Modulation of Tmem135 Leads to Retinal Pigmented Epithelium Pathologies in Mice

Abstract

Purpose: Aging is a critical risk factor for the development of retinal diseases, but how aging perturbs ocular homeostasis and contributes to disease is unknown. We identified transmembrane protein 135 (Tmem135) as a gene important for regulating retinal aging and mitochondrial dynamics in mice. Overexpression of Tmem135 causes mitochondrial fragmentation and pathologies in the hearts of mice. In this study, we examine the eyes of mice overexpressing wild-type Tmem135 (Tmem135 TG) and compare their phenotype to Tmem135 mutant mice.

Methods: Eyes were collected for histology, immunohistochemistry, electron microscopy, quantitative PCR, and Western blot analysis. Before tissue collection, electroretinography (ERG) was performed to assess visual function. Mouse retinal pigmented epithelium (RPE) cultures were established to visualize mitochondria.

Results: Pathologies were observed only in the RPE of Tmem135 TG mice, including degeneration, migratory cells, vacuolization, dysmorphogenesis, cell enlargement, and basal laminar deposit formation despite similar augmented levels of Tmem135 in the eyecup (RPE/choroid/sclera) and neural retina. We observed reduced mitochondria number and size in the Tmem135 TG RPE. ERG amplitudes were decreased in 365-day-old mice overexpressing Tmem135 that correlated with reduced expression of RPE cell markers. In Tmem135 mutant mice, RPE cells are thicker, smaller, and denser than their littermate controls without any signs of degeneration.

Conclusions: Overexpression and mutation of Tmem135 cause contrasting RPE abnormalities in mice that correlate with changes in mitochondrial shape and size (overfragmented in TG vs. overfused in mutant). We conclude proper regulation of mitochondrial homeostasis by TMEM135 is critical for RPE health.

Aging imposes significant stress on the retina and contributes to the development of age-related macular degeneration (AMD),¹ diabetic retinopathy,² and glaucoma.³ Although there is an excellent understanding of the pathways disturbed by aging,⁴ it is unclear how aging influences the risk for these debilitating age-dependent retinal diseases. Studying how aging perturbs retinal homeostasis may lead to novel therapies for age-dependent retinal diseases because the prevalence of such diseases are expected to increase in the future dramatically.^5–7 

To identify critical pathways involved in retinal aging, we used forward genetics to uncover genes harboring mutations responsible for accelerated aging retinal phenotypes in mice. Using this approach, we studied an ENU-induced mutant mouse line, FUN025, displaying visual loss, photoreceptor degeneration, thicker retinal pigmented epithelium (RPE), increased ectopic synapses, and neuroinflammation.⁸ Positional cloning of FUN025 revealed a mutation in transmembrane protein 135 (Tmem135) as causal for the accelerated aging retinal phenotype in the FUN025 line.⁸ Tmem135 encodes a 52 kilodalton protein with five transmembrane domains that share similarities with the Tim17 protein family, a group of proteins important for mitochondria biogenesis.^8,9 TMEM135 can localize with mitochondria,⁸ as well as peroxisomes,^10,11 and lipid droplets.¹⁰ The molecular function of TMEM135 is unclear but has a role in regulating mitochondrial dynamics.⁸ “Mitochondrial dynamics” is the collective term for the fusion, fission, and mitophagy events mitochondria undergo to preserve their shape, number, and function.¹² Primary-cultured murine fibroblasts expressing mutant TMEM135 and overexpressing wild-type (WT) TMEM135 show increased numbers of elongated and fragmented mitochondria, respectively.⁸ Elucidating the function of TMEM135 on mitochondrial dynamics in the murine eye may shed insight into how aging affects retinal homeostasis. 

Tmem135 overexpression leads to excess cardiac mitochondrial fission and pathologies, including hypertrophy, fibrosis, and collagen accumulation in mice.¹³ It is unknown whether Tmem135 overexpression affects the eye. In this study, we evaluated the eyes of mice overexpressing WT Tmem135 (Tmem135 TG) for pathologies and compared our findings to those in Tmem135^{FUN025/FUN025} mutant mice. 

Materials and Methods

Animals

All experiments performed in this study were in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals and were authorized by the Animal Care and Use Committee at the University of Wisconsin-Madison. Generation of Tmem135 TG and Tmem135^{FUN025/FUN025} mice was previously described.^8,13 For the current study, we generated Tmem135 TG mice congenic on the C57BL/6J background by crossing Tmem135 TG mice on the mixed background with C57BL/6J mice for at least 10 generations. None of the mice used in this study carried either the Ped6b^rd1 or the Crb1^rd8mutation.^14,15 Age-matched littermate WT mice served as controls for Tmem135 TG mice, and Tmem135^FUN025/+littermates were used as controls for Tmem135^{FUN025/FUN025} mice. Both female and male mice were used in this study, but no gender differences were observed, and figures include data from both genders. All ages, genders, and cohorts of WT and Tmem135 TG, as well as Tmem135^FUN025/+ and Tmem135^{FUN025/FUN025} mice, are summarized in Supplementary Tables S1 and S2, respectively. All numbers and ages of mice used for experiments are also listed in the figure legends. All animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

RNA Isolation, cDNA Preparation, and Quantitative PCR

Tissues were collected from the mice and kept at –80°C degrees. RNA was extracted using an RNeasy lipid tissue mini kit (#74804; Qiagen, Germantown, MD, USA) according to the manufacturer’s instructions. RNA concentrations were determined using the Nanodrop 2000 UV-Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Equal RNA concentrations were used to make cDNA using an Oligo d(T)18 primer (#S1316S; NEB, Ipswich, MA, USA) and the ProtoScript II Reverse Transcriptase (#M0368L; NEB) following the manufacturer’s protocol. Expression of each gene was assessed in triplicate reactions in the Roche Lightcycler 480 system using cDNA, 200 nmol/L of each primer, and Lightcycler 480 SYBR Green I Supermix (#507203180; Roche, Indianapolis, IN, USA). Primer sequences used in this study are described in Supplementary Table S3. Relative mRNA expression was normalized to the ribosomal protein lateral stalk subunit P0 (Rlpl0) using the quantitative 2^−ΔΔCT method.¹⁶ 

Histology

Eyes were fixed in 2% paraformaldehyde (PFA) and 2% glutaraldehyde solution overnight at 4°C and submitted to the University of Wisconsin-Madison’s Translational Research Initiatives in Pathology (TRIP) core for paraffin embedding, processing, and sectioning. Sections were stained with hematoxylin and eosin to visualize retinal architecture and imaged using an Axio Imager 2 microscope (Carl Zeiss MicroImaging, White Plains, NY, USA) at a 40X magnification. Images were analyzed using ImageJ software (NIH, Bethesda, MD, USA). The thickness of the RPE, outer nuclear layer (ONL), and inner nuclear layer (INL) were measured on a single section from each eye that had a visible optic nerve head. Measurements were started from the optic nerve head and taken at seven consecutive 300-µm intervals. Quantification of RPE pathologies (i.e., migratory cells and vacuolization) were assessed on six retinal sections that were separated by 50 µm. The total number of RPE pathologies were divided by the total number of slides to generate the frequency of RPE pathologies per eye. 

