Retinoschisin Gene Therapy and Natural History in the Rs1h-KO Mouse: Long-term Rescue from Retinal Degeneration

Abstract

purpose. The authors characterized the natural history of a retinoschisin gene knockout (Rs1h-KO) mouse model and evaluated the long-term effects of retinal rescue after AAV(2/2)-CMV-Rs1h gene delivery.

methods. Full-field scotopic electroretinograms (ERGs) were recorded from 44 male hemizygous Rs1h-KO and 44 male wild-type (WT) C57BL/6J mice at six ages between 1 and 16 months. Retinal morphometry included outer segment layer (OSL) width, photoreceptor cell count, and grading of schisis cavity severity. One eye each of seven Rs1h-KO mice at age 14 days was injected with AAV(2/2)-CMV-Rs1h, and retinal histology and ERG findings at 14 months were analyzed.

results. The outer nuclear layer (ONL) of 1-month-old Rs1h-KO mice was disorganized but had nearly normal cell counts. The OSL was thinned, rod outer segments were misaligned, and abundant schisis cavities spanned the inner nuclear and outer plexiform layers in all retinas. ERG a- and b-wave amplitudes at this age were reduced by 33% and 50%, respectively. ERG and ONL cell numbers decreased further between 1 and 16 months, with unequal changes in the a- and b-waves with age. The a-wave reduction correlated well with the steady decline in ONL cell number, whereas a rapid decline in the b-wave and a (b/a-wave) ratio less than in WT were associated with increasing severity of schisis cavities at young ages. At 4 months, the cavities were maximal, but they coalesced and disappeared at older ages. The (b/a-wave) ratio was inversely correlated with cavity severity across all ages (r = −0.74; P < 0.0001; n = 22). Considerable heterogeneity was observed at each age in the ERG amplitudes and retinal morphology. Mice injected with AAV-Rs1h at 14 days showed considerable structural and functional rescue at age 14 months, including improved rod outer and inner segment integrity, less photoreceptor cell loss, and larger ERG amplitudes compared with untreated fellow eyes.

conclusions. The ERG of the Rs1h-KO mouse at early ages reflects disruption of photoreceptor and second-order neuron function. In mid to late ages, the ERG decline reflects primarily photoreceptor degeneration. The Rs1h-KO mouse is consistent with human clinical X-linked juvenile retinoschisis (XLRS) in showing schisis cavities, which affect primarily the b-wave, the regression of schisis cavities at older ages, and a considerable range in phenotypic severity across individuals. This mouse model also indicates the critical roll of RS-protein in photoreceptor survival consistent with decreased a-waves in some patients with XLRS. Long-term rescue of retinal morphology and function by AAV-Rs1h gene transfer may provide a basis for considering intervention in the homologous human XLRS condition.

X-linked juvenile retinoschisis (XLRS) causes juvenile macular degeneration in males and is characterized by structural changes in the retina, including microcystic changes in the macula and schisis or splitting within the inner layers of the peripheral retina (for reviews, see Sikkink ¹ and Tantri ² ). Recent evidence using ocular coherence tomography (OCT) shows that the intraretinal splitting can involve all retinal layers. ^{3 4} One hallmark of XLRS is an electronegative electroretinogram (ERG) response in which the b-wave is reduced disproportionately to the a-wave. ⁵ Although the manifestation of the phenotype is consistent, the severity is highly variable, even within families; hence, severity is not determined solely by the specific mutation. ^{6 7 8} XLRS is diagnosed at about the time patients reach school age, but some severe cases with nystagmus and bilateral bullous schisis cavities are present at birth. 

The classical understanding of human retinoschisis posits a relatively stationary condition with minimally progressive clinical retinal changes with age: XLRS typically exhibits modest severity in patients at a young age, grows worse through the teenage years, and stabilizes in adulthood. ^{9 10} By the time the patient is 40 to 50 years old, the macular schisis becomes less obvious, but macular atrophy may occur in later age and cause additional visual failure. Complications of the disease can include vitreous hemorrhage and retinal detachment. Despite this, RS disease is not typically considered a progressive degenerative retinopathy, distinguishing it from retinitis pigmentosa. 

The cause of the disease has been linked to mutations in the gene encoding retinoschisin, a 24-kDa secreted protein found in the retina ¹¹ and pineal. ¹² The retinoschisin molecule contains a conserved discoidin domain (DD) ¹¹ and is a member of the DD family of proteins that are involved in cell adhesion and cell-cell interactions. ¹³ Retinoschisin is expressed in the mouse retina as early as postnatal day 1 (P1). During development, all retinal neurons express RS after differentiation, beginning with the ganglion cells, which are the first to mature, followed by neurons of each of the more distal layers. ¹⁴ From P14 onward, it is particularly strongly expressed in the outer half of the inner nuclear layer (INL) and by photoreceptor inner segment (RIS) but continues to be expressed in all classes of retinal neurons to a lesser degree even in adults. 

