DNA Methylation Dynamics in a Mouse Model of Retinitis Pigmentosa

Abstract

Retinitis pigmentosa (RP) is a group of sight-threatening genetic diseases characterized by progressive degeneration of photoreceptors, leading to severe vision loss from childhood to adulthood. With limited treatment options, understanding the molecular mechanisms underlying RP is crucial. Increased DNA methylation, especially in degenerating photoreceptors, is a contributing factor to retinal damage in RP. To exploit the molecular insights into methylation-driven pathways, this study investigated the DNA methylation patterns and their potential roles in photoreceptor degeneration in a mouse model of RP, specifically mice carrying a rhodopsin deficiency (Rho^–/–). Elevated levels of DNA methyltransferases (DNMTs) and DNA methylation were observed during photoreceptor degeneration. Importantly, weekly intravitreal injections of the pan DNMT inhibitor decitabine in Rho^–/– mice significantly improved photoreceptor morphology and visual function, as evidenced by electroretinogram, spectral-domain optical coherence tomography, and optomotor response-based visual behavior assays. Further histologic and immunohistochemical assessments revealed increased survival of cone photoreceptors and thicker outer nuclear layers in decitabine-treated mice compared with controls. Together, these findings revealed that the dynamics of DNA methylation correlate with photoreceptor degeneration. Inhibition of DNMTs mitigated the morphologic and functional impairments associated with the genetic defects in photoreceptors, suggesting that targeting DNA methylation could be a viable therapeutic strategy for neuroprotection in RP.

Retinitis pigmentosa (RP), encompassing a group of genetically heterogeneous conditions, is the leading cause of inherited retinal dystrophies worldwide. It is characterized by progressive degeneration of photoreceptor cells, leading to debilitating visual impairment and blindness in its advanced stages. More than 1.5 million patients are affected by RP worldwide.¹ The hallmark clinical features of RP typically manifest as night blindness, followed by gradual constriction of the visual field, decreased visual acuity, and, in some cases, eventual loss of central vision.

Despite decades of research and significant advancements in understanding the molecular mechanisms underlying RP, efficacious treatment options are yet to be discovered. Despite the significant burden placed by RP on affected individuals and their families, only one gene therapy, approved for patients with RP with RPE65 gene mutation, is currently available. However, these patients represent only 2% of cases of recessive RP and approximately 16% of Leber congenital amaurosis.² The majority of patients with RP have no option but to rely on supportive care, including low-vision aids and adaptive techniques. These interventions do not alter the natural course of the disease or halt its progression, representing a significant unmet clinical need to preserve existing vision and even improve visual function in patients with RP.³

Methylation of cytosine nucleotides is a covalent epigenetic modification of DNA usually associated with transcriptional silencing. In sensory neurons, including photoreceptors, DNA methylation is dynamically regulated during development,⁴ differentiation,⁵ and stress-induced death of sensory neurons.⁶ Improper DNA methylation is associated with developmental disorders of photoreceptors.⁷ DNA hypermethylation in photoreceptors contributes to the onset and progression of RP, occurring in degenerating photoreceptors without inheritable genetic mutations.⁸ How DNA methylation changes in RP or whether genetic mutations in photoreceptors influence DNA methylation patterns that, in turn, affect other genes involved in retinal survival, repair, and degeneration is not known. In addition, whether such methylation changes can affect disease progression is still unknown. This study was designed to study the roles of DNA methylation in photoreceptor cell degeneration associated with RP.

Because mutations of the Rhodopsin gene account for approximately 25% of dominant RP cases,⁹ and mice with Rhodopsin deficiency (Rho^–/–) develop a similar pattern of photoreceptor degeneration as patients with RP,¹⁰ Rho^–/– mice were used as a model of RP. Rho^–/– mice lack functional rhodopsin at birth and develop progressive cone photoreceptor degeneration, leading to complete blindness after postnatal week 13.¹¹ In the present study, DNA methylation changes in wild-type and Rho^–/– retinas were compared and the pan DNA methyltransferase (DNMT) inhibitor, decitabine,¹² was used to determine the functional roles of DNA methylation during photoreceptor degeneration. During the early stages of retinal development, DNMTs—including DNMT1, DNMT3A, and DNMT3B—are expressed at high levels.¹³ Although DNMT1 is present in most retinal cells, its expression pattern varies by cell type, with cones typically showing stronger nuclear staining than rods.¹⁴ Here, decitabine reduced the DNA methylation and photoreceptor cell loss, and improved visual function in Rho^–/– mice. The findings of this study offer promising new perspectives for protecting and restoring residual visual function in patients with RP.

