Cat LCA-CRX Model, Homozygous for an Antimorphic Mutation Has a Unique Phenotype

Abstract

Purpose: Mutations in the CRX transcription factor are associated with dominant retinopathies often with more severe macular changes. The CRX-mutant cat (Rdy-A182d2) is the only animal model with the equivalent of the critical retinal region for high-acuity vision, the macula. Heterozygous cats (CRX^Rdy/+) have a severe phenotype modeling Leber congenital amaurosis. This study reports the distinct ocular phenotype of homozygous cats (CRX^Rdy/Rdy).

Methods: Gene expression changes were assessed at both mRNA and protein levels. Changes in globe morphology and retinal structure were analyzed.

Results: CRX^Rdy/Rdy cats had high levels of mutant CRX mRNA and protein. The expression of photoreceptor target genes was severely impaired although there were variable effects on the expression of other transcription factors. The photoreceptor cells remained immature and failed to elaborate outer segments consistent with the lack of retinal function. The retinal layers displayed a progressive remodeling with cell loss but maintained overall retinal thickness due to gliosis. Rapid photoreceptor loss largely occurred in the macula-equivalent retinal region. The homozygous cats developed markedly increased ocular globe length.

Conclusions: The phenotype of CRX^Rdy/Rdy cats was more severe compared to CRX^Rdy/+ cats by several metrics.

Translational Relevance: The CRX-mutant cat is the only model for CRX-retinopathies with a macula-equivalent region. A prominent feature of the CRX^Rdy/Rdy cat phenotype not detectable in homozygous mouse models was the rapid degeneration of the macula-equivalent retinal region highlighting the value of this large animal model and its future importance in the testing of translational therapies aiming to restore vision.

Introduction

Cone-rod homeobox (CRX) is a transcription factor essential for normal photoreceptor development, function, and survival.^1–3 The N-terminus of the protein contains a homeodomain for DNA binding whereas the C-terminus contains WSP and Orthodenticle Homeobox 2 (OTX)-tail domains and controls gene transactivation. CRX directly regulates many genes that are necessary for retinal function, such as those involved in the phototransduction, and the visual cycle are directly regulated by CRX.^4–9 

In humans, CRX mutations result in a spectrum of mostly dominant retinopathies with variable severity, ranging from severe childhood-onset forms, such as Leber congenital amaurosis (LCA7), to adult-onset forms such as cone–rod dystrophies, retinitis pigmentosa, and macular dystrophies.^10–13 LCA represents approximately 5% of all human inherited retinopathies with a prevalence of 1 in 30,000 to 81,000 newborns.^14,15 CRX mutations accounts for approximately 2.35% of the cases of LCA.¹⁶ 

CRX mutations have been categorized into four different classes by the proposed molecular disease mechanism.¹⁷ From Tran et al.¹⁷, class I mutations are hypomorphic with reduced DNA binding; class II are antimorphic mutations with variable DNA binding; class III are antimorphic frameshift/nonsense mutations with intact DNA binding and class IV are antimorphic mutations with reduced DNA binding. Some mutations, such as class I mutations that are null or hypomorphic mutations or those that ablate DNA binding, when present in the heterozygous state, result in a mild phenotype (such as a macular dystrophy or mild foveal abnormalities) or no phenotype, whereas patients who are homozygous for the same mutations have a severe (LCA) phenotype.^18,19 Similarly, mice heterozygous for a knockout mutation in Crx have a mild phenotype with a slight delay in photoreceptor maturation, whereas mice homozygous for the null mutation have a severe phenotype, equivalent to LCA, with a lack of photoreceptor outer segment development.²⁰ 

The severe dominant LCA phenotypes in patients, and the equivalent in animal models, are associated with expression of a mutant protein with retained DNA binding that has an antimorphic effect.^9,17 As stated above, class III mutations are antimorphic frameshift/non-sense mutations that escape nonsense-mediated decay and produce proteins that have intact DNA binding but lack transactivation activity. In heterozygous animals, this dominant-negative CRX protein interferes with the function of the WT allele, resulting in a more severe dominant phenotype. Three animal models are available for this class of mutations: a knock-in mouse model, E168d2, that produces a truncated protein (at 171 of 299 amino acids),⁹ a mouse with a L253X mutation,²¹ and the Rdy (rod-cone dysplasia) cat, which has a spontaneous frameshift mutation (c.546del, p.Ala185LeuTer2) resulting in a truncated protein of 185 amino acids.²² The L253X mouse produces a protein with less truncation of the transactivation domain and a milder phenotype than the other class III models. The E168d2/+ mice and CRX^Rdy/+ cats have a similar and more severe phenotype associated with overexpression of the mutant allele and higher levels of the mutant than the wild-type protein.^9,23 Development of photoreceptors in CRX^Rdy/+ cats is halted resulting in only stunted photoreceptor outer segments. They lack cone electroretinograms (ERG) but initially have delayed and markedly reduced rod-mediated ERGs that are extinguished by 20 weeks of age as photoreceptor degeneration progresses. Degeneration is most rapid in the cone-rich area centralis, which also has a higher packing of photoreceptors than the peripheral retina and is the equivalent of the human macula. In this region the outer nuclear layer is lost by 25 weeks of age.²³ The presence of a macula-equivalent retinal region gives the cat model a distinct advantage over the mouse models that do not have similar regional differences in photoreceptor number and distribution. In both humans²⁴ and cats²⁵ the central region (center of area centralis and center of macula) is cone-rich, having peak cone density, and is surrounded by a region of peak rod density (the perimacular rod-rich region in the human). Despite the similarities in photoreceptor distribution, the cat does not have a cone-only fovea, which is a feature of the human retina. 