The retinoschisin-knockout (Rs1h-KO) mouse that we created, evaluated between 1 and 6 months of age, displays structural and functional features similar to those of human XLRS, ¹⁵ including the electronegative ERG waveform and splitting, or gaps, in the INL similar to so-called retinoschisis cavities. Similar retinal morphology was reported in an earlier study of a separate Rs1h-KO mouse model at 2 months of age in which the ERG b-wave was greatly reduced, with relative sparing of the a-wave. ¹⁶ The histologic observations and the reduction in the b-wave greater than in the a-wave presented in both studies suggested an effect on bipolar cell function, perhaps because of the disruption of synaptic transmission at the photoreceptor/bipolar cell synapse in the absence of retinoschisin protein. Our Rs1h-KO mouse also demonstrated displacement of cells from the photoreceptor nuclear layer (ONL) and reduced thickness of the outer segment layer by 6 weeks, but with a near normal ONL thickness out to 6 months. ¹⁵ The earlier study, in addition to reduced outer segment layer thickness, also showed some photoreceptor loss by 2 months of age that affected cones more than rods. ¹⁶ Neither study systematically looked at the progression of ERG or morphologic changes with age. Our previous study also showed that intraocular gene delivery using AAV(2,2)-CMV-Rs1h injections into 3-month-old mice resulted in a return of the normal ERG waveform configuration by 6 months of age. ¹⁵ A somewhat longer term of treatment showed modest recovery of scotopic and more pronounced recovery of photopic ERG function to 5 months in mice injected with AAV5-mOP-Rs1 at 15 days of age. ¹⁷  

In the present study, we systematically evaluated the natural history of retinal structural and functional abnormalities in Rs1h-KO mice to 16 months of age and investigated the long-term effects of treatment using AAV(2,2)-Rs1h in Rs1h-KO mice. The results showed clear evidence of progressive changes in both the inner and the outer retina and a rough correlation between structural abnormality and functional impairment. Treatment with a single intravitreal injection of AAV(2,2)-RS at P14 (at 2 weeks age) greatly reduced the structural and functional loss of the retina in mice when they were evaluated at 14 months age. 

Methods

Mice

Animal procedures were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and with the National Institutes of Health Animal Care and Use Committee Protocols 1126–03 and 1101–02. Approximately 50 Rs1h-KO mice, ¹⁵ ranging in age from 1 to 16 months, were studied. All were back-crossed approximately four generations onto C57BL/6 background and genotyped to verify that they carried the Rs1h-KO construct. Age-matched C57BL/6 wild-type (WT) mice from Jackson Laboratory (Bar Harbor, ME) were used as controls. 

Immunohistochemistry of Mouse Retina

A rabbit polyclonal RS antibody against the N terminus of retinoschisin (amino acid residues 24 to 37, translated from Rs1h exons 2 and 3) was used for immunofluorescence of mouse retinal specimens, as described. ¹⁴ Retinal sections cut at 8-μm thickness were blocked in 10% normal goat serum in PBS and then incubated with RS antibody at 4°C overnight. After washing in phosphate-buffered saline (PBS), sections were incubated with Alexa 568 goat anti-rabbit IgG/DAPI (Invitrogen, Carlsbad, CA). Images were collected with laser confocal microscope (SP2; Leica, Wetzlar, Germany). 

Histologic Evaluation of the Mouse Retina

Eyes from WT and Rs1h-KO male mice were enucleated from freshly euthanatized animals fixed for histology either by transcardial perfusion with 2.0% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M cacodylate buffer, followed by overnight immersion in the same fixative, or by immersion for 24 hours. Perfusion-fixed eye tissue was trimmed, postfixed in 1% osmium tetraoxide/dH₂O for 1 hour, and embedded in Araldite resin (Electron Microscopy Science, Hatfield, PA). Sections 0.5-μm thick were cut along the vertical meridian passing through the optic nerve and were stained with 0.1% toluidine blue for light microscopy. Photoreceptor cell loss with age was evaluated by counting ONL cells on photomicrographs of retinal sections taken with the 20× objective of a photomicroscope (E800; Nikon, Tokyo, Japan) and a digital camera (DXM1200; Nikon). Inferior and superior retina were counted separately between 200 and 1200 μm from the optic nerve (ON) using an automated method with ImageJ software (http://www.bioimage.ucsb.edu/software.html) with the ITNC nuclei detector plug-in (http://rsb.info.nih.gov/ij), as previously described. ¹⁸ The width of the photoreceptor outer segment layer (OSL) was measured perpendicularly between the retinal pigment epithelium (RPE) and the inner segment layer at 100-μm intervals in the same retinal regions. This measurement was used rather than the standard technique of measuring single rod outer segment (ROS) lengths because ROS in Rs1h-KO mice were often disrupted and misaligned. The severity of retinal schisis cavities was assessed by five examiners who individually scored the size and extent of cavities on a scale of 0 to 4, zero indicating no cavities, using digital photomicrographs of retinal sections taken at 2×. This was done for both the inferior and superior halves of the retinal sections. Values for the size and the extent of cavities were multiplied together and normalized to a 0% to 100% scale (0% meant not affected). A similar technique has been used to assess photoreceptor survival in histologic sections in light damage experiments. ¹⁹ Measurements were made on four to five animals at each age. 