Materials and Methods

Animals

Rhodopsin knockout (Rho^–/–) mice were generated in Humphries laboratory (Trinity College, Dublin, Ireland).¹⁵ Rho^–/– and age-matched C57BL/6J wild-type (Jackson Laboratory, Bar Harbor, ME; 000664) mice (aged 6 to 16 weeks) of both sexes were used in the experiments. All animal procedures were approved by the Institutional Animal Care and Use Committees of the Schepens Eye Research Institute of Mass Eye and Ear and adhered to the Association for Research in Vision and Ophthalmology (Rockville, MD) Statement for the Use of Animals in Ophthalmologic and Vision Research. Animals were kept in a 12-hour light-dark cycle with free access to food and water. Genotyping for Rho^–/– mice was performed using Transnetyx's outsourced PCR genotyping services (Transnetyx, Inc., Cordova, TN). At the experiment end point, mice were euthanized by CO₂ inhalation following the NIH Animal Research Advisory Committee Guidelines for Euthanasia of Rodents Using Carbon Dioxide.¹⁶

Human Donor Eyes and Histology

Human donor eyes were obtained from the Iowa Lions Eye Bank (Iowa City, IA) and were preserved within 8 hours of death after informed consent of the donor families. The use of human donor eyes conformed to the Declaration of Helsinki. Eyes were assigned to either dry age-related macular degeneration or age-matched control donors based on chart review. The age of donors was approximately 60 years.

Seven 6-mm punches centered or peripheral on the fovea of each donor were collected and fixed in 4% paraformaldehyde in 10 mmol/L phosphate-buffered saline, before sucrose cryoprotection and cryostat sectioning, using methods described previously.¹⁷

Intravitreal Injection in Mice

The procedure for intravitreal injection was essentially described.^18,19 Mice were anesthetized by i.p. injection of a mixture of ketamine (100 mg/kg; Ketaset; Fort Dodge Animal Health, Fort Dodge, IA) and xylazine (9 mg/kg; TranquiVed; Vedco, Inc., St. Joseph, MO) supplemented by topical application of proparacaine (0.5%; Bausch & Lomb, Tampa, FL). During intravitreal injection, a pulled microcapillary tube connected to a Hamilton syringe was inserted posterior to the limbus. Mice were randomly divided into two groups: one group received intravitreal injection of 1 μL phosphate-buffered saline (Sigma-Aldrich, St. Louis, MO) as controls, and the other group was injected with 1 μL of a decitabine solution (100 μmol/L; MedChem Express, Monmouth Junction, NJ) that was dissolved in phosphate-buffered saline. Weekly injection was started in 6-week–old Rho^–/– mice, before they exhibited visual function loss,¹⁰ and continued for 4 weeks.

Total Nucleic Acid Extraction

As previously described,²⁰ total RNA and DNA were extracted using the Quick-DNA/RNA Microprep Plus Kit (Zymo Research, Irvine, CA; D7005) and Zymo Quick-RNA MicroPrep Kit (Zymo Research; R1051). The quantity and quality of both nucleic acids were measured using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA). RNA was converted into cDNA with the PrimeScript RT Master Mix (Takara Bio, Kusatsu, Japan; RR036A) for quantification of mRNA levels using real-time quantitative PCR. Primers used for real-time quantitative PCR are listed in Table 1. DNA was processed for assessment of epigenetic modifications.

Table 1.

Primers Used for Real-Time Quantitative PCR in the Study

Target	Sequence
Mouse Dnmt1 forward	5′-GGACAAGGAGAATGCCATGAAGC-3′
Mouse Dnmt1 reverse	5′-TTACTCCGTCCAGTGCCACCAA-3′
Mouse Dnmt3a forward	5′-GAGGGAACTGAGACCCCAC-3′
Mouse Dnmt3a reverse	5′-CTGGAAGGTGAGTCTTGGCA-3′
Mouse Opn1sw forward	5′-TCTGCTACCTCCAAGTGTGGCT-3′
Mouse Opn1sw reverse	5′-GCAGTATGCGAAGACCATCACC-3′
Mouse Opn1mw forward	5′-GTCGCCATGTTTGTGCTCTGGA-3′
Mouse Opn1mw reverse	5′-GCTTGGAGTTGAAGCGGATGCT-3′
Mouse Gapdh forward	5′-CATCACTGCCACCCAGAAGACTG-3′
Mouse Gapdh reverse	5′-ATGCCAGTGAGCTTCCCGTTCAG-3′

Global DNA Methylation Assay

DNA methylation and demethylation levels were determined using the 5-methylcytosine (5-mC) DNA enzyme-linked immunosorbent assay kit (Zymo Research; D5325) and Quest 5-hydroxymethylcytosine (5-hmC) DNA ELISA Kit (Zymo Research; D5425), per the manufacturer's instruction individually, as described before.²¹ Absorbance was recorded at 405 nm with a microplate reader (Agilent, Santa Clara, CA; BioTek Synergy H1 Multimode Reader). The percentage of methylated cytosines (% 5-mC) and hydroxymethylated cytosines (% 5-hmC) in the total DNA content was determined by calculating the levels of 5-mC and 5-hmC using a standard curve generated with the kit controls.

Immunofluorescence Labeling and Quantification for Cone Survival

Mouse eyes were post-fixed in 4% paraformaldehyde at 4°C overnight, and then cryoprotected in 20% sucrose overnight at 4°C, as described before.²² Corneas and lens were removed from the eye before being embedded in OCT media (VWR, Radnor, PA; 25608-930) and frozen. The frozen tissue block was then sliced sagittally into serial cross-sections (10 μm thick), collected on Superfrost Plus Slides (Thermo Fisher Scientific, Portsmouth, NH), and stored at –20°C until processed.