The purpose of the current study was to report the phenotype of the homozygous (CRX^Rdy/Rdy) cat. As anticipated, the homozygous cat had a more severe decrease in mRNA levels of photoreceptor expressed CRX target genes than the heterozygous (CRX^Rdy/+) cat. In contrast to the phenotype of the heterozygote, the homozygotes failed to develop photoreceptor outer segments and had no detectable photoreceptor function. They initially developed relatively normal retinal lamination, although the outer nuclear layer (ONL) showed a bilaminar cell body arrangement. With progression they had a loss of photoreceptor nuclei and a marked loss of normal retinal lamination with cell migration and loss and extensive gliosis. The blind homozygous cats developed a marked globe enlargement because of expansion of the vitreal cavity also making them a potential model for abnormal globe growth resulting in myopia. 

Material and Methods

Ethics Statement

All procedures were performed in accordance with the ARVO statement for the Use of Animals in Ophthalmic and Vision Research and approved by the Michigan State University Institutional Animal Care and Use Committee. 

Animals

A colony of CRX^Rdy cats maintained at Michigan State University was used for this study and bred to obtain homozygote-affected (CRX^Rdy/Rdy) and heterozygote-affected (CRX^Rdy/+) kittens and wild-type (WT) control cats. Animals were housed under 12-hour dark/12-hour light cycles during breeding and 14-hour dark/10-hour light cycles the rest of the time (facility lighting level averages 8 × 10³ lux). They were fed a commercial feline dry diet (Purina One Smartblend and Purina Kitten Chow; Nestlé Purina, St. Louis, MO, USA). Animals studied ranged from two weeks to six years of age. For numbers used in the different experiments, please refer to Supplementary Tables S1, S2, and S3. 

Ophthalmic Examination and Fundus Imaging

Full ophthalmic examinations were performed, including indirect ophthalmoscopy and wide-field color fundus imaging (Ret-Cam II, Clarity Medical Systems, Inc., Pleasanton, CA, USA). Confocal scanning laser ophthalmoscope (cSLO) fundus imaging (Spectralis OCT+HRA; Heidelberg Engineering Inc., Heidelberg, Germany) was performed with the animals under general anesthesia concurrently with spectral domain-optical coherence tomography (SD-OCT) examination (see below). 

In vivo fluorescein angiography imaged by cSLO was performed in a few animals. Anesthesia, pupil dilation, and globe positioning were performed as previously described.²³ A bolus of 20 mg/kg of 10% sodium fluorescein (Fluorescite 10%; Alcon Laboratories Inc, Fort Worth, TX, USA) was injected through a 20-gauge catheter in the left cephalic vein followed by a 2 mL bolus of Ringer Lactate. Images or video were recorded using a 55° wide-field lens. 

Measurement of Globe Length

Axial globe length was measured using a combined A- and B-mode ultrasound scanning (A/B Scan System 835; Humphrey, Dublin, CA, USA) with the animals under anesthesia. Initially, only the axial globe length was measured, but in later studies the cornea-anterior segment anterior to posterior width, lens width, and posterior segment depth were also measured. Measurements in millimeters were taken from the best A scan and B scan combined on the same images (Fig. 9A). 

Intraocular Pressure

Intraocular pressure (IOP) was measured using a TonoVet (Icare Finland Oy, Helsinki, Finland). Measurements were performed at the same time in the morning and three recordings averaged for each eye. 

Refraction

Refractive error was assessed using a retinoscope and standard refractive bars. 

ERG

Electroretinography was performed on CRX^Rdy/Rdy kittens under general anesthesia as previously described.²³ 

Assessment of Retinal Morphology and Vasculature

In Vivo SD-OCT

In vivo retinal morphology was assessed by SD-OCT (Spectralis OCT + HRA, Heidelberg Engineering Inc., Heidelberg, Germany) as previously described.²³ Total retinal thickness, Receptor+ (REC+; including layers between retinal pigmentary epithelium and outer plexiform layer included)²⁶, inner nuclear layer (INL), ganglion cell complex (including inner plexiform layer, ganglion cell layer and internal limiting membrane) and inner retina (layers between inner nuclear layer and internal limiting membrane) thickness were measured in the center of the area centralis and in the four retinal quadrants (at four optic nerve head diameter distances from the edge of the optic nerve head superiorly, inferiorly, nasally, and temporally) using the Heidelberg Eye Explorer software. 

Immunohistochemistry (IHC)

After humane euthanasia, eyes from CRX^Rdy/Rdy and wild-type kittens were collected (at two, six, 12, and 20 weeks, and two and 3.5 years of age) and processed as previously described.²³ The antibodies used are listed in Supplementary Table S4. 