Immunohistochemistry

Eyes were fixed in 4% PFA overnight at 4°C, processed for cryoprotection and embedded in Tissue-Tek O.C.T. Compound (Sakura Finetek USA, Torrance, CA, USA). Sections were cut at 10-µm thickness on a cryostat. We labeled subretinal immune cells with rabbit anti-ionized calcium binding adaptor molecule 1 (IBA-1) (# 019–19741; Wako, Richmond, VA) as previously described.^8,17 One retinal section per mouse was imaged using a Nikon A1RS confocal microscope (Nikon, Tokyo, Japan) at a 40X magnification, and IBA-1 positive immune cells between the RPE and ONL were counted using ImageJ software (NIH). 

RPE Flat Mounts

Eyes were enucleated and dissected to remove the anterior segment, lens, and retina. The remaining posterior eyecup was fixed in methanol overnight and then washed with 1X PBS the next day. The eyecup was incubated with 10% normal donkey serum (#017-000-121; Jackson ImmunoResearch, West Grove, PA, USA) for at least 30 minutes and then switched to a 1:50 dilution of a rabbit polyclonal zonula occludens-1 antibody (#40-2200; Thermo Fisher Scientific) overnight at 4°C. Then, the eyecups were rinsed with 1X PBS and incubated with a 1:250 dilution of an Alexa Fluor 488 conjugated donkey anti-rabbit IgG (#A-21206; Thermo Fisher Scientific) and 4′,6-Diamidine-2′-phenylindole dihydrochloride (DAPI) (#D9542; Sigma Aldrich, St. Louis, MO, USA) for 2 hours. The eyecups had a final rinse with 1X PBS and were mounted on slides. The RPE flat mounts were imaged with a Nikon A1RS confocal microscope at a 20X magnification. For each RPE flat mount, an image was obtained from the north, east, south, and west quadrants surrounding the optic nerve. RPE flat mounts were analyzed to determine RPE cell size and density. RPE cell size was calculated using the RPE Cell Profiler program (Broad Institute, Boston, MA, USA) that was previously used to measure RPE cell size in human donor eyes.¹⁸ RPE cell density was assessed by counting the number of RPE cells using the Cell Counter Plugin in NIH’s ImageJ program and dividing the number of total counted RPE cells by the total image area. 

Electron Microscopy (EM)

Eyes were fixed with 2% PFA and 2% glutaraldehyde and submitted to the Electron Microscope Core at the University of Wisconsin-Madison for transmission EM processing as previously described.¹³ Eye sections were mounted on a 400-mesh thin bar grid and images were collected where the grid bars intersected the RPE using a Phillips CM120 STEM microscope (FEI Company, Hillsboro, OR, USA) at a 15,000X magnification. For this study, we collected 163 EM images for WT RPE and 99 images for Tmem135 TG RPE. Mitochondria number, surface area, pigment granule number, Bruch's membrane (BrM) thickness, and basal laminar deposit frequency were determined using NIH's ImageJ software. 

Electroretinography (ERG)

Prior to the ERG recording, mice were dark-adapted overnight. They were anesthetized with a cocktail of ketamine (80 mg/kg) and xylazine (16 mg/kg) diluted in PBS by subcutaneous injection under dim red light. The pupils were dilated with 1% tropicamide and placed on a heating pad set to 37°C to maintain body temperature and later placed on a temperature-regulated heating plate equipped with the system during the ERG recording. 

A drop of sterile 2.5% hypromellose ophthalmic solution (Goniovisc; HUB Pharmaceuticals LLC, Rancho Cucamonga, CA) was applied to the cornea of the dilated eyes to prevent dehydration for the electrical contact and placement of the recording electrode. Needle electrodes were inserted through the nostril and on the back as reference and ground electrode, respectively. ERGs were recorded using the Espion system (Diagnosys LLC, Lowell, MA, USA). Scotopic recordings were obtained from dark-adapted animals, and the eyes were exposed to a sequential increment of flash intensities (0.1–30 cd.s/m²) for 300 ms with a 2 second interval between the flashes. For the acquisition of c-wave, the eyes were flashed with light intensities of 2.5 and 25 cd.s/m² for 4 seconds. Photopic recordings were performed after the light adaptation of mice at 30 cd.m² background and also during the photopic recordings. The flash intensities used for the photopic recordings were 0.78, 1, 2.25, 5, and 10 cd.s/m². Only the 10 cd.s/m² recordings were used in this study. ERG components were measured using the Espion software (Diagnosys LLC) and analyzed using Origin2020 (OriginLab Corp., Northampton, MA, USA). 

Western Blot Analysis

Neural retina, eyecup (RPE/choroid/sclera), and heart were isolated from mice and stored at –80°C. Neural retina, eyecup, and heart lysates were made with a tissue homogenizer in 100, 80, and 1000 µL of RIPA buffer (#P189901; Thermo Fisher Scientific) containing protease inhibitors (#11836170001; Thermo Fisher Scientific), respectively. Protein concentrations were quantified using a BCA Protein Assay Kit (#P123228; Thermo Fisher Scientific). Equal protein amounts were aliquoted and loaded onto 10% Bis-Tris Criterion XT gels (#3450112; Biorad, Hercules, CA, USA) in MOPS buffer (#1610788; Biorad) and transferred to nitrocellulose membranes (#102673-324; Biorad). A total of 20 µg of each eyecup lysate, 30 µg of each neural retina lysate, and 30 µg of each heart lysate was used for Western blot analysis in this study. Membranes were blocked with milk and probed overnight with their respective primary antibody at 4°C. The primary antibodies used in this study were anti-CRALBP (#ab15051; Abcam, Cambridge, MA, USA), anti-Rhodopsin (#ab5417; Abcam), and total rodent OXPHOS Western blot antibody cocktail (#ab110413; Abcam). For loading controls, blots containing neural retinal samples were reprobed for tubulin (#3873S; Cell Signaling Technologies, Danvers, MA), blots containing eyecup lysates were probed for succinate dehydrogenase A (SDHA) (#11998S; Cell Signaling Technologies) or beta-actin (ACTB) (#4970S; Cell Signaling Technologies), and blots containing heart lysates were probed for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (#MCA-1D4; EnCor Biotechnology, Gainesville, FL, USA). Blots were washed with TBST buffer and incubated with their corresponding secondary antibody. Secondary antibodies used in this study were either donkey antimouse or antirabbit and were obtained from LI-COR Biosciences (Lincoln, NE, USA). Blots were washed again with TBST and imaged using the Odyssey Imaging System (LI-COR Biosciences) and analyzed using NIH's ImageJ software. All immunobands were normalized to the loading control on the immunoblot. 