Frozen mice or human sections were blocked with 5% bovine serum albumin (Sigma-Aldrich) and 1% Triton X-100 (MilliporeSigma, Burlington, MA) in phosphate-buffered saline for 1 hour at room temperature, followed by incubation with the primary antibodies overnight at 4°C and then the proper secondary antibodies for 2 hours at room temperature in the same blocking solution. Between changes of antibody incubation, all sections were washed three times, 5 minutes each. Staining for 5-mC, 5-hmC, DNMT1, and DNMT3A required 2N HCl treatment for 30 minutes at 37°C, followed by neutralization with 0.1 mol/L Tris-HCl (pH 8.3) for 10 minutes at room temperature before the blocking.²³ Immunostained samples were mounted with Fluoromount-G (Dako, Carpinteria, CA; S3023) and imaged using Leica DMi8 fluorescence microscope and Leica SP8 confocal microscope (Leica Microsystems, Wetzlar, Germany) with 20× and 40× lens, as described previously.¹⁰ Tile scans of four mice with three to four consecutive retinal sections per animal were imaged and used for quantification, where data from the scans were averaged for individuals. Antibodies used for staining are listed in Table 2. For quantification of cone density and outer nuclear layer (ONL) thickness, retinal sections containing the thickest ONL (thus crossing over the uneven degenerated and proliferated retinal regions) were used for the vehicle and decitabine group. The total number of cones, the thickness of cones, the total layer number of ONL (number of DAPI-positive nuclei row), and the thickness of ONL in each retinal section were recorded. The quantifications were performed by individuals (including L.H. and A.A.) blinded to the experimental conditions.

Table 2.

Details of Antibodies in the Study

Antibody	Host species	Dilution	Manufacturer	Catalog no.	Usage
DNMT1	Mouse	1:500 (IF), 1:1000 (WB)	Active Motif, Carlsbad, CA	39204	IF, WB
DNMT3A	Rabbit	1:500 (IF), 1:1000 (WB)	Cell Signaling Technology, Danvers, MA	3598S	IF, WB
β-Actin	Mouse	1:1000	Cell Signaling Technology	3700	WB
5-Methylcytosine	Mouse	1:500	Active Motif	39649	IF
5-Hydroxymethylcytosine	Rabbit	1:500	Active Motif	39791	IF
Alexa 488–conjugated peanut agglutinin	Mouse	1:200	Life Technologies, Carlsbad, CA	L21409	IF
DAPI and Hoechst Nucleic Acid Stains	/	1:1000	Thermo Scientific, Waltham, MA	62248	IF
AffiniPure Alexa Fluor 488 Anti-Rabbit IgG (H+L) F(ab')2∗	Donkey	1:1000	Jackson ImmunoResearch, West Grove, PA	711-545-152	IF
Cy3-conjugated AffiniPure Anti-Mouse IgG (H+L)∗	Donkey	1:1000	Jackson ImmunoResearch	715-165-150	IF
Horseradish Peroxidase–conjugated Goat Anti-Mouse IgG	Goat	1:4000	Proteintech, Rosemont, IL	SA00001-1-A	WB

DNMT, DNA methyltransferase; IF, immunofluorescence; WB, Western blot analysis.

^∗Secondary antibody.

Hematoxylin and Eosin Staining

Frozen sections of the eyecup were made at 10 μm and stained with hematoxylin and eosin. All slides were examined using NanoZoomer 2.0-HT (Hamamatsu Photonics, Hamamatsu, Japan). The thickness of the ONL was measured using ImageJ software version 1.53t (NIH, Bethesda, MD; https://imagej.nih.gov/ij). The quantifications were performed by individuals blinded to the experimental conditions (including L.H. and A.A.).

TUNEL Assay

Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay was performed with the in situ cell death detection kit, tetramethylrhodamine red terminal deoxynucleotidyl transferase dUTP nick-end labeling (Roche, Mannheim, Germany). The percentage of apoptotic cells was determined using Leica DMi8 fluorescence microscope and Leica SP8 confocal microscope (Leica Microsystems, Wetzlar, Germany) with 40× lens, as described before.¹⁰ On each coverslip, the total number of cells, identified by positive staining with DAPI (blue), and the number of cells that were TUNEL positive were counted. Four fields per coverslip were counted. The data were expressed as the percentage of total cells that were apoptotic (n = 3 to 4).