Electroretinography

Full-field scotopic ERGs were recorded from 44 male Rs1h-KO and 44 male WT C57BL/6 mice at six ages between 1 and 16 months. Recordings were obtained only once from each mouse. Mice were dark adapted for 12 hours before anesthesia with intraperitoneal administration of ketamine (80 mg/kg) and xylazine (4 mg/kg). The pupils were dilated with topical 0.5% tropicamide and 0.5% phenylephrine HCl. Mice were placed on a heating pad to maintain body temperature near 38°C. ERGs were recorded with gold wire loops placed on the cornea with a drop of methylcellulose after application of 1% proparacaine topical anesthetic. Gold wires were placed on the sclera at the limbus as the differential electrodes, and a ground wire was attached to the left ear. Scotopic ERG responses were elicited using single flashes from a Xenon discharge source (Grass Photic Stimulator PS33; Astro-Med Inc., West Warwick, RI) from −6.9 to +0.6 log cd · s/m² in 0.5-log steps or with bright photostrobe flashes (model 283; Vivitar, Santa Monica, CA) from +1.4 to +2.4 log cd · s/m². Stimuli were delivered in a Ganzfeld (full-field) sphere. Interstimulus intervals lasted 3 to 180 seconds, depending on stimulus intensity. A stimulus intensity range of −6.9 to +2.4 log cd · s/m² was obtained using neutral density filters (Wratten; Eastman Kodak, Rochester, NY). Responses were amplified 5000 times and filtered using a 0.1-Hz to 1-kHz bandpass and a 60-Hz line-frequency filter (Grass CP511 AC amplifier; Astro-Med Inc.). A-waves were measured from the prestimulus baseline to the initial trough. B-waves were measured from the baseline or from the a-wave trough when present. Measurements were made on six to eight animals per age group. 

AAV-Rs1h Construct and Delivery

The Cis pAAV(2/2)-CMV-Rs1h vector, in which the Rs1h cDNA was driven by the CMV promoter, was made by inserting the 705-bp EcoRI fragment of the pCR-Rs1h plasmid into the EcoRI restriction sites of the pZac2.1 vector provided by the Vector Core, Medical Genetics Division, Department of Medicine, Medical School, University of Pennsylvania, using methods previously described. ^{20 21} Briefly, HEK-293 cells were triple transfected with three plasmids. The first plasmid encoded the Rs1h expression cassette packaged between the AAV2 internal terminal repeats; the second encoded the rep and AAV2 cap genes; and the third encoded the adenoviral helper function genes. The AAV(2/2)-Rs1h construct was purified by heparin column chromatography. The virus titer was assessed by real-time PCR, and the infectivity was assessed by an infectious center assay as described previously. ^{22 23} The ratio of the genomic copy number to the infectious center assay for this AAV(2/2)-Rs1h construct was 39. We previously reported use of this vector for Rs1h gene delivery. ¹⁵  

Animals were anesthetized, and intraocular injections were performed using a technique similar to that described earlier ²⁴ : a 33-gauge needle was inserted into the eye posterior to the limbus to deliver the vector-gene construct into the vitreous. Seven Rs1h-KO mice at P14 were given 1.5 μL AAV-Rs1h at a titer of 2.3 × 10¹⁰ GC/μL into the right eye. The contralateral left eye served as the control and remained untouched. 

Statistical Analysis

A quantitative assessment of retinal structure and ERG amplitudes is presented as the mean ± SE for each age group. Student’s t-test was used to calculate significant differences of Rs1h-KO ONL cell count at each age from the average WT cell count from 1 to 16 months and of the ERG amplitudes at 1 month. 