RPE Cell Culture and Mitochondria Staining

The primary RPE cultures were carried out using a previously described method with a slight modification.^19,20 Specifically, RPE cells were continued to be cultured in T75 flasks in complete Dulbecco's modified Eagle's medium after the initial harvest and subcultured at a 1:3 dilution every 3 days until they were used at passage 5 and reached at least 80% confluence for the mitochondrial morphological analysis. We immunolabeled a subset of our cultures with antibodies against beta-catenin (#C2206; Sigma Aldrich) to image tight junctions and RPE65 (#ab13826; Abcam), OTX1/OTX2 (#ab21990; Abcam), and CRALBP (#ab15051; Abcam) as RPE markers. RPE cells were stained with MitoTracker Red CMXRos (#M7512; Thermo Fisher Scientific) and DAPI using the same protocol as previously described.⁸ 

Statistics

GraphPad Prism software (GraphPad, San Diego, CA, USA) was used to perform analysis of variance (ANOVA) tests between tissues, genotypes at each age point, and Tmem135 TG mice across ages in this study. Significance and error bar information is presented in each figure legend. 

Results

Tmem135 Expression in Mouse Eye

Tmem135 is produced in the eyes and hearts of mice,^8,13 but the relative expression of Tmem135 across various tissues in mice is not known. Multiple tissues, including eyecup, neural retina, heart, kidney, brain, and liver from WT C57BL/6J mice, were collected to evaluate Tmem135 expression. The fold change of Tmem135 expression relative to eyecup was 8.66 ± 1.82 in neural retina; 1.86 ± 0.72 in heart; 3.84 ± 0.35 in kidney; 5.72 ± 0.71 in brain; and 6.93 ± 0.36 in liver (Fig. 1A). These results suggest similar levels of Tmem135 expression in the eyecups and hearts, which are lower than the neural retinas and other tissues of WT mice. 

Figure 1.

 

Tmem135 expression in WT and Tmem135 TG tissues. (A) Tissues were harvested from three 180-day-old WT mice for RNA isolation and quantitative PCR to assess Tmem135 expression. Data were normalized to the eyecup (RPE/choroid/sclera) value. Tmem135 is highly expressed in the neural retina compared with the eyecup and heart. (B) The eyecups and neural retinas were collected from four 30-day-old WT and Tmem135 TG mice to confirm the Tmem135 overexpression driven by the chicken beta-actin promoter. Both the eyecups and neural retinas of Tmem135 TG mice had amplified Tmem135 expression that were not different. Data were normalized to the WT eyecup value. ***P < 0.001, ANOVA. NS, not significant. All data are mean ± SD.

We also examined Tmem135 levels in the neural retinas and eyecups of WT and Tmem135 TG mice to confirm Tmem135 overexpression. There was a dramatic augmentation in Tmem135 expression in the eyecup (49.91 ± 4.12) and neural retina (43.05 ± 9.06) of Tmem135 TG mice compared with the WT mice (Fig. 1B). The increased Tmem135 expression was not statistically different between the eyecup and neural retina (P = 0.22). These data show comparable Tmem135 overexpression in posterior eye tissues of Tmem135 TG mice. 

Histological Examination of Tmem135 TG eyes

To determine if there is an effect of Tmem135 overexpression on the retina, we collected eyes from 20-, 30-, and 240-day-old Tmem135 TG mice for histology and compared them to age-matched WT littermates. At 20 days of age, we observed no differences between the retina of WT and Tmem135 TG mice (Fig. 2A). At 30 days of age, there were RPE abnormalities in Tmem135 TG mice, such as vacuolization, that were absent in WT mice (Fig. 2A, arrow). These pathologies along with RPE degeneration were noted in 240-day-old Tmem135 TG mice (Fig. 2A). We measured the thickness of the RPE to quantify RPE degeneration. No differences were found in the RPE thickness between WT and Tmem135 TG mice at 20 and 30 days of age (Figs. 2B, 2C). At 240 days of age, there were significant decreases in the RPE height in Tmem135 TG mice compared with WT mice (Fig. 2D). We normalized the ONL thickness (ONLT) to the INL thickness to generate the ONLT ratio to investigate if there is photoreceptor degeneration in the Tmem135 TG mice. We detected no differences in the ONLT ratios between Tmem135 TG and WT mice at 20, 30, and 240 days of age (Figs. 2E–G). Our survey of Tmem135 TG eyes revealed signs of RPE degeneration but no signs of photoreceptor degeneration. 

Figure 2.

 

Retinal architecture in WT and Tmem135 TG mice. (A) Representative images of the retina from P20, P30, and P240 WT and Tmem135 TG mice. Note the large vacuolization (yellow arrow) in P30 Tmem135 TG mice and thin RPE in P240 Tmem135 TG mice. Magnification = 40X. Scale bar = 50 µm. (B–D) RPE thickness. There was a significant decrease in RPE thickness in P240 Tmem135 TG mice. (E–G) ONLT ratios. There were no changes in ONLT ratios. Data from n = 6 WT (P20), n = 4 Tmem135 TG (P20), n = 5 WT (P30), n = 4 Tmem135 TG (P30), n = 3 WT (P240), and n = 5 Tmem135 TG (P240). *P < 0.05, **P < 0.01, and ***P < 0.001, ANOVA. All data are mean ± SEM. RGC, retinal ganglion cell layer; IPL, inner plexiform layer; OPL, outer plexiform layer; IS, inner segments; OS, outer segments.