Western Blot Analysis

To assess the protein expression levels of DNA methylation-associated factors in the retina, mice were euthanized and retinas were isolated immediately and homogenized in radioimmunoprecipitation assay lysis buffer with protease inhibitors (1% phosphatase inhibitor cocktail and 1 mmol/L phenylmethylsulfonyl fluoride). Then centrifuged and collected supernatants were subject to Western blot analysis. Proteins were loaded in SDS-PAGE gels (10%, 10 μg per lane) and then transferred to polyvinylidene difluoride membrane (MilliporeSigma). The membrane was blocked with 5% nonfat dry milk in Tris-buffered saline with 0.1% Tween-20 (pH = 7.4) for 2 hours at room temperature and incubated overnight with primary antibodies at 4°C. Then, the membrane was incubated with the appropriate peroxidase-linked secondary antibodies (Proteintech, Rosemont, IL) for 2 hours at room temperature. Antibodies used for Western blot analysis are listed in Table 2. Immunoblots were examined using an electrochemiluminescence detection reagent (Epizyme Biotech, Cambridge, MA), and images were captured and analyzed with ChemiDoc Touch Imaging System with Image Lab Touch Software version 2.3 (Bio-Rad, Hercules, CA).

OCT Imaging

Spectral-domain optical coherence tomography (OCT) was performed using a Bioptigen Spectral Domain Ophthalmic Imaging System (Bioptigen, Inc., Durham, NC), as previously described.¹⁰ The system has a platform designed for easy orientation and aligning of mice for retinal imaging and provides a high resolution of 2 μm. The mice were anesthetized by i.p. injection with a mixed solution of ketamine and xylazine. The pupils were dilated with 1% tropicamide eye drops (Bausch & Lomb Inc., Vaughan, ON, Canada) before imaging. Radial volume scan (centered on the optic disc, consisting of 100 B-scans), and all layer thickness measurements were acquired using image analysis software (InVivoVue Clinic version 2.4; Bioptigen, Inc.). ONL thickness is an indicator of photoreceptor survival.

Electroretinography

The electroretinography (ERG) was performed as previously reported.¹⁰ For photopic ERG recording, mice were light adapted for 7 minutes before the experiment started. Mice were anesthetized with ketamine/xylazine (100 mg/kg; Dechra Vet Products, Overland Park, KS, 383017-01; 20 mg/kg; Covetrus North America, Dublin, OH, 1XYL006), and the pupils were dilated using 1% tropicamide eye drops. Mice were kept on a built-in warming pad in the Ganzfeld ColorDome (ColorDome LabCradle mouse ERG testing; Diagnosys LLC, Lowell, MA) to prevent hypothermia. Gold wire ring electrodes were used for ERG recording, positioned on each cornea with lubrication by GenTeal gel (Novartis, Basel, Switzerland) for optimal contact. Silver needle electrodes served as reference and ground, inserted subcutaneously between the eyes near the base of the skull and at the base of the tail, respectively. Single-flash stimuli were presented with a light intensity of 3.0 cds/m². Responses were recorded by the ERG system (Espion Electroretinography System; Diagnosys LLC), and the b-wave amplitude was measured from the cornea-negative peak to the first major cornea-positive peak after oscillatory potentials.

Optomotor Response–Based Visual Tests

The optomotor response–based spatial frequency threshold test was built as previously described to measure the visual acuity and contrast sensitivity of mice.¹⁰ Briefly, it consisted of a pedestal in the center of an area surrounded by four identical, connected liquid crystal display screens (15 inch; Acer Inc., New Taipei City, Taiwan). Mice were placed on the central pedestal and allowed to move freely while exposed to optomotor stimuli displayed on the monitors. The optomotor stimulus consisted of vertical black and white bars in a grating pattern that rotated clockwise (for the right eye test) or counterclockwise (for the left eye test) at a constant speed (12°/second). Each eye was tested separately based on the direction of the rotating stripes. A positive optomotor response was defined by mouse head tracking behavior, a smooth reflexive head movement with a velocity and direction in concert with the rotating bars, recorded and verified by two experienced observers blinded to the treatment group (L.H. and W.L.T.), and a minimum of three instances of head tracking observed by both observers was required to conclude that the animal could detect the stimulus. The width and contrast of the bars were varied to test how fine the mouse could see, and the corresponding bar width was calculated in terms of cycle/degree where 0.4 to 0.5 cycle/degree reflects normal visual acuity in the mouse. Mice with hemorrhage or inflammation after surgical procedures were excluded from the analysis. The exclusion criteria were established before the experiment.

Statistical Analysis

The statistical analysis in this study was conducted using GraphPad Prism version 5 (GraphPad Software Inc., San Diego, CA). The Mann-Whitney rank-sum test was used for categorical variables or non-normally distributed data or data without same SD. For the quantitative figures, each dot represents one mouse. Analysis was performed using GraphPad Prism software version 5. P < 0.05 was considered as a statistically significant difference. In the figures, statistical significance is indicated as follows: ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Not significant was not labeled.