Plastic Embedded Sections

Eyes were processed for plastic histologic sections and imaged as previously described.²³ Samples from the dorsal, ventral and area centralis regions of the glutaraldehyde fixed eye cup were processed and imaged. 

Quantitative Reverse Transcriptase–Polymerase Chain Reaction

Retinal samples were collected immediately following euthanasia and globe removal from 2-week-old CRX^Rdy/Rdy, CRX^Rdy/+ and wild-type kittens. Two areas (central and peripheral areas) were dissected as previously described.²³ Retinal samples were flash frozen and stored at −80°C until RNA extraction, and the remaining retina was stored for protein assay. RNA extraction, cDNA synthesis, and quantitative reverse transcriptase–polymerase chain reaction were performed as previously described.⁹ RNA quality was assessed, and only samples with an RNA integrity number RIN > 7.0 were used to evaluate gene expression changes. The target genes and primer sequences are shown in Supplementary Table S5. 

Western Blot Assay

Processing of retinal samples for Western blot was as previously described.^9,23 Monoclonal mouse anti-β-actin antibody (Sigma-Aldrich, St. Louis, MO, USA) and polyclonal rabbit anti-CRX 119b1 at 1:1000 dilution were used to probe the membranes. Secondary antibodies were goat anti-mouse IRDye 680LT and goat anti-rabbit IRDye 800CW (LI-COR Biosciences, Lincoln, NE, USA), respectively. Fluorescence was detected using the Odyssey Infrared Imager (LI-COR Biosciences). Quantification was performed using Image J (http://imagej.nih.gov/ij/; provided in the public domain by the National Institutes of Health, Bethesda, MD, USA).²⁷ 

Statistical Analysis

Statistical analysis of IOP, refraction, cDNA level, and Western blot fluorescence level data differences were tested for normality (Shapiro-Wilk test for normality). Normally distributed data was analyzed by unpaired two-tailed Student's t-testing (significance level set at P < 0.05) and nonparametric data by a Mann-Whitney rank sum test (SigmaPlot 12.0; Systat Software, Inc., San Jose, CA, USA). Student's t-testing was performed when comparing only two groups. A mixed-effect model using R studio was used to analyze the data for globe length and SD-OCT measurements; SD-OCT measurements as data were evaluated over time. This was also used to analyze the effect of other factors on IOP and refraction (age) using the equation below.²⁸ 

\begin{eqnarray*}{Y_i} = \mathop \sum \limits_{i = 0}^n \beta {\rm{X\;}} + {\rm{\;}}{\alpha _i} + {\varepsilon _i},\end{eqnarray*}

where β is the parameter vector, X is the independent variable matrix, α_i is the cat level residual, and Ɛ_i is the individual observation level residual. 

Results

CRX^Rdy/Rdy Retinas Overexpressed Mutant CRX and Had a Marked Reduction in Expression of Photoreceptor Specific Genes and Variable Changes of Other Retinal Transcription Factors

To investigate molecular changes in the maturing CRX^Rdy/Rdy retina, we measured mRNA levels of CRX and other transcription factors involved in photoreceptor development, as well as rod and cone–specific genes in two-week-old CRX^Rdy/Rdy, CRX^Rdy/+, and WT kittens (an age at which photoreceptors are developing inner and outer segments in WT cats) (Fig. 1. Supplementary Table S6). 

Figure 1.

 

Mutant CRX is overexpressed and alters expression levels of target genes. (A) QRT-PCR for total CRX. Overexpression of CRX is apparent in both CRX^Rdy/+ and CRX^Rdy/Rdy kittens during retinal maturation (two weeks of age). Shown relative to WT cat CRX mRNA levels. (B1) Western blot for nuclear CRX protein (immunolabeled with antibody 119b1). The truncated mutant CRX is present in the CRX^Rdy/Rdy kitten retina (two weeks of age). (B2) Quantification of total CRX levels. Mutant CRX accumulates to high levels in the CRX^Rdy/Rdy samples compared to the level of WT CRX in the WT control kittens (two weeks of age). **P < 0.01. (B3) IHC for CRX. Labeling for total CRX shows accumulation in CRX^Rdy/Rdy samples that becomes more pronounced with age. (C) Changes in mRNA expression (qRT-PCR) in CRX^Rdy/Rdy retinas compared to CRX^Rdy/+ retinas (two weeks of age). Expression of photoreceptor genes is further decreased in CRX^Rdy/Rdy retinas compared to CRX^Rdy/+ retinas (Rho, rhodopsin; Arr3, cone arrestin; MLO, medium-long wavelength opsin; SO, short wavelength opsin). Expression of transcription factors is also altered: OTX2 expression is significantly increased compared to WT retinas. NRL and NR2E3 are expressed at lower levels in the CRX^Rdy/Rdy compared to the CRX^Rdy/+ and WT kitten, whereas TRβ2 and RORβ were expressed at higher levels compared to the WT kitten's retinas. P values comparing the means CRX^Rdy/Rdy, CRX^Rdy/+ and WT expression levels are *P < 0.05, **P < 0.01, and ***P < 0.001. (D) IHC for rhodopsin (RetP1 antibody). In the WT animal this labels the outer segments. In CRX^Rdy/Rdy there is some labeling of the very small inner segments (IS) at the two- to three-week timepoint. There is also some mislocalization with some expression in the ONL cell bodies that was more apparent at 12 weeks. In the adult a few RetP1-positive cells were still apparent.