We further investigated the type and extent of RPE pathologies in 30-day-old Tmem135 TG mice. At this age, several changes were noted in the RPE of Tmem135 TG mice including migratory cells (Fig. 3A), and small (Fig. 3B) and large vacuolization (Fig. 3C). We examined whether the migratory cells were subretinal immune cells engulfing sick RPE cells by labeling 30-day-old WT and Tmem135 TG retinal cryosections with anti-IBA-1 (Supplementary Fig. S1). We did not see any IBA-1 positive immune cells in the subretinal space of 30-day-old Tmem135 TG eyes (Supplementary Fig. S1), suggesting that these migratory cells are RPE cells. The frequency of the RPE pathologies were quantified for each 20-, 30- and 240-day-old WT and Tmem135 TG mouse eye. Our analysis revealed that the frequency of RPE migratory cells were not different between 20-day-old WT and Tmem135 TG mice but significantly increased in 30- and 240-day-old Tmem135 TG mice compared with WT mice (Fig. 3D). We observed a similar trend in the frequency of small and large RPE vacuolization in 240-day-old Tmem135 TG mice (Figs. 3E, 3F), although the increase in extensive RPE vacuolization does not reach statistical significance (P = 0.15) (Fig. 3F). Interestingly, there were significant decreases in the frequency of small and large RPE vacuolization in 240-day-old Tmem135 TG mice compared with 30-day-old Tmem135 TG mice (Figs. 3E, 3F), which may be due to RPE degeneration that occurs between these time points. In summary, overexpression of Tmem135 leads to a variety of RPE pathologies. 

Figure 3.

 

RPE pathologies in Tmem135 TG mice. Representative images of P30 Tmem135 TG RPE (A–C). RPE pathologies in Tmem135 TG mice include migratory RPE cells (A) and vacuoles in RPE that were small (B, yellow arrows) and large (C). Magnification = 40X. Scale bar = 100 µm. (D–F) Quantification of migratory RPE cells, small and large RPE vacuolization in WT and Tmem135 TG mice. There were significant increases in these RPE pathologies in P30 and P240 Tmem135 TG mice although large RPE vacuolization in P240 Tmem135 TG mice did not reach statistical significance. There were significant increases in these RPE pathologies between P20 and P30 Tmem135 TG mice. Small and large RPE vacuolization significantly decreased between P30 and P240 Tmem135 TG mice. Data from n = 8 WT (P20), n = 7 Tmem135 TG (P20), n = 5 WT (P30), n = 6 Tmem135 TG (P30), n = 4 WT (P240), and n = 6 Tmem135 TG (P240). *P < 0.05, **P < 0.01, and ***P < 0.001, ANOVA. All data are mean ± SEM. ND, not detected; NS, not significant; V, vacuole.

Size and Density of RPE Cells in Tmem135 TG Mice

RPE degeneration and abnormalities correlate with RPE dysmorphogenesis, including disorganization and enlargement of RPE cells in mice.^21,22 To determine if there is RPE dysmorphogenesis in the Tmem135 TG eyes, we prepared and labeled RPE flat mounts from 20-, 30- and 240-day-old mice with an antibody against the tight junction-associated protein ZO-1 to delineate RPE cells and DAPI to stain nuclei. The RPE of 20-day-old Tmem135 TG and WT mice were indistinguishable in their morphology and size (Fig. 4A). Conversely, there was disorganization of the RPE in 30- and 240-day-old Tmem135 TG mice with loss of their classic hexagonal patterning (Fig. 4A). Analysis of RPE cell size revealed similar averages between 20-day-old Tmem135 TG and WT mice (Fig. 4B). At 30 and 240 days of age, there was a significant increase in RPE cell size in Tmem135 TG mice relative to WT mice (Fig. 4B). We observed significant increases in Tmem135 TG RPE cell size across the ages used in this study (Fig. 4B). Evaluation of RPE cell density uncovered similar densities between Tmem135 TG and WT mice at 20 days of age (Fig. 4C). It significantly decreased in Tmem135 TG mice compared with age-matched WT mice at 30 and 240 days of age (Fig. 4C). We found RPE cell density decreased in Tmem135 TG mice between the ages examined in this experiment (Fig. 4C, a statistically significant change between P20 and P30 TG, and a decreasing trend between P30 and P240 TG [P = 0.066]). Together, these analyses showed that Tmem135 TG mice have progressive RPE dysmorphogenesis and cell enlargement. 

Figure 4.

 

RPE dysmorphogenesis and cell enlargement in WT and Tmem135 TG mice. (A) Representative images of WT and Tmem135 TG RPE flat mounts at P20, P30, and P240 with ZO-1 in green and DAPI in blue. Magnification = 20X. Scale bar = 100 µm. (B) RPE cell size and (C) RPE cell density in WT and Tmem135 TG mice. The RPE cell size in P30 and P240 Tmem135 TG mice were greater than WT mice. The RPE cell density in P30 and P240 Tmem135 TG mice were smaller than WT mice. There were age-dependent increases in RPE cell size and decreases in RPE cell density in Tmem135 TG mice. Data from n = 6 WT (P20), n = 4 Tmem135 TG (P20), n = 4 WT (P30), n = 4 Tmem135 TG (P30), n = 4 WT (P240), and n = 6 Tmem135 TG (P240). *P < 0.05, **P < 0.01, and ***P < 0.001, ANOVA. All data are mean ± SEM.

Ultrastructure of the RPE in Tmem135 TG mice

EM was undertaken to visualize if there were ultrastructural abnormalities in Tmem135 TG RPE (Fig. 5A). This analysis showed the presence of migratory RPE cells (Supplementary Fig. S2A); hyperpigmented RPE cells (Supplementary Fig. S2B); and small (Supplementary Fig. S2C), medium (Supplementary Fig. S2D), and large vacuolization (Supplementary Fig. S2E) within RPE cells of 60-day-old Tmem135 TG mice. These pathologies were not observed in WT RPE at 60 days of age. We delved deeper into the Tmem135 TG RPE phenotype by assessing mitochondria, pigment granules, and BrM thickness using a previously described systematic method to quantitate ocular pathologies in mice using EM.²³ In brief, retinal sections were mounted on 400-nm mesh thin bar grids, and images were taken at every place where a grid line intersected the RPE. We counted the number of mitochondria per image and found significantly reduced mitochondria in Tmem135 TG RPE (Fig. 5B). Measurement of the mitochondrial surface area revealed significantly smaller mitochondria in Tmem135 TG RPE compared with WT controls (Fig. 5C). Because there were hyperpigmented RPE in Tmem135 TG eyes, we counted the number of pigment granules per image and saw a trend of increased pigment granules that did not reach statistical significance (P = 0.07) (Fig. 5D). The thickness of BrM was measured on each EM image, which was similar between WT and Tmem135 TG (Fig. 5E). Through our ultrastructural examination of Tmem135 TG RPE, we found the presence of basal laminar deposits, a pathological hallmark often seen in aged WT²⁴ and AMD mouse models,²⁵ in 10% of the Tmem135 TG RPE images (Fig. 5F). These data show overexpression of Tmem135 affects RPE ultrastructure in mice. 