Results

Dynamics of DNMT Expression and DNA Methylation during Photoreceptor Degeneration

To determine whether Rhodopsin deficiency alters DNA methylation patterns and/or the expression of DNMTs, mRNA levels of DNMTs in the retinas of 12-week–old Rho^–/– and age-matched wild-type mice were compared by real-time quantitative PCR. This time point was chosen to ensure capturing the effects on DNMTs resulting from the loss of rhodopsin, a key photoreceptor gene. Four DNA methyltransferases—DNMT1, DNMT3A, DNMT3B, and DNMT3L—have been identified as critical enzymes establishing DNA methylation patterns that dynamically regulate chromatin remodeling and gene expression in the retina.^24,25 Expressions of Dnmt1, Dnmt3a, and Dnmt3b in the mouse retina from embryonic day 10.5 through 10 months of age indicate that all three genes are highly expressed during early retinal differentiation, with expression levels markedly reduced after birth.¹⁴ Here, significant up-regulation of Dnmt1 and Dnmt3a was detected in the retina of Rho^–/– mice compared with that in wild-type mice (Figure 1, A and B), supporting that Rhodopsin deficiency drives changes in DNA methylation enzymes. Expressions of Dnmt3b and Dnmt3l were below a detectable level in the retina (data not shown).

Figure 1.

Rhodopsin deficiency drives DNA methyltransferase (DNMT) expression and DNA methylation changes in the retina. A and B: Results of real-time quantitative PCR quantification of Dnmt1 and Dnmt3a mRNA levels in the retinas of 12-week–old Rho^–/– age-matched wild-type (WT) mice. Data were normalized to the internal control of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) levels (rank-sum test was used). C–E: Quantification of levels of global DNA methylation [5-methylcytosine (5-mC) %; C], demethylation [5-hydroxymethylcytosine (5-hmC) %; D], and the ratio of 5-mC/5-hmC (E) in retinas of Rho^–/– mice compared with the age-matched WT mice (rank-sum test was used). F: Photomicrographs showing DNMT1 (red) and DNMT3A (green) immunolabeling in the retinas of 12-week–old Rho^–/– and aged-matched WT mice. Retinal sections were counterstained with a nuclei marker DAPI (blue). G: Photomicrographs of 5-mC (red) and 5-hmC (green) labeling in the retinas of Rho^–/– mice and aged-matched WT mice. Retinal sections were counterstained with a nuclear marker DAPI (blue). H–J: Western immunoblot images (H) and quantitative analysis (I and J) of DNMT1 and DNMT3A proteins in the retinas of 12-week–old Rho^–/– age-matched WT mice. Equal sample loading was confirmed by immunostaining against β-actin (rank-sum test was used). n = 3 to 7 mice per group (A and B); n = 6 mice per group (C–E); n = 4 to 10 mice per group (H–J). ∗P < 0.05, ∗∗P < 0.01. Scale bars = 100 μm (F and G). INL, inner nuclear layer; ONL, outer nuclear layer.

To determine whether the observed increases in Dnmt1 and Dnmt3a expression were associated with DNA methylation changes in the retina, the retinal DNA methylation state was analyzed using immunodetection of 5-mC and 5-hmC. These two most common cytosine modifications in DNA are key markers of DNA methylation, with 5-mC typically associated with gene repression and 5-hmC often representing an intermediate stage in the DNA demethylation process.²⁶ Significant increases of 5-mC (Figure 1C), and decreases of 5-hmC (Figure 1, D and E) were observed in the retinas of Rho^–/– mice compared with those of wild-type mice, indicating dynamic regulation of DNA cytosine methylation. These findings were verified by subsequent immunofluorescence tests (Figure 1, F and G) and Western immunoblot assays (Figure 1, H–J).

To determine the association between photoreceptor degeneration and DNA methylation, human retinas were collected and analyzed. Loss of photoreceptors and rhodopsin content is observed during the progression of age-related macular degeneration, a prevalent retinal degenerated disease.²⁷ This was the impetus for examining the DNA methylation state in human dry age-related macular degeneration retinas. Significant increases of 5-mC/5-hmC were observed in the ONL of patients with dry age-related macular degeneration compared with those in healthy human retinas (Supplemental Figure S1), further supporting the association between DNA methylation and photoreceptor degeneration. The results suggest a potential involvement of DNA methylation in mediating photoreceptor survival and/or function.

Attenuation of Photoreceptor Cell Loss in Rho^–/– Mice by Inhibition of DNMT Function with Decitabine

To determine the functional significance of Rhodopsin deficiency–driven increases in DNA methylation in photoreceptor degeneration, 6-week–old Rho^–/– mice were treated, before the onset of cone photoreceptor cells, with a pan DNMT inhibitor, decitabine. Between postnatal weeks 4 and 6, Rho^–/– mice displayed normal cone ERG responses but were devoid of rod activity due to Rhodopsin deficiency. By week 7, cone degeneration began as a result of rod photoreceptor degeneration.¹¹ Hence, 6-week–old Rho^–/– mice were selected to start the decitabine treatment. Down-regulation of retinal DNA methylation following decitabine injection was confirmed by 5-mC and 5-hmC immunostaining in retinal sections (Figure 2, A and B) and immunodetection by enzyme-linked immunosorbent assay in DNA extracted from retina lysis (Figure 2, C–E). Drastic decrease of 5-mC staining and increase of 5-hmC were shown in the retina, particularly in the ONL. The expression pattern of 5-mC was concentrated in the central heterochromatin of photoreceptor nuclei, a finding consistent with previous reports.²⁸

Figure 2.