Results

Morphology

Mice were examined out to 16 months of age. Retinas of Rs1h-KO mice showed major changes in morphology as early as 1 month of age (Fig. 1) , with large and frequent cavities within the outer plexiform layer (OPL) and the inner nuclear layer (INL). The size and distribution of these cavities changed considerably with age. Cavities were most severe at 4 months of age, were present throughout the retinal sections, and frequently were wider than the normal width of the INL or the OPL. By 6 months of age, the number and size of the cavities were decreased, and many areas of the inner retina now appeared mostly intact. By 16 months of age, no cavities were observed anywhere in the retinal sections. Except for irregular boarders of the INL at the OPL junction, these layers looked relatively normal at older ages. There was no apparent loss of cells in the INL, even at later ages. An assessment of the severity of these schisis cavities in retinal sections (size x extent) by age (Fig. 2a)confirmed the impression from the photomicrographs that there was an increase in cavities between 1 and 4 months, after which they began to decline and were greatly reduced by 8 months. 

The photoreceptor layer of Rs1h-KO mice showed a slowly progressive loss of cells and consequent thinning to approximately one quarter of normal thickness by 16 months of age (Figs. 1h 1l) . At 1 month (Figs. 1c 1i 2b) , the photoreceptor cell number was not significantly different from WT. However, the structure of the ONL was disrupted, and some nuclei were displaced into the OPL and the photoreceptor inner segment (RIS)/ROS layer (Figs. 1c 1i) . At older ages, the cell number and ONL width declined steadily, but beyond 4 months the overall structural integrity of the layer improved: the borders were more distinct, and there were fewer displaced cells. The width of the OSL at 1 month was approximately half that of normal but was normal by 4 months of age, suggesting a developmental delay in normal ROS elongation in Rs1h-KO mice. This layer became progressively thinner thereafter (Fig. 2c) . At younger ages (1 and 4 months), the regular alignment of the ROS was disrupted, and it was difficult to find any well-aligned inner or outer segments. Like the ONL, however, they were more orderly at later ages. 

ERG

Figure 3shows Rs1h-KO a- and b-wave amplitude-intensity graphs over a 16-month age range. Amplitudes of the a-waves (Fig. 3a)and b-waves (Fig. 3b)throughout the stimulus intensity range decreased with age. As early as 1 month, the a-wave was reduced 33% and the b-wave was reduced 50% compared with those in WT mice at 0.6 log cd · s/m² stimulus intensity (a-wave mean ± SE: Rs1h-KO 418 ± 47, WT 627 ± 36, P < 0.005, n = 7; b-wave mean ± SE: Rs1h-KO 672 ± 74, WT 1357 ± 85, P < 0.0001, n = 7). The general reduction in amplitude technically elevated the threshold up to 1 log unit using a criterion amplitude determination. However, at all ages, responses could be tracked back to the same threshold intensity as for WT. The threshold reverted to normal if the maximum amplitudes were normalized. b-Wave intensity-response functions showed a different pattern of amplitude change with age from that of a-wave. The maximum b-wave response (Fig. 3c)declined considerably between 1 and 4 months and changed little between 4 and 8 months; after 8 months the amplitude continued to decline, but at a slower rate than earlier. The decline in a-wave saturated response was steadier with age. As a consequence, the (b/a-wave) ratio in Rs1h-KO mice varied with age (Fig. 3d) . It was smallest at 4 months of age after the rapid decline in the b-wave amplitude. The (b/a-wave) ratio increased rapidly between 4 and 8 months, when the b-wave was not changing and the a-wave continued to decline. At older ages, the (b/a-wave) ratio was often greater than in WT mice. 

The scotopic a-wave reflects primarily the rod photoreceptor light response, ^{25 26} and the b-wave reflects postsynaptic bipolar cell activity, albeit driven by photoreceptor input. ^{27 28 29 30} As shown in Figure 4a , the decline in a-wave amplitude with age coincided well with the loss of ONL cells, presumably because of its dependence on photoreceptor responses. In light of the progressively declining photoreceptor responses with age, the (b/a-wave) ratio might be a purer reflection of the response of the b-wave-generating mechanism because it normalizes for the photoreceptor response amplitude. The severity of schisis cavities in the inner retina was inversely correlated with the (b/a-wave) ratio (Fig. 4b ; r = −0.74), and the fit to the linear regression line appeared to be best (most points within the 95% confidence interval) at the highest severity, which occurred between 1 and 6 months. 