Figure 5.

 

Ultrastructure of WT and Tmem135 TG RPE. (A) Representative EM images of 60-day-old WT and Tmem135 TG RPE. Mitochondria (M), pigment granules (P), and nucleus (N) are labeled. Mitochondrial surface area is outlined in yellow. BrM thickness is delineated by black arrows. Magnification = 15,000X. Scale bar = 800 nm. CC, choriocapillaris. Quantification of mitochondria number (B), mitochondrial surface area (C), pigment granule number (D), and BrM thickness (E) per EM image. We measured the surface area of 3057 mitochondria in WT RPE and 1730 mitochondria in Tmem135 TG RPE. There were fewer and smaller mitochondria in Tmem135 TG RPE at 60 days of age. All data are mean ± SEM. (F) Representative EM images of basal laminar deposits in Tmem135 TG mice. Basal laminar deposit is outlined in yellow. Some 10% of Tmem135 TG EM images had basal laminar deposits. Data from n = 3 WT (P60), n = 3 Tmem135 TG (P60). *P < 0.05 and ***P < 0.001, ANOVA. All data are mean ± SEM.

Visual Function Analysis of Tmem135 TG Mice

Although we did not observe photoreceptor degeneration, we sought to investigate whether photoreceptor activity was affected by the RPE pathologies in the Tmem135 TG mice. To examine photoreceptor activity, we measured the ERG responses of dark-adapted 60-, 180- and 365-day-old Tmem135 TG and WT mice. There were no changes in the scotopic a- or b-wave in Tmem135 TG and WT mice at 60 days (Figs. 6A, 6B) and 180 days (Figs. 6D, 6E) of age. We also detected no differences in the photopic a- and b-wave responses at 60 days of age in these genotypes (Supplementary Figs. S3A, S3B). At 180 days of age, although the photopic a-wave was not different (Supplementary Fig. S3C), there was a significant decrease in the photopic b-wave of Tmem135 TG mice relative to WT mice at the 10 cd.s/m² flash intensity (Supplementary Fig. S3D). In 365-day-old Tmem135 TG mice, the scotopic a-wave responses were significantly decreased at the 0.30, 1.00, and 30.00 cd.s/m² intensities (Fig. 6G), and the scotopic b-wave responses were significantly lowered at the 0.3 cd.s/m² through 3.0 cd.s/m² intensities (Fig. 6H). The photopic a-wave (Supplementary Fig. S3E) and b-wave (Supplementary Fig. S3F) did not differ between Tmem135 TG and WT mice at 365 days of age. The c-wave of Tmem135 TG mice were only significantly decreased at 365 days of age compared with age-matched control, but not earlier (Figs. 6C, 6F, 6I). We conclude that the visual function of Tmem135 TG mice remains generally intact past 180 days of age, and show decreased scotopic a-, b-, and c-wave responses by 365 days of age. 

Figure 6.

 

Visual function in WT and Tmem135 TG mice. Transcorneal ERG recordings for P60 (A–C), P180 (D–F), and P365 (G–I) WT and Tmem135 TG mice. No significant changes were noted in the scotopic a-, scotopic b-, and c-waves at P60 and P180 between these genotypes. There were significant decreases in the scotopic a- and b-wave at some light intensities, as well as the c-wave of P365 Tmem135 TG mice compared with WT mice. Data from n = 5 WT (P60), n = 10 Tmem135 TG (P60), n = 5 WT (P180), n = 3 Tmem135 TG (P180), n = 5 WT (P365), and n = 4 Tmem135 TG (P365). *P < 0.05, ANOVA. All data are mean ± SEM.

RPE and Photoreceptor Markers in Aged Tmem135 TG mice

To test whether there was photoreceptor degeneration along with RPE degeneration in Tmem135 TG mice at 365 days of age in which there are differences in ERG responses, we isolated RNA from separated eyecups and neural retinas from 365-day-old Tmem135 TG and WT mice for quantitative PCR (qPCR) measuring RPE markers (retinaldehyde binding protein 1 [Rblp1], bestrophin 1 [Best1], and potassium inwardly rectifying channel subfamily J member 13 [Kcnj13]) and photoreceptor markers (rhodopsin [Rho], opsin 1 short wave sensitive [Opn1 sw], opsin 1 medium wave sensitive [Opn1 mw], arrestin 3 [Arr3], G protein subunit alpha transducin 2 [Gnat2], elongation of very long-chain fatty acids-like 4 [Elovl4], adenosine triphosphatase [ATP] binding cassette subfamily A member 4 [Abca4], retinal degeneration slow [Rds], surface antigen protein 1 [Sag1], G protein subunit alpha transducin 1 [Gnat1], and retinal outer segment membrane protein 1 [Rom1]). There were significant decreases in Rblp1, Best1, and Kcnj13, but no changes in any of the photoreceptor markers (Fig. 7A). We confirm our qPCR results through Western blot analysis in which there were significant decreases in CRALBP in the eyecups (Fig. 7B) and no changes in RHO in the neural retinas (Fig. 7C) of 365-day-old Tmem135 TG mice. Our data suggest that there is no significant photoreceptor degeneration in TG mice at 365 days of age. 

Figure 7.

 

RPE and photoreceptor markers in 365-day-old WT and Tmem135 TG mice. (A) Quantitative PCR analysis of RPE and photoreceptor markers revealed decreased RPE but not photoreceptor markers. Values in Tmem135 TG mice are normalized against those of WT mice. n = 3 for each genotype. (B) Western blot analysis of eyecup lysates of 365-day-old WT and Tmem135 TG mice for cellular retinaldehyde-binding protein (CRALBP). N = 5. (C) Western blot analysis of neural retinal lysates of 365-day-old WT and Tmem135 TG mice for rhodopsin (RHO). N = 5. Succinate dehydrogenase A (SDHA) and tubulin (TUB) served as loading controls for eyecup and neural retina lysate blots, respectively. *P < 0.05 and **P < 0.01, ANOVA. All data are mean ± SD.