Decitabine treatment rescues DNA methylation changes in the retina. Photomicrographs (A) and quantitative analysis (B) of 5-methylcytosine (5-mC; red) and 5-hydroxymethylcytosine (5-hmC; green)–immunolabeled retinal sections taken from Rho^–/– mice subjected to decitabine (DB) or phosphate-buffered saline (PBS) vehicle (Veh) treatment. Retinal sections were counterstained with a nuclear marker DAPI (blue). The t-test was used. Quantification of levels of global DNA methylation (5-mC%; C), demethylation (5-hmC%; D), and the ratio of 5-mC/5-hmC (E) in retinas of Rho^–/– mice subjected to decitabine or PBS vehicle treatment (absence of symbol indicates no statistical significance between groups by rank-sum test). n = 4 to 10 mice per group (A and B); n = 3 to 5 mice per group (C–E). ∗P < 0.05. Scale bar = 50 μm (A). INL, inner nuclear layer; ONL, outer nuclear layer.

To determine the effect of decitabine on photoreceptor cell survival, the thickness changes of cones and the retina was monitored by OCT in the decitabine-treated and control groups. Before treatment, no detectable differences were observed between the two groups, as assessed by B-scan OCT (Supplemental Figure S2, A–F). By 4 weeks after treatment, OCT images revealed an apparent thicker ONL, implicating improved photoreceptor survival, in decitabine-injected Rho^–/– mice compared with vehicle-injected controls (Figure 3A). Measurements of the maximal thickness of the ONL outer plexiform layer–ONL and the total retina confirmed this observation (Supplemental Figure S3). Quantification of B-scans from OCT images further demonstrated that decitabine treatment ameliorated the thinning of ONL, outer plexiform layer, and total retina thicknesses in Rho^–/– mice (Figure 3, B–D). The increased thickness of ONL in decitabine-treated Rho^–/– mice was further validated by hematoxylin and eosin staining (Figure 3, E and F). Additionally, the thickness of other layers were also measured as assessed by B-scan OCT (Supplemental Figure S4). To clarify whether the increased thickness attributed to the neuroprotective effect of decitabine on photoreceptors, photoreceptor cell death was assessed using TUNEL assay. In Rho^–/– mice, TUNEL-positive cells were predominantly observed in the ONL. Decitabine administration significantly reduced the TUNEL-positive cells compared with controls even at 4 weeks after initial treatment (Figure 3, G and H). These structural and cell death measurements indicate the neuroprotective effect of decitabine on photoreceptors.

Figure 3.

Decitabine preserves the retinal thickness in Rho^–/– mice. A: Photomicrographs of cross-sectional retinal B-scan images showing segmentation of optical coherence tomography images from Rho^–/– mice treated with decitabine (DB) and vehicle (Veh) controls. Yellow dotted line highlights the outer nuclear layer (ONL). Lines indicate the ONL [between the outer plexiform layer (OPL) and the external limiting membrane]. B–D: Decitabine prevented the thinning of the ONL (B), total retina (C), and OPL-ONL (D) of Rho^–/– mice (rank-sum test was used; the maximum thickness was calculated automatically using InVivoVue software based on data collected at 121 points in the B-scan images). E and F: Photomicrographs (E) and ONL thickness quantitative analysis (F) of hematoxylin and eosin (H&E)–stained retinal sections taken from Rho^–/– mice subjected to decitabine or phosphate-buffered saline (PBS) vehicle treatment (t-test was used). G and H: Photomicrographs (G) and quantitative analysis (H) for terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL; red) immunolabeled retinal sections taken from Rho^–/– mice subjected to decitabine or PBS vehicle treatment. Retinal sections were counterstained with a nuclear marker DAPI (blue). The t-test was used. n = 6 to 9 mice per group (B–D); n = 3 to 4 mice per group (E–H). ∗P < 0.05, ∗∗P < 0.01. Scale bars: 50 μm (A and G); 20 μm (E). GCC, ganglion cell complex; INL, inner nuclear layer.

The photoreceptor survival and function were then measured in retinal sections of control and decitabine-treated Rho^–/– mice. Retinal sections were stained for the cone photoreceptor marker peanut agglutinin, which labels the extracellular matrix surrounding cone photoreceptors (Figure 4A). Rhodopsin deficiency induced progressive loss of rod photoreceptor cells, as measured by the ONL thickness and/or the number of ONL cell layers. By 4 weeks after decitabine injection, a significant reduction of the ONL thinning was detected compared with that in phosphate-buffered saline–treated controls (Figure 4, B and C). Along with it, there were significant increases in cone photoreceptor cell counts and the lengths of the cone outer segments that were measured by the peanut agglutinin–stained proximal-to-distal axis (Figure 4, D and E). The increases in the mRNA levels of cone opsin expression, Opn1sw (Figure 4, F and G), further supported the protection of cone photoreceptors. From the data, decitabine treatment did not stir evident change to the Opn1mw gene expression, which codes for the medium-wavelength cone opsin. The data indicate the neuroprotective potential of inhibiting retinal DNA methylation by decitabine in RP.

Figure 4.

