Cell Transplantation to Arrest Early Changes in an Ush2a Animal Model

Abstract

Purpose.: Usher's syndrome is a combined deafness and blindness disorder caused by mutations in several genes with functions in both the retina and the ear. Here the authors studied morphologic and functional changes in an animal model, the Ush2a mouse, and explored whether transplantation of forebrain-derived progenitor cells might affect the progress of morphologic and functional deterioration.

Methods.: Ush2a mice were tested at postnatal days (P) 70 to P727 using an optomotor test, which provides a repeatable method of estimating rodent visual acuity and contrast sensitivity. A group of mice that received grafts of forebrain-derived progenitor cells at P80 was tested for up to 10 weeks after grafting. At the end of testing, animals were killed, and eyes were processed for histology.

Results.: The optomotor test showed that both acuity and contrast sensitivity deteriorated over time; contrast sensitivity showed a deficit even at P70. By contrast, photoreceptor loss was only evident later than 1 year of age, though changes in the intracellular distribution of red/green cone opsin were observed as early as P80. Mice that received transplanted cells performed significantly better than control mice and no longer demonstrated abnormal distribution of red/green opsin where the donor cells were distributed.

Conclusions.: This study showed that vision impairment was detected well before significant photoreceptor loss and was correlated with abnormal distribution of a cone pigment. Cell transplantation prevented functional deterioration for at least 10 weeks and reversed the mislocalization of cone pigment.

Usher's syndrome is the most common cause of combined deafness and blindness. Several gene mutations are responsible for this disabling disorder. The prevalence of the disease is estimated in the range of 2 to 6.2 per 100,000 people. ^1–3 More than 50% of deaf-blind adults and approximately 20% of persons with retinitis pigmentosa have Usher's syndrome. ⁴ Of the individual mutations causing these conditions, the most prevalent is Usher's syndrome type 2 (USH2); the subset, USH2A, accounts for approximately 70% of USH2 cases. ³ At present there is no effective treatment. Recently, a line of mice was generated that harbors a disruption in the Ush2a gene, resulting in a deletion of the Usherin protein (called Ush2a-null mice). ⁵ Importantly, the targeted disruption of the Ush2a gene led to progressive photoreceptor degeneration, though this was not seen until after 10 months of age: only when photoreceptor loss at 20 months of age was clearly evident were the ERG a- and b-wave amplitudes diminished. There was moderate but nonprogressive hearing impairment. As such the model mimics the visual and hearing deficits in USH2A patients. The large size of the usherin gene (171-kDa protein) limits its current applicability for gene therapy. Therefore, we have explored whether a cell-based therapy might be an effective alternative in this condition. 

In a series of studies, we have explored the use of a number of cell types, ^6–10 including human forebrain-derived progenitor cells, ^11,12 injected into the subretinal space between the photoreceptors and adjacent retinal pigment epithelium (RPE) to slow the progress of deterioration of visual functions over the long term. For most of this work, we used the Royal College Surgeons (RCS) rat, an animal with an RPE defect leading to photoreceptor loss. ^13,14  

One crucial issue that has slowed progress in using animal models for exploring potential treatments for Usher's disease is the fact that photoreceptor degeneration is very protracted or fails to occur. In the Ush2a-null mice, although the thickness of the outer nuclear layer remains unchanged beyond 1 year of age, photoreceptor loss does eventually develop. In this study, we have used behavioral measures of function, including acuity and contrast sensitivity, and have found that both show early deterioration, making them more practical indices for exploring the value of cell transplantation and indeed of other potential therapeutic approaches. Furthermore, though the gross morphology of the outer retina was unchanged at these earlier times, there were subtle changes in the intracellular distribution of a cone-specific antigen, detected with an antibody against red/green cone opsin. Even more important, we have shown that transplantation of forebrain-derived progenitor cells to the subretinal space sustained the visual functions at preoperative levels and reversed the cone pigment mislocalization. 

Materials and Methods

Animals

For the background study examining the progress of functional and morphologic changes over time, Ush2a knockout mice were tested at the following ages: postnatal day (P) 70 (n = 7), P80 (n = 8), P180 (n = 7), P360 (n = 4), P566 (n = 4), P623 (n = 5), P677 (n = 3), and P727 (n = 3) for visual acuity and contrast sensitivity using an optomotor test device. C57/BL6 wild-type (WT) mice were used as normal controls. Eyes were taken for histology after functional testing. 