RPE in Tmem135 Mutant Mice

Given that the RPE of Tmem135 TG mice exhibited numerous abnormalities, we revisited RPE phenotypes in Tmem135^{FUN025/FUN025} mutant mice. Our previous study found photoreceptor degeneration, infiltration of immune cells into the subretinal space, RPE hypertrophy, and visual loss in Tmem135^{FUN025/FUN025} mice relative to WT mice.⁸ Age-dependent photoreceptor degeneration, subretinal immune cells, and RPE hypertrophy were observed in Tmem135^{FUN025/FUN025} mice compared with Tmem135^FUN025/+littermate controls at 45 and 365 days of age (Supplementary Fig. S4) consistent with our previous report.⁸ Photoreceptor degeneration in 365-day-old Tmem135^{FUN025/FUN025} translated into lowered scotopic ERG a-wave (Supplementary Fig. S5A) and b-wave (Supplementary Fig. S5B) responses, although the changes in the scotopic b-wave did not reach significance. The c-wave (Supplementary Fig. S5C), photopic a-wave (Supplementary Fig. S5D), and b-wave (Supplementary Fig. S5E) were significantly attenuated in Tmem135^{FUN025/FUN025} mice compared with Tmem135^FUN025/+mice. 

We evaluated RPE phenotypes of Tmem135^FUN025/+and Tmem135^{FUN025/FUN025} mice at 45 and 365 days of age. RPE flat-mount analysis showed no signs of RPE dysmorphogenesis in Tmem135^{FUN025/FUN025} mice (Fig. 8A). Quantification of the RPE flat mounts revealed decreased cell size (Fig. 8A) and increased cell density (Fig. 8B) in Tmem135^{FUN025/FUN025} mice compared with Tmem135^FUN025/+mice at 45 and 365 days of age. These findings show that the Tmem135 mutation affects RPE cells differently than the overexpression of Tmem135. 

Figure 8.

 

RPE size and density in Tmem135^FUN025/+and Tmem135^{FUN025/FUN025} mice. (A) Representative images of P45 and P365 Tmem135^FUN025/+ and Tmem135^{FUN025/FUN025} RPE flat mounts with ZO-1 in green and DAPI in blue. Magnification = 20X. Scale bar = 100 µm. (B) RPE cell size and (C) RPE cell density in Tmem135^FUN025/+ and Tmem135^{FUN025/FUN025} mice. The RPE cell size in P45 and P240 Tmem135^{FUN025/FUN025} mice were greater than Tmem135^FUN025/+control mice. The RPE cell density in P45 and P240 Tmem135^{FUN025/FUN025} mice were smaller than Tmem135^FUN025/+control mice. Data from n = 4 Tmem135^FUN025/+ (P45), n = 4 Tmem135^{FUN025/FUN025} (P45), n = 9 Tmem135^FUN025/+ (P365), and n = 10 Tmem135^{FUN025/FUN025} (P365). *P < 0.05 and ***P < 0.001, ANOVA. All data are mean ± SEM.

Mitochondrial Morphology in RPE

The RPE phenotypes in Tmem135 TG and Tmem135^{FUN025/FUN025} mice mirror the overly fragmented and fused mitochondrial phenotypes in primary-cultured mouse fibroblasts from these mice, respectively.⁸ To evaluate the mitochondrial morphologies in RPE cells, we isolated and cultured mouse RPE cells from 60-day-old WT, Tmem135 TG, and Tmem135^{FUN025/FUN025} mice. We verified our cell cultures for tight junctions and RPE protein expression (Supplementary Fig. S6). Using these cultures, we stained the RPE cells with Mitotracker (Thermo Fisher Scientific, Waltham, MA, USA) to visualize the mitochondria. We found normal mitochondria in WT RPE cells (Fig. 9A) but overly fragmented mitochondria in Tmem135 TG RPE cells (Fig. 9B) and overly fused mitochondria in Tmem135^{FUN025/FUN025} RPE cells (Fig. 9C). Therefore the RPE phenotypes in Tmem135 TG and Tmem135^{FUN025/FUN025} mice are associated with mitochondria morphological changes. 

Figure 9.

 

Mitochondrial morphology in primary RPE cultures. Representative images of mitochondria (red) and nucleus (blue) in RPE cells obtained from 60-day-old (A) WT, (B) Tmem135 TG, and (C) Tmem135^{FUN025/FUN025} mice. Note the fragmented mitochondria in the Tmem135 TG RPE cell, and elongated mitochondria in the Tmem135^{FUN025/FUN025} RPE cell. Magnification = 60X. Scale bar = 10 µm.

Oxidative Phosphorylation Proteins in Tmem135 TG and Tmem135^{FUN025/FUN025} Tissues

Mitochondria modify their morphology through mitochondrial dynamics to sustain energy homeostasis within cells.²⁶ To test whether there was altered mitochondrial energy production in Tmem135 TG and Tmem135^{FUN025/FUN025} mice, we evaluated the levels of proteins involved in mitochondrial oxidative phosphorylation (OXPHOS), the main source of cellular ATP.²⁷ Using a commercially available antibody cocktail, we examined the protein levels of ATP synthase subunit alpha (ATP5A), cytochrome b-c1 complex subunit 2 (UQCRC2), cytochrome c oxidase subunit 1 (MTCO1), succinate dehydrogenase subunit B (SDHB), and NADH dehydrogenase beta subcomplex subunit 8 (NDUFB8) in eyecup and neural retina lysates of 365-day-old WT, Tmem135 TG, Tmem135^FUN025/+, and Tmem135^{FUN025/FUN025} mice. We found no changes in OXPHOS proteins in Tmem135 TG eyecups versus WT littermates (Fig. 10A) and Tmem135^{FUN025/FUN025} eyecups relative to Tmem135^FUN025/+ (Fig. 10B). In the Tmem135 TG neural retinas, we measured significant decreases in ATP5A, UQCRC2, MTCO1, SDHB, and NDUFB8 compared with WT (Fig. 10C). In contrast, there were significant increases in ATP5A and UQCRC2 in the Tmem135^{FUN025/FUN025} neural retinas (Fig. 10D). As a positive control, we quantitated the level of these OXPHOS proteins in heart lysates because expression of OXPHOS genes is decreased in Tmem135 TG hearts.¹³ There were significant decreases in ATP5A, UQCRC2, MTCO1, and SDHB in Tmem135 TG hearts (Fig. 10E) but significant increases in ATP5A, UQCRC2, and NDUFB8 in Tmem135^{FUN025/FUN025} hearts (Fig. 10F). Our results indicate that Tmem135 overexpression and mutation leads to changes in mitochondrial OXPHOS. 