Considerable variability in the ERG waveform and in structural changes was observed at each of the ages of Rs1h-KO mice. For example, at 1 month of age, mouse 2709 had severe inner retinal cavities whereas mouse 2702 did not (Figure 5a) . The mouse with the most severe retinal cavities also had a more diminished b-wave relative to the a-wave. Similar spreads of cavity formation (not shown) and ERG response (Fig. 5b)were observed for each of the other age groups. At older ages the ERG shape was more normal, though the amplitudes were substantially reduced. By 12 and 16 months, only 4 of 7 in each age group had a recordable rod ERG, and some of these even had amplitudes comparable with 4-, 6-, and 8-month-old mice. However, in all mice with recordable ERGs, thresholds did not differ substantially. Thus, as previously shown for age-related effects, there was a considerable variability in ERG within an age group that seemed to be related to the severity of histologic alterations in the retina. ERG waveforms at ages up to 6 months were more similar to those typically observed in ERG waveforms of patients with retinoschisis in that b-wave reductions were larger than a-wave reductions. 

AAV-RS Treatment

We explored the long-term efficacy of Rs1h gene delivery therapy in ameliorating the extent of progressive retinal changes in the Rs1h-KO mouse retina. Seven animals were given intravitreal injections of AAV-Rs1h in the right eyes at 14 days of age and were evaluated at 14 months of age. The fellow eye was untouched. Four mice showed a substantially larger ERG in the treated eye than in the fellow control eye. Figure 6shows the retinal morphology and ERG from the animal with the greatest rescue. As described, Rs1h-KO mice at 14 months normally had severe reductions in the number of photoreceptor cells, and few had well-formed inner and outer segments. As seen in Figure 6 , the ONL cell layer in the untreated eye was reduced 90% to approximately one cell thickness. The treated eye, however, had four to five rows of photoreceptor nuclei, and the RIS and ROS looked essentially like those of WT but were one third to one half the length, which, because of their structural preservation, could now be easily measured. The full-field ERG of this treated Rs1h-KO retina had a waveform similar to that of WT, and the ERG b-wave intensity response relationship of the 14-month-old treated eye was nearly identical with that of an average 1-month-old Rs1h-KO mouse (Fig. 6d) . The a-wave amplitude, though lower than in 1-month-old animals, was also substantially larger than any we recorded from Rs1h-KO mice between 12 and 16 months of age (Fig. 6c) . 

Three additional mice of the seven treated also showed successful rescue based on substantially improved ERGs compared with the untreated eye (Fig. 7) . Though the absolute amplitudes of the treated eyes where not as great as in the mouse depicted in Figure 6 , improvement in the treated eye compared with the fellow eye in all three mice was greater than any difference observed between the two eyes in untreated animals at any age. In 24 Rs1h-KO mice for which neither eye was treated, the mean ratio of the b-wave amplitude in the best eye compared with the worst eye was 1.24 (SD ±0.20). By comparison, ratios for the four mice treated at P14 that showed rescue were 3.05 to 9.41, or 9 to 40 SD from the mean for the untreated mice, and thus were well outside the range of normal variability. 

Immunohistochemistry (IHC) for retinoschisin protein was performed on one injected eye that showed rescue, and this eye was found to have strong expression of RS in the inner segment layer (Fig. 7) . On single slides in a z-series of confocal images (not shown), weak expression in the inner nuclear layer and inner plexiform layer was also seen. We performed IHC on the treated eyes of the three mice that failed to show rescue after AAV-Rs1h. None showed expression of retinoschisin, leading us to conclude that gene delivery failed in these three attempts. 

Discussion

Retinal morphology in Rs1h-KO mice between 1 and 16 months of age showed two major pathologic changes. The most striking was the presence of many large cavities spanning the INL and OPL, reminiscent of the splitting in the retinal layers seen in OCT images of human XLRS retinas. ^{4 31 32} These cavities were maximal at 4 months but disappeared over approximately the next 8 months. In humans, macular schisis is often found during childhood and adolescence but can disappear in adulthood, leaving only a blunted macular reflex. ³³ In Rs1h-KO mice, no notable lasting effects of schisis cavity formation and collapse were observed at the light microscopy level except for the loss of a well-defined OPL and the presence of irregular INL boarders. No obvious cell loss occurred in the inner retina, though we did not conduct formal cell counts to confirm this. Others have also shown little or no cell death in the INL in another retinoschisis knockout mouse model up to 9 months of age. ³⁴ The mechanism of disappearance of the cavities is not clear, but one possibility is that Müller cells eventually retract and pull the margins of the cavities together in a mechanism similar to scar formation. 