Levels of total retinal CRX mRNA were about 1.8 times higher than WT in both the CRX^Rdy/Rdy and CRX^Rdy/+ animals (Fig. 1A). The increased mRNA levels in CRX^Rdy/Rdy animals led to mutant CRX protein levels about three times higher than that of the normal CRX in the retinas of WT controls (P = 0.005) (Figs. 1B1, 1B2). Immunolabeling with an antibody that specifically detects both wild-type and mutant CRX showed that the mutant CRX protein was present across inner and outer nuclear layers in the CRX^Rdy/Rdy cat (shown in Fig. 1B3 and Supplementary Fig. S1 at 12 weeks of age, which is shortly after the age at which the retina is functionally mature in the WT cat). The increased expression of the mutant CRX was maintained with age, with strong immunolabelling of retinal cell nuclei across the inner and outer nuclear layers in adult CRX^Rdy/Rdy cats (see representative section at 3.5 years of age in Fig. 1B3 and Supplementary Fig. S1). In contrast, the CRX immunolabeling in the WT adult cat was not so strong and was predominantly detected in the outer nuclear layer, although a subset of inner nuclear layer cells was labeled for CRX (Supplementary Fig. S1, upper panel). Many but not all of the CRX-positive nuclei in the INL appeared to have cytoplasm labeled by PKCalpha. 

The CRX^Rdy/Rdy and CRX^Rdy/+ mutant cats had marked reduction in mRNA levels of the rod (rhodopsin) and cone (cone arrestin and cone opsins) markers assessed (Fig. 1C). The levels in the CRX^Rdy/Rdy kitten samples were significantly lower than in the CRX^Rdy/+ kitten samples. In young animals, despite the marked reduction in rhodopsin mRNA, rhodopsin protein was detected by immunostaining in small protuberances from the ONL into the subretinal space (Fig. 1D). By 12 weeks, rhodopsin signal expanded and mislocalized to remaining cell bodies in the outer retina (ONL and INL). In adult animals, only a small number of rhodopsin-positive cells remained (Fig. 1D). IHC for cone markers (arrestin, and cone opsins) did not label any cells at any ages (Fig. 2). 

Figure 2.

 

Immunohistochemistry for cone markers (human cone arrestin, ML-opsin, S-opsin, and the lectin PNA). There was no labeling in the CRX^Rdy/Rdy cat retina. WT control retina shown for comparison. The hCAR labels the entire cone photoreceptor cell body in WT cats. PNA labels the cone matrix in WT cats. ML-Opsin and S-Opsin are visible within the outer segment of the cone photoreceptor cells in WT cats.

The mRNA levels of CRX-interacting retinal transcription factors involved in photoreceptor differentiation were altered in different ways in the CRX^Rdy/Rdy kitten retina compared to WT controls (Fig. 1C): OTX2, TRβ2 and RORβ, which were normally expressed in immature rods/cones were increased, whereas rod-specific transcription factors NRL and NR2E3 were decreased. These results were distinct from those of the CRX^Rdy/+ retinas, in which OTX2 mRNA levels were increased whereas those for NRL, NR2E3, TRβ2, and RORβ were not significantly different from WT levels. The differences between CRX^Rdy/Rdy and CRX^Rdy/+ animals in transcription level changes of all genes examined are shown in Figure 1C. All P values are shown in Supplementary Table S6. 

CRX^Rdy/Rdy Retinas Developed Immature Photoreceptors Lacking Outer Segments and Underwent Extensive Retinal Remodeling

SD-OCT imaging of young CRX^Rdy/Rdy animals showed the development of major retinal layers, but the zones such as the ellipsoid zone and interdigitation zone, representing the region of the photoreceptor inner/outer segments and interaction with the retinal pigment epithelium,²⁹ could not be detected at any age (Figs. 3A, 4A). Plastic-embedded semi-thin sections and transmission electron microscopy (TEM) sections in young kittens (Figs. 3B, 3C) showed the presence of small rudimentary inner segments surrounded and in contact with RPE villosities, but no outer segment material was detected. The rudimentary inner segments corresponded to the rhodopsin labelled protuberances from the ONL seen on IHC (Fig. 1D). On SD-OCT, as well as histology and immunohistochemistry sections, the ONL had a bilayered appearance at a young age (four to six weeks). This corresponded to two ONL cell nuclei populations: the outer portion of the ONL had cell bodies with elongated nuclei whereas the nuclei of cell bodies in the inner portion of the ONL were more circular (Fig. 3B). 

Figure 3.