Figure 10.

 

Relative levels of OXPHOS proteins in 365-day-old Tmem135 TG and Tmem135^{FUN025/FUN025} tissues. Representative Western blots for ATP synthase subunit alpha (ATP5A), cytochrome b-c1 complex subunit 2 (UQCRC2), cytochrome c oxidase subunit 1 (MTCO1), succinate dehydrogenase subunit B (SDHB), and NADH dehydrogenase beta subcomplex subunit 8 (NDUFB8) using eyecup lysates (A,B), neural retina lysates (C,D) and heart lysates (E,F) for WT and Tmem135 TG (A,C,E) as well as Tmem135^FUN025/+ and Tmem135^{FUN025/FUN025} mice (B,D,F). We did not detect NDUFB8 on our Western blots containing eyecup lysates, which may be owing to the lower protein concentrations used for these blots. Immunobands were normalized to loading control. Beta-actin (ACTB), tubulin (TUB), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as loading controls for eyecup, neural retina, and heart blots, respectively. Data from Tmem135 TG were normalized to WT and Tmem135^{FUN025/FUN025} were normalized to Tmem135^FUN025/+. N = 5–7. All information on the samples used for these Western blots are included in Supplementary Tables S1 and S2. ND, not detected. *P < 0.05, **P < 0.01, and ***P < 0.001, ANOVA. All data are mean ± SD.

Discussion

Imbalanced mitochondrial dynamics are thought to accumulate dysfunctional mitochondria and contribute to aging and disease.²⁸ In support of this hypothesis, we identified Tmem135 as a gene important in regulating mitochondrial dynamics, maintaining the mitochondrial function, and inhibiting the formation of age-dependent retinal pathologies in mice.^8,13 Our current study furthers our understanding of TMEM135 biology by phenotyping the eyes of Tmem135 TG mice. In Tmem135 TG mice, the full-length murine Tmem135 cDNA is driven by a chicken beta-actin promoter, which results in overexpression of Tmem135 in the eyecup and neural retina at equal levels. However, we detect pathologies only in the RPE, including degeneration, vacuolization, migration, dysmorphogenesis, and cell enlargement. In contrast, the neural retina does not appear to be sensitive to the upregulation of Tmem135, as we did not observe any pathologies in the retinas of Tmem135 TG mice. The dampened scotopic ERG a- and b-waves in 365-day-old Tmem135 TG mice is more likely due to functional RPE deficits, as shown by significant reduction of c-wave response. Our qPCR analysis uncovered a decrease of both Kcnj13 and Best1 expression in Tmem135 TG eyecups but not in photoreceptor genes in Tmem135 TG neural retinas relative to WT tissues. Kcnj13 encodes Kir7.1 channel critical for the efflux of potassium ions and generation of the RPE-derived c-wave.²⁹ Hence a decrease in Kcnj13 would explain the reduced c-wave and scotopic ERG responses in Tmem135 TG mice because their RPE is unable to restore potassium ions in the subretinal space required for visual transduction.³⁰ However, the Tmem135^FUN025 mutation leads to contrasting changes in the RPE. Tmem135^FUN025 mutant mice have thicker, smaller, and denser RPE, whereas their photoreceptor cells progressively degenerate.⁸ Thus we conclude that the balance of functional TMEM135 is essential for RPE health. 

The RPE phenotypes of Tmem135 TG and Tmem135^FUN025 mutant mice correlate with disruptions in mitochondrial dynamics. We confirm smaller mitochondria in EM images of 60-day-old Tmem135 TG RPE that is in contrast with our previous study finding larger mitochondria in EM images of Tmem135^FUN025 mutant RPE.⁸ These results are consistent with the mitochondrial staining of primary-cultured RPE indicating overly fragmented mitochondria in Tmem135 TG RPE and overly fused mitochondria in Tmem135^FUN025 mutant RPE. Mitochondrial dynamics is essential for mitochondrial homeostasis and thought to contribute to tissue aging,³¹ although it is unknown whether mitochondrial dynamics affects RPE aging. It has been postulated that mitochondrial fission facilitates the removal of damaged mitochondrial membranes to preserve mitochondrial function.^32,33 Overfragmentation of mitochondria due to Tmem135 overexpression may result in mitochondrial dysfunction in the RPE, which in turn could lead to RPE degeneration. Consistent with this hypothesis, loss of mitochondrial function in the RPE either by ablating the mitochondrial transcription factor Tfam or mitochondrial biogenesis gene Pgc-1⍺ leads to RPE degeneration.^21,34 On the contrary, mitochondrial fusion is thought to intermix proteins and DNA between mitochondria under conditions of nutrient deprivation.³⁵ Enhancing mitochondrial fusion in Tmem135^FUN025 mutant mice may trigger increased nutrient intake and metabolic stress in the RPE that causes RPE thickening. In support of this idea, a nuclear magnetic resonance–based metabolomics study revealed an accumulation of metabolites linked to glucose, amino acid, and fat metabolic pathways in primary-cultured Tmem135^FUN025 mutant RPE cells compared with WT and Tmem135 TG RPE cells.²⁰ Furthermore, inducing metabolic stress through hypoxia³⁶ or oxidative stress³⁷ can cause thickening of the RPE in mice, similarly to what we observe in Tmem135^FUN025 mutant mice. RPE phenotypes of Tmem135 TG and Tmem135^FUN025 mutant mice align with other published reports of RPE phenotypes in mice with known mitochondrial dysfunction, a cellular consequence of aging.⁴ 