Inhibition of DNA methyltransferases by decitabine (DB) attenuates photoreceptor degeneration in Rho^–/– mice. A: Photomicrographs of peanut agglutinin–immunolabeled retinal sections taken from Rho^–/– mice subjected to decitabine or phosphate-buffered saline vehicle (Veh) treatment. Retinal sections were counterstained with a nuclear marker DAPI (blue). The zoom-in images were posted right below the originals. Brackets indicated the thickness of outer nuclear layer (ONL). B and C: Significant improvement in ONL thickness (B) and the number of cell layers (C) in the retinas of decitabine-treated Rho^–/– mice compared with the vehicle-treated controls (rank-sum test was used). D and E: Significant increases in the number of cones (D) and the length of cone outer segment (E) in the retinas of decitabine-treated Rho^–/– mice compared with the vehicle-treated controls (rank-sum test was used). F and G: Results of real-time quantitative PCR showing significantly increased expression of Opn1sw mRNA in the retinas of decitabine-treated Rho^–/– mice compared with the vehicle-treated controls. Data were normalized to the intensity of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (absence of symbol indicates no statistical significance between the groups by rank-sum test). n = 4 mice per group (B–E); n = 6 to 13 mice/group (F and G). ∗P < 0.05. Scale bar = 100 μm (A). INL, inner nuclear layer.

Rescue of Retinal and Visual Functions in Rho^–/– Mice by Administration of Decitabine

To investigate whether the improved survival of photoreceptor cells, particularly cones, resulted in enhanced function in Rho^–/– mice, ERG was performed to electrophysiologically record photoreceptor function. Rho^–/– mice do not have functional rods, so the focus was on assessing photopic ERGs that measure cone functions. At baseline, the decitabine group and the control group displayed no significant differences in visual function (Supplemental Figure S2, G–K). At 4 weeks after treatment, decitabine-treated Rho^–/– mice showed evident increases in the amplitude of photopic 600, m-cone, and s-cone (Figure 5, A–F). To determine whether enhanced cone function is translatable to visual perception, mouse visual acuity and contrast sensitivity were also examined using optomotor response–based assays. Decitabine treatment preserved vision in Rho^–/– mice as shown by their improved visual acuity and contrast sensitivity (Figure 5, G-I). This indicated that DNMT inhibition through pharmacologic tools empowers structural and functional cone neuroprotection in adult RP mice.

Figure 5.

Inhibition of DNA methyltransferases by decitabine (DB) attenuates visual function defect in Rho^–/– mice before detectable visual function defect occurs. Representative electroretinography (ERG) waveforms for photopic 600 (A), M-cone (B), and S-cone (C) taken from Rho^–/– mice before and after treatment with decitabine and vehicle (Veh) controls. ERG B-wave amplitudes for photopic 600 (D), M-cone (E), and S-cone (F) (rank-sum test was used). Schematic of the optomotor response (OMR; G) and assessment of visual acuity (H) and contrast sensitivity (I) by OMR assay in Rho^–/– mice treated with decitabine and vehicle controls (rank-sum test was used). n = 8 mice per group (D–F); n = 12 mice per group (G–I). ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001, and ∗∗∗∗P < 0.0001.

Discussion

The current data reveal an epigenetic component or DNA methylation in the pathogenesis and propagation of photoreceptor degeneration in RP. The observed increase in DNMT expression following rhodopsin deficiency can be attributed to secondary effects arising from photoreceptor degeneration. Specifically, when rhodopsin loss initiates degenerative cascades, this causes widespread disruption in retinal cellular homeostasis that is not limited to photoreceptors, leading to perturbation of DNA methylation patterns. We propose that this homeostatic disruption potentially triggers compensatory responses and changes in DNMT expression levels, as observed in the study. These results provide a possible mechanistic framework in which photoreceptor gene mutations disrupt the pattern of DNA methylation, leading to deleterious effects on photoreceptors and resulting in debilitating visual impairment. Notably, DNMT inhibition is a promising strategy to restore DNA methylation homeostasis and provide neuroprotection for cone photoreceptors. Therefore, maintaining the homeostasis of DNA methylation represents a potential therapeutic avenue for photoreceptor neuroprotection in retinitis pigmentosa.