 

Photoreceptor development is halted in CRX^Rdy/Rdy cats. (A) SD-OCT is of a six-week-old CRX^Rdy/Rdy cat; the infrared scanning laser ophthalmoscope fundus image shows the region of the SD-OCT image (1) with the green line. The white boxes on the SD-OCT image show the regions magnified in A2 and A3. The overlaid IHC with DAPI fluorescent staining are from a two-week-old in A2 and 20-week-old in A3. The white box on the IHC image in A2 represents the region imaged in B1. Relatively normal retinal lamination from the vitreal face of the retina to ONL is apparent in the younger animals. Although the ONL has uneven reflectance. By 20 weeks, remodeling of retinal layers is apparent on SD-OCT and in the DAPI image (A3). (B) Plastic sections at two weeks of age. (1) Low-power view showing relatively normal retinal lamination. (2) Magnified view of the outer retina. The ONL nuclei show a bilaminar arrangement outlined by the black ellipses. The inner half of the ONL has nuclei that are round, similar to mature photoreceptors, whereas the outer half has more elongated nuclei similar to immature photoreceptors. (3). Higher magnification of the region indicated by the black box on B2 shows the presence of short vestigial inner segments (white arrowheads) but no obvious outer segment development. The retinal pigment epithelium in indicated by a black arrow (C). TEM showing the presence of inner segments (white arrowhead) and RPE villosities (white asterisk) but no outer segments. GCL/NFL, ganglion cell layer/nerve fiber layer; IPL, inner plexiform layer; OPL, outer plexiform layer; IS, inner segments; RPE, retinal pigment epithelium.

Figure 4.

 

Extensive remodeling of CRX^Rdy/Rdy retina occurs with age. (A). SD-OCT comparing dorsal region imaging of retina of CRX^Rdy/Rdy cats with WT controls. At six weeks the SD-OCT of the CRX^Rdy/Rdy cat and WT are similar except for the absence of zones representing the outer segments. With age the definition of layers is lost in the CRX^Rdy/Rdy cat, although overall retinal thickness is maintained and a demarcation between inner and outer retina can still be discerned. (B) Plastic sections showing the loss of cell bodies and loss of retinal lamination with age. Overall retinal thickness is maintained most likely due to glial activation (black arrowheads indicate glial extensions between remaining cell bodies; also see Figure 5A). (C) Thicknesses of retinal layers in the dorsal region from SD-OCT imaging from CRX^Rdy/Rdy, CRX^Rdy/+ and WT cats (mean ± SD). Note the maintenance of the total retinal thickness but thinning of REC+ and thickening of the IR in the CRX^Rdy/Rdy cat. In the CRX^Rdy/+ cat progressive outer retinal thinning occurs as shown on REC+ graph. TR, total retina; REC+; receptor+; IR, inner retina; GCC, ganglion cell complex; GCL/NFL, ganglion cell layer/nerve fiber layer; IPL, inner plexiform layer; OPL, outer plexiform layer; OS, outer segment; IS, inner segment; RPE, retinal pigment epithelium.

The normal lamination of the retina became less apparent as early as 12 weeks of age and further deteriorated with age, although the overall retinal thickness was maintained. Figure 4 illustrates the changes in SD-OCT appearance and retinal morphology of the dorsal central retinal region with age. Changes in thicknesses in the other 3 quadrants examined showed a similar pattern (data not shown). The area centralis showed a different pattern of retinal layer thickness changes and is considered separately below. The maintenance of overall retinal thickness with age was different from the findings in the CRX^Rdy/+ cats which had a progressive retinal thinning predominantly driven by more severe outer retinal thinning (Fig. 4C). Both CRX^Rdy/+ and CRX^Rdy/Rdy cats showed thickening of the inner retina compared to wildtype controls. This became more pronounced after one year of age in the CRX^Rdy/Rdy cats whereas in CRX^Rdy/+ cats it tended to decrease after one year of age. In the CRX^Rdy/Rdy cats this compensated for the outer retinal thinning meaning that the overall retinal thickness remained close to that of WT cats. 

The plastic sections in Figure 4B show the loss of retinal cells in the ONL and INL and the apparent appearance of glial cell processes extending to within the ONL (Fig. 4B, black arrowheads). IHC (Fig. 5A) showed marked GFAP immunolabelling by 12 weeks of age. In the more mature animals GFAP signal was even more extensive throughout the retina showing extensive Müller cell activation and retinal gliosis.^30–32 

Figure 5.

 

IHC showing inner retinal cell changes with time in CRX^Rdy/Rdy cat. (A). GFAP is a glial cell marker, and increased labeling can be an indication of Müller glia activation. The normal cat has labeling in the region of the ganglion cell layer. With age there is a progression of GFAP-positive processes throughout the retina (arrowheads show Müller cell processes which with age tend to replace other cell types.) (B). PKCalpha which labels bipolar cells shows that with age there is disorganization of the labeled cells with dendrites invading the ONL and forming a matrix within it (white star). (C) Immunolabeling with NeuN antibody showed apparently normal labeling of ganglion cells and some INL cells as in the WT retina, but there was also some abnormal labeling through the ONL nuclei. OS, Photoreceptor outer segment; IS, Photoreceptor inner segment; OPL, Outer plexiform layer; IPL, Inner plexiform layer; GCL, ganglion cell layer.

