Clic4 knockout mice showed deconvolution of basal infoldings in conjunction with similar microvilli appearance.^8,23 Contrary to the ezrin and Clic4 deficiency, however, the basal infoldings of RPE were not affected in Clic6^−/− mice as demonstrated on TEM. This finding implies that Clic4 and Clic6 are not redundant and function differently in regulating RPE morphogenesis with Clic6 being more specific to the apical microvilli. 

Indeed, the morphological changes in microvilli in conjunction with decreased ONL thickness on OCT did not cause any apparent functional disturbance in Clic6^−/− mice, because no differences in ERG recordings were observed between genotypes. It was reported that ezrin and Clic4 deficiency significantly inhibited PR development with disorganization of PR outer segments and ONL, respectively.^22,23 By contrast, the relatively intact basal infoldings of the RPE and the less severe retinal phenotype observed in Clic6^−/− mice is likely the reason that functional abnormalities were not seen. Age may also be a factor; the Clic6^−/− mice used in this study were only 4 months of age. In the longitudinal study with RPE-specific Clic4 knockout mice, significantly lower ERG b-wave amplitudes were detected beginning at 6 and 9 months of age for rods and cones, respectively.⁸ We presume that accumulated cellular stress exceeding the normal homeostasis of RPE in aged Clic6^−/− mice may eventually affect the overall PR turnover, causing clinically significant PR degeneration and a decrease in vision. 

This finding, in turn, may suggest that the role of apical microvilli in the early disease process is under-recognized. Despite the recent advancements in high-resolution adaptive optics imaging systems in combination with lipofuscin autofluorescence allowing an observation of the RPE to the single cell level, in vivo RPE imaging has been confined to en face images of the RPE mosaic or 3D reconstructed images of the PR–RPE complex,^24,25 which is insufficient to visualize apical microvilli pathology in the living human retina. There is growing evidence that dysregulated cytoskeletal dynamics of apical microvilli is associated with retinal diseases. The ERM-binding phosphoprotein 50 links intracellular retinaldehyde binding protein to ezrin and the actin cytoskeleton at the RPE apical microvilli.^26,27 Intracellular retinaldehyde binding protein is known to be responsible for retinoid trafficking between the RPE and PRs in the visual cycle.²⁷ It was demonstrated that ERM-binding phosphoprotein 50 expression was upregulated in patients with BEST1 mutations as well as during RPE aging.^28,29 The importance of apical microvilli and RPE polarity was well-described in Bbs8^−/− mice having primary cilium defects. Bbs8^−/− mice with abnormal microvilli showed RPE maturation defects with compromised phagocytosis and decreased low potassium response and subsequently developed PR defects.³⁰ Another mouse model with altered microvilli by RPE-specific Ift20 knockout demonstrated microvilli defects preceded the retinal degeneration.³¹ Along similar lines, iPSC-RPE derived from a ciliopathy patient with CEP290 mutation exhibited lower ezrin expression, smaller microvilli size, and decreased phagocytic ability.³⁰ These studies suggest that aberrant RPE microvilli may contribute to the initiation and/or progression of retinal degeneration. And we highlight that CLIC6 expression was identified in apical RPE microvilli of human fetal retinas, which implies its significance on normal RPE and retinal development. 

One of the major limitations of the present study is the rd8 mutation background of Clic6^−/− mice that were used. Multiple white to tan retinal spots were observed on fundus examination in both wild-type and Clic6^−/− mice, which correlated with lamination defects on OCT, as previously described in C57BL/6N mice owing to the homozygous rd8 mutation in the Crb1 gene.^11,14,15,32 It is likely that the fundus lesions observed in this study is not a direct consequence of Clic6 deficiency, and we presume that Clic6 deficiency may have potentiated the rd8 phenotype as demonstrated in Cygb knockout mice with the same rd8 background.³² The lamination defects may arise from aberrant cell polarity between progenitor cells, Muller glial cells, and newborn PRs during retinal development. An important recent report suggests that there may be a gut-derived microbiome component to rd8-associated retinal lesions as well.³³ Although ezrin expression was absent in PRs and all other cell types of the retina in rats, it was detected in the microvilli of Muller glial cells, as well as the apical microvilli of RPE.¹⁸ Given the altered ezrin expression in Clic6^−/− mice, the Clic6 gene may have played a role in retinal lamination as a potential genetic modifier, worsening the rd8 phenotype. A transcriptional study to determine the impact of Clic6 on the regulation of Crb1-related cell polarity would be helpful to understand the molecular mechanism. Furthermore, a longitudinal study with Clic6 knockout mice free of the rd8 mutation is also warranted to better characterize the functional consequence of Clic6 deletion and, thereby, clarify the genotype–phenotype relationship and to investigate RPE phagocytosis in Clic6^−/− mice as a cause of potential retinal degeneration. 

In conclusion, we demonstrated that Clic6 deficiency induced aberrant apical microvilli in the RPE and worsened the rd8 phenotype that was inherent in the background of the mice used. Despite the limitations, this study provides insights into the role of Clic6 and RPE microvilli in the pathophysiology of retinal degeneration. 

Acknowledgments

Supported by the NIH/NEI EY033733 (SMT) and P3012576 (SMT). 

Disclosure: S. Park, None; N. Suvarnpradip, None; N. Kasiri, None; G. Yiu, None; S.M. Thomasy, None; D.M. Imai, None; K.C.K. Loyd, None; B.C. Leonard, None; A. Moshiri, None