The Rho^–/– mouse model targets the rhodopsin gene,^29,30 mutations of which are involved in 25% of dominant RP.⁹ This model exhibits a pattern of photoreceptor degeneration similar to that observed in patients with RP, including bone-spicule pigmentation in the fundus, marked thinning of the retina as detected by OCT, significant degeneration of the retinal neurons and photoreceptors, and functional impairment of cone photoreceptors as determined by ERG.^1,10,31 These characteristics make the Rho^–/– mouse model a valuable platform for translational research on human retinal degeneration. RP and its association with DNA methylation has been studied using the rd1 and rd2 gene-specific models, and the methyl methanesulfonate–induced photoreceptor injury model.32, 33, 34 The distinct causative gene of Rho^–/– model provides valuable tools for advancing the understanding of RP pathogenesis and evaluating targeted treatments of RP. The research explores the use of decitabine as a novel approach to address heritable gene defects leading to RP. Notably, rhodopsin deficiency affects the DNA methylation states of not only photoreceptors but also retinal ganglion cells and the inner nuclear layer in Rho^–/– mice. This effect likely arises from the extensive network of epigenetic regulation within the retina.35, 36, 37, 38 When photoreceptors degenerate, they induce cellular stress signals throughout the retinal structure, affecting the photoreceptors themselves but also neighboring cells, including the retinal ganglion cells and inner nuclear layer.³⁹ This damage likely triggers adaptive changes in DNA methylation as a cellular response to injury, aimed at preserving remaining cellular function.³⁹ Moreover, the loss of photoreceptors disrupts normal intercellular communication and signaling processes. Such disruptions can alter the epigenetic landscape of adjacent cells, including those in the retinal ganglion cells and inner nuclear layer. These cells might alter their DNA methylation patterns in response to signals from degenerating photoreceptors or as part of compensatory mechanisms to maintain overall retinal integrity and function.⁴⁰ As photoreceptor loss progresses, these changes become more pronounced, leading to further epigenetic dysregulation. By targeting disease-specific epigenetic changes, the research supports the development of tailored therapies that could potentially halt the progression of RP in Rho^–/– mice, as demonstrated by the improved morphologic condition of both the outer segments and ONL.

The present studies have important caveats to consider. First, data obtained from the mouse model may not fully reflect the changes occurring in human patients. Second, the specific genes most affected or hypermethylated in association with rhodopsin deficiency remain unidentified. Although a correlation between DNA methylation changes and photoreceptor degeneration has been observed, establishing a causal role remains a challenge. Moreover, although decitabine has shown promise in short-term studies, its long-term safety and efficacy require further investigation. These limitations highlight the urgent need for more comprehensive studies to elucidate the underlying mechanisms and develop effective treatments. Despite these challenges, the present study opens a promising avenue for neuroprotection and the potential treatment of photoreceptor degeneration.

Disclosure Statement

None declared.

Supplemental Data

Supplemental Figure S1

DNA methylation changes in the human degenerated retina. A: Photomicrographs of 5-methylcytosine (5-mC; red) and 5-hydroxymethylcytosine (5-hmC; green) labeling in the retinas obtained from a male without known age-related macular degeneration (AMD) and another male with dry AMD. Retinal sections were counterstained with a nuclear marker DAPI (blue). B: Quantification of ratio of relative fluorescence of 5-mC/5-hmC per DAPI-positive nucleus in retinas of human patient with dry AMD compared to human without known AMD. n = 7 areas per retina. ∗∗∗P < 0.001 by Mann-Whitney rank-sum test Scale bar = 50 μm (A). INL, inner nuclear layer; ONL, outer nuclear layer.

Supplemental Figure S2

Baseline electroretinography (ERG) and visual function assessments in 6-week–old Rho^–/– mice before treatment. The baseline thickness of the outer nuclear layer (ONL; A), total retina (B), and outer plexiform layer (OPL)–ONL (C) of Rho^–/– mice before the start of treatment (no statistical significance was detected between groups by rank-sum test; the thickness was calculated automatically using InVivoVue Software based on data collected at 121 points in the B-scan images). The maximal thickness of ONL (D), total retina (E), and OPL-ONL (F) in Rho^–/– mice before the start of treatment (no statistical significance was detected between groups by rank-sum test; the maximum thickness was calculated automatically using InVivoVue software based on data collected at 121 points in the B-scan images). ERG B-wave amplitudes of photopic 600 (G), M-cone (H), and S-cone (I) of Rho^–/– mice before the start of treatment (no statistical significance was detected between groups by rank-sum test). Visual acuity (J) and contrast sensitivity (K) assessed by optomotor response (OMR) assay in Rho^–/– mice before the start of treatment (no statistical significance was detected between groups by rank-sum test). n = 6 to 9 mice per group (A–F); n = 8 to 9 mice per group (G–I); n = 12 mice per group (J and K). DB, decitabine; OCT, optical coherence tomography; Veh, vehicle.

Supplemental Figure S3

Improved maximal retinal thickness in Rho^–/– mice following decitabine (DB) treatment. The maximal thickness of outer nuclear layer (ONL; A), total retina (B), and outer plexiform layer (OPL)––ONL (C) in control and decitabine-treated Rho^–/– mice (the maximum thickness was calculated automatically using InVivoVue software based on data collected at 121 points in the B-scan images). n = 6 to 9 mice per group (A–C). ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001 by Mann-Whitney rank-sum test. Veh, vehicle.

Supplemental Figure S4

Decitabine (DB) does not induce detectable differences in the averge thickness (A–C) or maximum thickness (D–F) of retinal nerve fiber layer (RNFL; A and D), inner plexiform layer (IPL; B and E), and inner nuclear layer (INL; C and F) thicknesses in Rho^–/– mice. (no statistical significance was detected between groups by rank-sum test; the average and maximum thickness was calculated automatically using InVivoVue software based on data collected at 121 points in the B-scan images). n = 6 to 9 mice per group (A–F). OCT, optical coherence tomography; Veh, vehicle.