Immunolabeling of bipolar cells showed extensive migration and branching of cells with age. By 3.5 years of age PKC alpha–positive bipolar cells extended throughout the remaining ONL (Fig. 5B, Supplementary Fig. S1). Some PKC alpha–labeled cells had CRX-positive nuclei. Immunolabeling for the synaptic marker CTBP2 (bassoon) showed reduce labeling of outer plexiform layer in both the heterozygous and homozygous cat compared to wildtype (Supplementary Fig. S2). PKC alpha labeling bipolar cell extension reached beyond the synaptic terminals of photoreceptors into the outer nuclear layer. NeuN, a neuronal marker, labeled ganglion cells as in normal retina, but it also abnormally labeled inner retinal cells and ONL nuclei at all age tested in the CRX^Rdy/Rdy retina (Fig. 5C). 

CRX^Rdy/Rdy Retinas Showed Early Degeneration of Photoreceptors in the Area Centralis

SD-OCT across the area centralis showed an early thinning of the ONL (Figs. 6A and B). This developed at an earlier age than seen in the CRX^Rdy/+ kittens (Fig. 6A). The SD-OCT finding was confirmed by histology (Fig 6B lower histology images). Figure 7 shows heat maps for retinal layer thickness of the area centralis from SD-OCT measure showing the progressive thinning of the outer retina while inner retina thickness increases. 

Figure 6.

 

Early degeneration in the center of the area centralis. (A) SD-OCT layer measurement in the center of the area centralis from CRX^Rdy/Rdy, CRX^Rdy/+ and WT cats (mean ± SD). Note the early and marked loss of REC+ in the CRX^Rdy/Rdy cat whereas the inner retina thickens and total retinal thickness is maintained. (B) SD-OCT scan vertically through the area centralis of a 20-week-old CRX^Rdy/Rdy cat. The insert shows a magnified image of the area centralis with outer retinal thinning. Below a DAPI labeled frozen section and plastic section are shown. These show the small region of marked ONL thinning present at the center of the area centralis.

Figure 7.

 

Total retina, Receptor+ and inner retina thicknesses color (heat) maps in the area centralis. (A) Shows color (heat) maps for a representative CRX^Rdy/Rdy cat from six weeks to four years of age. (B) The retinal location of the color map. (C) The color map for representative CRX^Rdy/+and WT cats at six weeks and four years of age. Note the thinner TR and REC+ in the CRX^Rdy/Rdy cat from an early age compared to WT cats. The IR of the CRX^Rdy/Rdy cat is thicker than that of the CRX^Rdy/+ and WT cats. With disease progression, thickening surrounding the center of the area centralis can be seen in the CRX^Rdy/Rdy cat, leading to slight TR and REC+ thickening, which is in contrast with the severe thinning of retinal layers in the CRX^Rdy/+ cat at four years of age.

CRX^Rdy/Rdy Kittens Lack Retinal Function

The CRX^Rdy/Rdy cats were blind. They had no dazzle reflex at any age, lacked visual tracking responses and did not develop a menace response. Dark- and light-adapted ERGs assessed at multiple time points from four to 20 weeks of age failed to elicit any measurable response (data not shown). A slow and reduced pupillary light reflex was initially present. Nystagmus was not noted at any age. 

CRX^Rdy/Rdy Cats Developed Tapetal Hyperreflectivity and Local Choroidal Atrophy but Maintained Superficial Retinal Vasculature

Color fundus images are shown in Figure 8. The important features are that hyperreflectivity of the tapetal fundus became apparent with age (Fig. 8A). This could be discerned as early as 12 weeks of age (Fig. 8A) and in fact the appearance of the tapetum was never normal having a generalized abnormal “sheen” when compared to WT cats. Tapetal hyperreflectivity indicates that there is less attenuation of light as it passes through the retinal layers and is reflected back from the tapetum. With progression some loss of tapetal tissue became apparent in a pattern radiating out from the optic nerve head; compare the fundus images in Figure 8A at 26 weeks and 3 years of age which are from the same animal and show the progression of these lesions. With the loss of tapetum, choroidal vessels could be visualized. SD-OCT imaging showed tapetal and choroidal thinning in those areas (data not shown). 

Figure 8.

 

CRX^{Rdy/ Rdy} cats have good preservation of superficial retinal vasculature compared to the CRX^Rdy/+ cats. (A) Fundus images of CRX^Rdy/Rdy show development of tapetal hyper-reflectivity with age, which usually indicates retinal thinning. There is also tapetal thinning close to the optic nerve head allowing tapetal vasculature to be visualized. (B) comparison of WT cat with CRX^Rdy/Rdy and CRX^Rdy/+ four-year-old cats. The color fundus images are followed by cSLO infrared, autofluorescence, and fluorescein angiography images. The CRX^Rdy/Rdy cat shows a lack of tapetum as dark streaks radiating from the optic nerve head. The white arrow indicated a choroidal vessel that is exposed. Compare the vasculature to that of the CRX^Rdy/+ cat, which has only the major vessels still detectable. The white arrowhead indicates the same vessel for the three types of imaging.