A commonly used hallmark in the diagnosis of retinoschisis is a reduced (b/a-wave) ratio as a result of a larger reduction in the b-wave than the a-wave. This has been understood to result from schisis or splitting of the inner retinal layers preferentially affecting signals involved in b-wave generation. ^{5 35 36} The correlation between the severity of schisis cavities in the inner nuclear and outer plexiform layers and the decline in the b-wave in Rs1h-KO mice supports this theory: between ages 1 and 4 months, when cavities were increasing, the b-wave declined rapidly; between 4 and 8 months, when cavity severity was decreasing rapidly, the b-wave amplitude remained steady; after 8 months, when inner retinal structure remained stable, the rates of b-wave and a-wave decline were similar, suggesting they were determined by the same process at that stage (photoreceptor degeneration). Because of the different rates of decline in the a- and b-wave of the Rs1h-KO-mouse at different ages, the (b/a-wave) ratio changed considerably with age, as indicated in Figure 3d . No clinical reports have thus far correlated the ERG changes with the degree of schisis by fundus imaging. However, a recent study shows a high incidence of lamellar schisis observed by OCT in XLRS across a broad extent of the retina, even though this was otherwise not evident on clinical examination. This could explain the apparent lack of correlation between ERG and the funduscopic appearance of schisis cavities in patients with XLRS. ³  

A second prominent feature of the Rs1h-KO mouse retinal morphology was the disruption and cell death in the photoreceptor layer. ^{16 34} We observed a progressive decline in ONL cell number between 1 and 16 months and in the OSL width after 4 months of age. Retinoschisin is highly expressed in rod inner segments ^{14 16 37} and is associated primarily with the outer leaflet of the RIS plasma membrane. ³⁸ Furthermore, the photoreceptor RIS plasma membrane and the mitochondria of Rs1h-KO mice have abnormal morphology at the electron microscopy level. That these changes could affect photoreceptor survival is not surprising. The a-wave amplitude decline correlated well with the reduction in ONL cell number, as has been found for other animal models with primary photoreceptor degeneration. ³⁹ Some rodent models of retinitis pigmentosa also exhibit an increase in the (b/a-wave) ratio as rod degeneration progresses. ^{39 40 41} Consequently, one can infer that in the Rs1h-KO mice, higher (b/a-wave) ratios at late stages reflect a predominance of photoreceptor degeneration over inner retinal changes at older ages. In human XLRS, photoreceptor degeneration, especially outside the macula, has not generally been considered a major component of the disease, but more recent reports of reduced a-waves indicate that photoreceptor degeneration may be more prevalent than previously thought. ^{42 43 44 45 46}  

In Rs1h-KO mice, widespread photoreceptor loss and differences in the time course and distribution of schisis cavities may partially explain why their ERG waveforms differed from those of typical XLRS patients in lacking a truly “electronegative” ERG ([b/a-wave ratio] <1) and in showing a progressive reduction in overall ERG response and changes in the (b/a-wave) ratios with age. The lack of longitudinal studies of human XLRS with modern imagining techniques (OCT) makes it hard to determine whether schisis cavities are generally reversible in humans. However, clinical evidence indicates XLRS disease evolution and variability with age. Progression to macular atrophy increases with age and, conversely, causes apparent improved visual acuity in some individuals. ³¹ One of the authors (PAS) has observed variable courses of schisis cavity in patients with XLRS and has also observed regression or collapse of XLRS cavities in patients, which, as we have seen in this mouse model, may signal actual progression over time. Such cases of human XLRS cavity regression have also been noted by others (Rafael Caruso, personal communication, July 25, 2006). ³³  

Human XLRS can exhibit considerable variation within the same genotype and across ages. ^{6 42} We also observed large variations in the ERG and morphology in Rs1h-KO mice within the same age groups and even within litters, as illustrated in Figure 5 , suggesting this model replicates the inherent phenotypic variation seen in XLRS. Heterogeneity of the Rs1h-KO retinal phenotype was also reported in a separate knockout mouse model. ¹⁶ This heterogeneity could help explain why we earlier reported a lack of significant progression of morphologic abnormalities between 1 and 6 months of age. The average maximal difference between age groups during this time period in cavity severity, cell count, and OSL thickness was approximately 30%. This degree of difference is within the range of individual variability seen in Figure 5and by Weber et al. ¹⁶ In our previous qualitative sampling of randomly selected time points, these differences, especially if not linearly progressive (cavities and outer segment layer that increase and then decline), would not necessarily be seen as other than individual variability. 

In mice, differences in genetic strain can have profound effects on the course of retinal degeneration. ⁴⁷ Because these mice were backcrossed with the C57BL/6 strain for 4 generations rather than the 8 to 10 generations necessary to produce an essentially pure C57BL/6 background, it is possible that some of this phenotypic variability resulted from variability in genetic background. However, in our experience thus far, the phenotype and phenotypic variation in animals that have now reached 8 to 10 generations of backcross do not differ from animals in this study. Possible environmental factors contributing to the variability we observed in Rs1h-KO mice could include light, diet, or stress, which should all be similar for littermates raised in the same cage. We hypothesize that there may be a critical period at a very early age, before the retina is fully developed, when some factor could set the outcome for further disease progression throughout life. 