Although it is clear that TMEM135 is important in mitochondrial dynamics, how mitochondrial dynamics is regulated by TMEM135 at the molecular level is unknown. Although TMEM135 may function at mitochondria to regulate their dynamics by associating with fission/fusion machineries, it is also possible that TMEM135 is involved in a different mechanism at mitochondria, which influences mitochondrial dynamics. In this study, we found that Tmem135 overexpression and mutation lead to changes in the level of mitochondrial OXPHOS proteins, which could be related to TMEM135 function at mitochondria. It has been shown that increased OXPHOS stimulates mitochondrial fusion,³⁸ whereas inhibition of OXPHOS by drug treatments are associated with enhanced mitochondrial fission.³⁹ However, dysregulated mitochondrial dynamics can also lead to changes in OXPHOS,³⁹ which may be more likely in this case considering the following findings. We observed decreased OXPHOS proteins in Tmem135 TG neural retina and heart lysates, as well as increased OXPHOS proteins in Tmem135^FUN025 mutant neural retina and heart lysates, whereas we did not detect changes in OXPHOS proteins in the Tmem135 TG and Tmem135^FUN025 mutant eyecup (RPE/choroid/sclera). The fact that we observed mitochondrial shape/size changes in Tmem135 TG and mutant RPE, but not OXPHOS changes, points to the scenario that TMEM135 function is more closely related to mitochondrial dynamics, which in turn influences OXPHOS in some tissues but not in the RPE. However, it is also possible that the heterogeneity of the eyecup tissue that includes the RPE, choroid, sclera, and choroidal immune cells, masked OXPHOS changes in the RPE alone. It is interesting to note that no pathologies were observed in the neural retina of Tmem135 TG mice in which OXPHOS proteins were significantly decreased. This may be explained by the metabolic pathway photoreceptor cells use to meet their energy demands. Photoreceptor cells are known to largely depend on aerobic glycolysis as the source of ATP,⁴⁰ and therefore the decrease in OXPHOS proteins may have less effect on the energy demand of these cells. In contrast, it is well established that the RPE relies on OXPHOS to utilize lactate produced by the photoreceptors⁴¹ and fatty acids derived from photoreceptor outer segments.⁴² Thus if there is indeed a decrease in OXPHOS in the RPE of Tmem135 TG mice concealed by the tissue heterogenuity of the eyecup as discussed earlier, then such a change in OXPHOS may have detrimental effects on RPE cells. 

It is also important to note that the TMEM135 protein was not always localized on mitochondria in our previous report.⁸ TMEM135 was also observed as punctate signals in the cytosol, which may represent vesicle-like structures.⁸ This observation led us to hypothesize that TMEM135 may be trafficked between mitochondria and cytosolic vesicles. Peroxisomes are one of such cytosolic vesicles in which TMEM135 was previously localized^10,11 and may be involved in the potential mechanism through which TMEM135 regulates mitochondrial dynamics. Physical contacts between peroxisomes and mitochondria have been well studied,^43,44 and more importantly, a recent study indicated that mitochondrial fission could be regulated by peroxisomes.⁴⁵ Based on these notions, it is possible that TMEM135 functions in peroxisomes and its perturbation may, in turn, affect mitochondrial dynamics. Alternatively, TMEM135 may have multiple molecular functions, and RPE phenotypes we have observed may be in part caused by a defect in a different function/pathway. Maharjan et al.¹¹ reported that TMEM135 might be a regulator of Rab8-mediated primary ciliogenesis in cultured RPE1 cells. However, this function could also be connected with mitochondrial function, which has been shown to dictate the length of cilia.⁴⁶ TMEM135 could also play a role in lipid metabolism. TMEM135 increases in the hearts of mitochondrial very-long-chain acyl-CoA dehydrogenase (Vlcad) deficient mice that lack the ability to break down very-long-chain fatty acids through mitochondrial ß-oxidation.¹⁰ These mice show lipid accumulation in the heart, as well as cardiac pathologies, including collagen deposition, vacuolated myocytes, and abnormal mitochondrial morphology,⁴⁷ which are similar to heart pathologies in Tmem135 TG mice. Further molecular analysis may help us to understand the molecular function of TMEM135 in RPE cells. 

Structural changes in the RPE of Tmem135 TG mice, including pigmentary changes, migratory cells, dysmorphogenesis, degeneration, are often observed in patients with AMD,^48–50 which is a common and complex photoreceptor degenerative disease strongly conferred by advanced aging but influenced by genetics and environmental stressors.⁵¹ These RPE changes in the Tmem135 TG mice are also similar to those observed in previous AMD mouse models, including sodium iodate–injected C57BL/6J,^52,53 Dicer1 knockout,^22,54 RPE^CreER/DTA,⁵⁵ and Sod2 knockout mice.^56–58 Interestingly, mitochondrial fragmentation has been documented to occur in human AMD donor eyes.⁵⁹ It is possible that upregulated TMEM135 activity may be involved in this mitochondrial fragmentation. Thus it will be valuable to evaluate TMEM135 expression and localization in AMD donor eyes to investigate whether there is a link between TMEM135 and AMD in a future study. 

Conclusions

We have identified the pathogenic effects of Tmem135 overexpression on mouse RPE cells. These RPE pathologies contrast those in Tmem135^FUN025 mutant mice consisting of thicker, smaller, and denser RPE. These divergent RPE phenotypes correlate with contrasting mitochondrial sizes and morphologies (overfragmented vs. overfused) in RPE cells due to Tmem135 overexpression and mutation. This study demonstrated that TMEM135-regulated mitochondrial dynamics are critical for the health of RPE in mice. 

Acknowledgments

The authors thank Satoshi Kinoshita and the University of Wisconsin Translational Research Initiatives in Pathology laboratory (TRIP), supported by the University of Wisconsin Department of Pathology and Laboratory Medicine, UWCCC (P30 CA014520) and the Office of the Director, National Institutes of Health (S10OD023526) for the use of facilities and services, as well as Randall Massey and the University of Wisconsin Electron Microscope Core for tissue processing, sectioning, and assistance for this study. The authors also want to recognize Freya Mowat and Janis Eells for their advice and feedback on this work. 

Supported by grants NIH R01 EY022086, R01EY024995, Timothy William Trout Chairmanship, NIH T32EY027721, NIH P30 EY016665, M.D. Matthews Research Professorship, and an Unrestricted Grant from Research to Prevent Blindness, Inc. 

Disclosure: M. Landowski, None; S. Grindel, None; P.K. Shahi, None; A. Johnson, None; D. Western, None; A. Race, None; F. Shi, None; J. Benson, None; M. Gao, None; E. Santoirre, None; W.-H. Lee, None; S. Ikeda, None; B.R. Pattnaik, None; A. Ikeda, None 

Amended November 2, 2020: In the Western Blot Analysis section, the sentence “The primary antibodies used in this study were anti-CRALBP (#ab15051; Abcam, Cambridge, MA, USA), anti-Rhodopsin (#ab5417; Abcam), and total rodent OXPHOS Western blot antibody cocktail, not total rodent OXPHOS whole-blood antibody cocktail (#ab110413; Abcam)” has been changed to “The primary antibodies used in this study were anti-CRALBP (#ab15051; Abcam, Cambridge, MA, USA), anti-Rhodopsin (#ab5417; Abcam), and total rodent OXPHOS Western blot antibody cocktail (#ab110413; Abcam).”