The superficial retinal vasculature was well maintained in the CRX^Rdy/Rdy cats. Figure 8B compares color fundus images and cSLO and fluorescein angiography images between a representative adult WT cat and age-matched CRX^Rdy/Rdy and CRX^Rdy/+ cats. The CRX^Rdy/Rdy cat has developed less hyperreflectivity of the tapetum (color images) than the CRX^Rdy/+ cat, and shows better preservation of superficial retinal vasculature. One region of tapetal loss in the CRX^Rdy/Rdy cat is indicated with a white arrow and fluorescein angiography highlights the exposed underlying choroidal vessel. The relative preservation of superficial retinal vessels in the CRX^Rdy/Rdy cats compared to CRX^Rdy/+ cats is clearly illustrated in the fluorescein angiography panel. The CRX^Rdy/+ cat had markedly attenuated retinal vasculature with mild leakage of fluorescein from remaining superficial retinal vessels. 

CRX^Rdy/Rdy Cats Develop an Increased Globe Length and a Myopic Refractive Error

CRX^Rdy/Rdy cats had a significantly increased axial globe length compared to CRX^Rdy/+ and control WT cats (Fig. 9) (P < 0.002 and = 0.003, respectively). For example, at one year of year the CRX^Rdy/Rdy, CRX^Rdy/+ and WT cats had an axial globe length of respectively 23.1 ± 0.4, 19.5 ± 0.3, and 20.1 ± 0.4 mm. This difference in globe size was obvious on ocular ultrasonography (Fig. 9A, bottom panel), and in the enucleated eye (Figs. 9B). Scatter plots in Figures 9C and 9D show the changes with age. The increase in axial globe length was due to an increase in the posterior segment length (P < 0.0001) (Fig. 9D). The anterior-posterior depth of the anterior chamber and the width of the lens did not differ between genotypes (data not shown). A difference in axial globe length was also present between CRX^Rdy/+ and WT cats (P = 0.02), with WT having a slightly greater axial length. It should be noted that the CRX^Rdy/Rdy group included 55 males and 13 females (each time point considered independently), the WT group included 94 M and 54 F and the CRX^Rdy/+ group in 44 males and 106 females. It is possible that the difference between the CRX^Rdy/+ and the WT group was due to the high number of females that were of smaller size compared to the males. For example, at three years of age female and male CRX^Rdy/+ cats had a mean axial globe length of 20.4 ± 0.13 and 21.50 ± 0.34 mm, respectively; a difference that was statistically significant (P ≤ 0.001). 

Figure 9.

 

CRX^{Rdy/ Rdy} cats develop globe enlargement. (A) Ultrasound images showing the measurements performed and below representative images from the different genotypes (each 12 weeks of age). (B) Gross pictures of enucleated globes from age-matched (12 weeks of age) animals show the much larger globe in the CRX^Rdy/Rdy compared to CRX^Rdy/+ cat. (C) Scatter plots of the axial length of the three genotypes with age. (D) Scatter plots of the posterior segment length of the three genotypes with age.

There was no difference in IOPs between CRX^Rdy/Rdy and WT cats (16.0 ± 1.8 and 15.4 ± 1.6 mm Hg, respectively; P = 0.577). The change in axial globe length did alter the refractive state of the eye. The CRX^Rdy/Rdy cats had a refractive error of a mean of −14 ± 1.1 D compared to −0.9 ± 0.4 D in the WT cats (P = 0.004). A refractive error was also present in the CRX^Rdy/+ cats with an average of −2.7 ± 0.8 D (P < 0.001 compared to CRX^Rdy/Rdy cats and P = 0.002 compared to WT cats). When analyzed with a mixed-effect model, age had a significant influence in the difference between the CRX^Rdy/+ cats and both the CRX^Rdy/Rdy and WT cats (P = 0.003). The eye (left or right) did not have an effect on the difference between groups, nor did it have an effect on refraction or IOP. 

Discussion

This study adds to previous studies characterizing the Rdy cat. Although phenotypic descriptions of the heterozygous CRX^Rdy/+ cat have been previously reported, there had been no studies reporting on the homozygous CRX^Rdy/Rdy cat.^23,33–36 In keeping with previous studies of Class III Crx mouse mutant models, the CRX^Rdy/Rdy cat showed an overexpression of the mutant CRX transcript and protein that we predict to exert an antimorphic effect on developing photoreceptor cells. The Rdy feline model offers important advantages because of the presence of a macula-like retinal region allowing for further characterization of the phenotype of Class III CRX mutations. 