The range of genotype/phenotype variability between strains and the fact that some strains carry mutations that can themselves cause retinal degeneration ⁴⁸ also had bearing on the choice of C57BL/6 mice as WT controls. We compared a group of six 20-month-old WT littermates of Rs1h-KO mice with our 16-month-old C57BL/6 in this study and found similar amplitude and variation in ERG responses and comparable retinal morphology. ERG responses of the KO littermates of the WT mice were completely absent. We conclude that C57BL/6 mice are useful as WT controls for Rs1h-KO mice and that genetic background alone does not cause an abnormal retinal phenotype. 

Intravitreal injection of AAV-Rs1h at 14 days in Rs1h-KO mice produced substantial long-term rescue of structure and of ERG amplitudes and waveforms. In the best case, ERG function at 14 months, after vector gene delivery, was similar to that in 1-month-old untreated RS-KO mice. Treated eyes also had photoreceptor cell numbers at 14 months comparable to those in 1-month-old Rs1h-KO mice and RIS and ROS even better organized than in 1-month-old Rs1h-KO retinas. Thus, viral-mediated delivery and expression of RS can significantly slow the degeneration of photoreceptors over an extended time period and perhaps can even preserve the state of the retina that existed when viral gene expression was turned on. It takes several weeks for AAV-mediated ²² expression of gene product to reach substantial levels. ⁴⁹ We evaluated treated eyes at advanced ages only when cavities in untreated eyes had already disappeared. Hence, we could not determine whether cavities were reduced by AAV-Rs1h treatment. However, our earlier study of more acute rescue in adult mice ¹⁵ and that of another group, ¹⁷ showed recovery of the b-wave and preservation of the a-wave compared with fellow untreated eyes over a 2- to 3-month period after AAV-RS injection. Min et al. ¹⁷ also showed the disappearance of cyst-like structures in the retina viewed by a scanning laser ophthalmoscope 6 months after treatment. This indicates that AAV-Rs1h treatment can reverse the effects of cavity formation and preserve photoreceptor cell loss when administered to the adult or developing retina by replacing deficient retinoschisin expression. 

A possible confounding explanation for some photoreceptor rescue by AAV-Rs1h is the previously reported effects of sham intraocular injection on retinal degeneration. Photoreceptor rescue by intravitreal or subretinal injection of saline, or even the insertion of a dry needle, was reported in light-damaged rats ⁵⁰ and in RCS rats with inherited retinal degeneration, ⁵¹ possibly resulting from the release of diffusible growth and neurotrophic factors in response to injury. ⁵² However, the protective response to intravitreal injection of saline in the RCS rat was more limited in extent than the response to growth factors, which, as we have shown with AAV-Rs1h in this study, was widespread throughout the retina. ⁵¹ Mice injected with neurotrophic factors also showed photoreceptor cell rescue from inherited degeneration compared with uninjected or saline-injected fellow eyes. The injury effect was not considered a significant component of this rescue ¹⁹ because sham injection did not produce photoreceptor protection in mice (Yasumura D, et al. IOVS 1995;36:ARVO Abstract 252). Furthermore, the upregulation of neurotrophic factors in mouse retina in response to injury is lower than in rat. ⁵³ Based on these results and on the duration of the treatment effect (16 months), it is unlikely that the extensive photoreceptor rescue in AAV-Rs1h-treated eyes was a response to injury. 

The treatment effect we observed was substantial but variable. This probably is partially related to factors affecting any viral-based gene delivery system, mainly those that affect the extent and strength of gene expression at the site where the protein is used. An additional factor in the Rs1h-KO mouse is the large degree of variability in the phenotype at any given time. Thus, the variability that we saw in the effect of treatment may also reflect the considerable differences that existed in the state of the retina in these four mice at an early age, before viral expression reached therapeutic levels. The degree of inner retinal disruption at the time of injection may also affect the extent of penetration of virus particles into deeper layers of the retina after intravitreal injection. 

Treatment in the Rs1h-KO mouse indicates efficacy of treatment with a viral vector expression system when no native WT or mutant protein is expressed. However, in human disease, this is rarely the cases because most mutations are thought to produce an abnormal protein. It has been suggested the expression of mutant protein in humans could reduce the efficacy of viral-mediated expression of WT Rs1 gene because of a dominant-negative effect. ¹ The lack of a dominant-negative effect in female heterozygous carriers who secrete a mutant protein suggests there is at least no suppression of the extracellular function of WT RS protein because both would coexist outside the cell in these patients. However, it is unknown how the mutant protein would interact with normal retinoschisin protein or its expression inside the cell. Studies are under way to gain insight into this. 