Molecular Mechanism Underlying CRX^Rdy/Rdy Phenotype

The Rdy mutation introduces a premature stop codon ²² in CRX with the mutant transcript escaping nonsense-mediated decay and resulting in overexpression of the mutant allele (Figs. 1A, 1B).^9,37,38 Lack of normal CRX activity resulted in a marked decrease in cone and rod transcripts in the CRX^Rdy/Rdy cat retinas (Figs. 1C, 1D). It also had effects on the levels of other transcription factors involved in photoreceptor development. CRX, OTX2 and RORβ showed an increase in expression. The increase in OTX2 expression may be due to a retroactive feedback mechanism compensating for the lack of other transcription factors. NRL and NR2E3 expression level was decreased, perhaps contributing to the partial failure in maturation of the photoreceptor nuclei. Although the cone maturation factor TRβ2 is still expressed, there was a lack of cone opsin expression and progressive cone degeneration, suggesting that CRX is required for cone maturation and maintenance. Although Rdy clearly inhibit the expression of many rod and cone–specific genes necessary for vision, photoreceptor transcription factors showed both positive and negative regulation; because these transcriptional network interactions are complex, this regulatory process is still not completely understood.^2,39 

The CRX^Rdy/Rdy Kitten Is a Model for Severe CRX-LCA Retinopathies

The CRX^Rdy/Rdy cat phenotype is characterized by blindness from birth, with a sluggish PLR, and no dazzle reflex or menace response. We suspected that the slow residual PLR was driven by melanopsin-containing ganglion cells.⁴⁰ Ophthalmoscopic features of the model included the development of tapetal hyper-reflectivity, which is considered an indicator of generalized retinal thinning. However, there was not a significant decrease in total retinal thickness so the tapetal hyper-reflectivity likely either resulted from the almost total lack of cone and rod opsins that in a normal eye would absorb photons passing through the retina or the extensive cellular degeneration and remodeling that occurred. The photoreceptors of the CRX^Rdy/Rdy cat only developed very short rhodopsin-positive inner segments (Fig. 1D) before photoreceptor maturation became halted. The photoreceptor nuclei consisted of two populations with different nuclear morphology: one which population had oval-shaped nuclei comparable to the appearance in an immature retina whereas the other showed a more mature morphology (round nuclei). It appeared that the normal migration of cell bodies of photoreceptors that occurs during maturation of the retina did not occur. The lack of photoreceptor maturation was reflected in the complete lack of detectable ERG responses. Interestingly, photoreceptor degeneration was not as rapid as in the CRX^Rdy/+ cat, although degeneration and marked cell migration occurred. There was initially relatively normal stratification of the retinal layers, but over time there was extensive activation of Müller cells and sprouting and migration of bipolar cells (Figs. 4, 5, Supplementary Figs. S1, S2). The normal interaction between photoreceptor synapses labeled with CTBP2 and PKC alpha–positive bipolar cells did not develop in the CRX^Rdy/Rdy cat, suggesting a lack of normal synapse formation. Retinal remodeling resulted in some thickening of the inner retinal layers, which counteracted the thinning of the outer retina and thus preserved the overall retinal thickness (Fig. 4).^30–32,41 This retinal remodeling was more severe than in the CRX^Rdy/+ feline phenotype ²³ but was similar to the E168d2/d2 homozygous mouse.¹⁷ The relative lack of overall retinal thinning may account for the striking persistence of retinal vasculature until advanced disease stages (Fig. 8). The area centralis had slightly different changes compared to the peripheral retina with thinning of the total retina and receptor+ layers (Figs. 6, 7) with early severe loss of photoreceptor nuclei in the center of the area centralis. 

The phenotype of the CRX^Rdy/Rdy cat is quite different from that of the heterozygote, not being simply a more-severe version. The absence of photoreceptor OS development might be responsible for the relatively slow photoreceptor nuclei loss compared to the CRX^Rdy/+ cat. In the CRX^Rdy/+ cats, the degeneration of photoreceptors expressing phototransduction proteins and retinal degeneration could result from altered phototransduction cascade dynamics and mislocalization of opsins.^42–47 

The CRX^Rdy/Rdy Kitten Develops Significant Globe Enlargement Without IOP Elevation

The CRX^Rdy/Rdy cat has a significant increase in axial globe length with resulting myopia caused by an increase in posterior segment length (Fig. 9).^48–51 Abnormal globe length is a well-recognized feature in animals with abnormal visual input.^52,53 The increase in globe size may also contribute to the choroidal thinning that was noted in the homozygous cats (Fig. 8).⁵⁴ The CRX^Rdy/Rdy cat provides a large animal model for investigating scleral growth factors implicated in myopia development and for the severe dominant CRX mutations associated with over expression of a mutant transcript with an antimorphic effect. 

Acknowledgments

The authors thank Cheryl Craft for the donation of the hCAR antibody, Alicia Withrow for her help with semi-thin sections, Wenjuan Ma for statistical advice, Paige Winkler-Smith and Kelian Sun for additional IHC, and the MSU CVM RATTS group for their help with animal handling. 

Supported by National Institutes of Health Grants EY012543 and EY025272-01A1 (SC), EY002687 (P30 Core Grant) (Washington University Department of Ophthalmology and Visual Sciences (WU-DOVS)), EY013360 (T32 Predoctoral Training Grant) (WU), unrestricted funds from Research to Prevent Blindness (WUDOVS), Foundation Fighting Blindness (SC), Hope for Vision (SC), George H. Bird and ‘‘Casper’’ Endowment for Feline Initiatives (LMO and SMPJ), Michigan State University Center for Feline Health and Well-Being (LMO and SMPJ), and Myers-Dunlap Endowment (SMPJ). 

Disclosure: L.M. Occelli, None; N.M. Tran, None; S. Chen, None; S.M. Petersen-Jones, None