Recently, treatment of XLRS patients with topical dorzolamide, a carbonic anhydrase inhibitor, has shown promise in improving visual acuity in association with reduction in schisis cavities as observed by OCT. ⁵⁴ Although this may produce short-term improvement, it is not expected to reduce the progressive photoreceptor degeneration that occurs in the Rs1h-KO mouse and is becoming more widely recognized as a component of XLRS. However, it might have a role to play in combination with gene therapy by quickly reducing cavity size, which would aid the healing process by bringing tissue together, and by setting the stage for inducing RS protein expression by gene transfer to stabilize the retina over a longer term. 

 Supported by the National Institutes of Health Intramural Research Program through the National Institute on Deafness and Other Communication Disorders and the National Eye Institute.

 Submitted for publication February 16, 2007; revised April 9, 2007; accepted May 29, 2007.

 Disclosure: S. Kjellstrom, None; R.A. Bush, None; Y. Zeng, None; Y. Takada, None; P.A. Sieving, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Natural history of structural changes in RS-KO mouse retina from 1 to 16 months. Representative sections from retinas of 1-month- (a, c, i), 4-month- (b, d, j), 6-month- (e, g, k) and16-month- (f, h, l) old mice. (a, b, e, f) Inferior half of the retina. (c, d, g, h) Higher magnification images of regions in these sections. (i–l) Taken 300 μm inferior to the optic nerve. White arrows: cells misplaced from the outer nuclear layer (ONL). One-half-micrometer sections were taken vertically through the optic nerve and stained with toluidine blue.

Figure 2.

 

Retinal morphometric changes with age in Rs1h-KO mice. (a) Semiquantitative assessment of inner retinal cavities. Size and extent of distribution of cavities was scored on a 4-point scale and then multiplied together and normalized to a 0% to 100% scale (0%, not affected). 1 month, n = 4; 4 months, n = 5; 6 months, n = 4; 8 months, n = 4; 12 months, n = 5; 16 months, n = 5. (b) ONL cell count per 100 μm retina measured from 200 to 1200 μm inferior and superior to the optic nerve (n = 4 at each time point). Dashed line: WT between 1 month and16 months (mean = 163.3 ± 9.43; n = 8). Rs1h-KO results compared to WT using Student’s t-test: *P < 0.05; **P < 0.001. (c) Width of the photoreceptor outer segment layer (OSL) measured at 100-μm intervals in the same retinal regions as cell count (1 month, n = 4; 4 months, n = 5; 6 months, n = 4; 8 months, n = 5; 12 months, n = 5; 16 months n = 5). Dashed line: WT 1 month to 16 months (mean = 24.5 mm ± 0.97; n = 15). Bar graphs = mean ± SE.

Figure 3.

 

ERG amplitude changes with age in Rs1h-KO mice. (a) a-wave V-log I. (b) b-wave V-log I. (c) a- and b-wave maximum with age. (d) (b/a-wave) ratio with age. Points represent average ± SE. 1 month, n = 7; 4 months, n = 8; 6 months, n = 7; 8 months, n = 6; 12 months, n = 7; 16 months, n = 7. Dashed line: WT (b/a-wave) ratio (mean = 1.88 ± 0.09; n = 37). (a, b) WT ERG data at 1 month (open squares).

Figure 4.

 

Comparison of morphologic and ERG changes with age in Rs1h-KO mice. (a) ONL cell number versus a-wave. (b) (b/a-Wave) ratio versus semiquantitative assessment of inner retinal cavities. Lines: linear regression (r = −0.74) and 95% confidence intervals.

Figure 5.

 

Variability of ERG waveform changes. (a) Representative waveforms at 1 month with corresponding morphology. (b) Representative ERG waveforms at 3 different ages.

Figure 6.

 

Long-term structural and functional rescue by AAV(2,2)-CMV-Rs1h in an Rs1h-KO retina. Photomicrographs of sections of retinas from the treated (right) eye (a) and the untreated eye (b) 14 months after intravitreal gene delivery at 14 days of age. Half-micrometer sections were taken vertically through the optic nerve and stained with toluidine blue. V-log I curves of ERG a-wave (c) and b-wave (d) from treated and untreated eyes compared with average responses of different age groups.

Figure 7.

 

Long-term rescue of ERG in RS-KO mice. ERG responses from treated and untreated eyes of three mice at 14 months of age. Treated eyes received intravitreal injections of AAV(2,2)-CMV-Rs1h at p14. ERG stimulus, 0.6 log cd · s/m². Immunohistochemistry staining for RS protein in Rs1h-KO mouse 1656 at the time of ERG is also shown. RS antibody (red). DAPI nuclear stain (blue).

The authors thank Yunqing Wang, Maria Santos-Muffley, and Jinbo Li for their excellent technical assistance.