Once it had been established that P70 to P80 was the age at which functional deterioration could be recognized, a second set of 12 Ush2a mice received injections of cortex-derived, human neural progenitor cells (hNPC^ctx) into the subretinal space; the contralateral eyes served as an in-animal control, either untreated (n = 6) or medium-only sham injections (n = 6). 

These studies were conducted with approval and under the supervision of the Institutional Animal Care Committee at the Oregon Health and Science University; all animals were treated in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Transplantation Procedure

hNPC^ctx, line M031 (passage 25), were cultured at the University of Wisconsin as previously described. ¹¹ The neurospheres were dissociated with cell detachment solution (Accutase; Millipore, Billerica, MA) and washed; a cell suspension containing approximately 3 × 10⁴ cells/2 μL was delivered into the subretinal space through a small scleral incision in carrying medium (balanced salt solution) with a fine glass pipette (internal diameter, 75–150 μm) attached by polyethylene tubing to a 10-μL Hamilton syringe (Reno, NV). Mice received daily intraperitoneal injections of cyclosporine A (CyA; 10 mg/kg; Novartis, Basel, Switzerland) for 2 weeks after transplantation and were subsequently maintained on oral CyA, administered in the drinking water (210 mg/L), until they were killed. 

Spatial Frequency and Contrast Sensitivity

Unoperated Ush2a mice at P70, P80, P180, P360, P566, P623, P677, and P727 and WT mice at P80, P180, and P360 were tested for spatial frequency using an optomotor test apparatus. ¹⁵ This consisted of a virtual rotating cylinder on which a vertical sine wave grating projected in virtual three-dimensional space on four computer monitors arranged in a square. Unrestrained mice were placed on a platform in the center of the square, where they tracked the grating with reflexive head movements. The “cylinder” was centered on the head of the test subject. Acuity was quantified by increasing the spatial frequency of the grating using a psychophysics staircase progression until the optokinetic reflex was no longer elicited, thereby obtaining a maximum threshold. Contrast sensitivity thresholds were achieved by a similar approach but by varying the relative light and dark of the stripes at a set spatial resolution (0.1 cyc/deg). 

The transplantation group, both operated and control eyes, was tested at 4 weeks and 10 weeks after surgery. 

Histology

Mice were given overdoses of pentobarbital sodium (Sigma-Aldrich, St. Louis, MO) and perfused with phosphate-buffered saline. The superior pole of each eye was marked with a suture to maintain orientation. The eyes were removed and immersed in 4% paraformaldehyde for 1 hour, after which they were infiltrated with 30% sucrose and embedded in optical cutting temperature (OCT) gel. Horizontal sections (10 μm) were cut on a cryostat for staining with cresyl violet and immunohistochemical markers. The following antibodies were used: red/green cone opsin (1:1000; Millipore); blue cone opsin (1:1000; Millipore); γ-transducin (1:1000; Cytosignal, Irvine, CA); cone arrestin (1:10,000; Millipore); PSD95 (1:2000; Chemicon, Temecula, CA); αPKC (1:1000; Stressgen, San Diego, CA); calbindin (1:2000; Swant, Bellinzona, Switzerland); cone arrestin (1:10,000; Millipore); bassoon (1:2000; Stressgen); GFAP (1:1000; Sigma); human-specific nuclear marker-MAB1281 (1:300; Chemicon); and human nestin (1:1000; Chemicon). Retinal sections were examined by regular and confocal microscopy. 

Results

Progress of Photoreceptor Changes

Compared with many retinal degeneration mutants, photoreceptor degeneration is very slow in Ush2a mice. At P80, the thickness of outer nuclear layer (ONL) was the same as for WT mice (Figs. 1Cvs. 1A); even at P360, it was still approximately 12 cells thick compared with 14 cells thick in WT mice (Figs. 1D vs. 1B); with age however, it gradually thinned to eight to nine cells deep at P623 and six to seven layers at P727 (Figs. 1E, 1F). Retinal lamination was comparable to that in WT mice, even at advanced ages; there were none of the irregularities in inner retinal lamination that can be seen in other mutations. ^16,17 However, outer plexiform layers (OPLs) and inner plexiform layers (IPLs) were greatly reduced in thickness compared with early time points (Figs. 1C, 1D) and WT mice (Figs. 1A, 1B). 

Figure 1.

 

Cresyl violet–stained retinal sections from WT (A, B) and Ush2a (C–F) mice. (A, B) WT retinas at P80 and P360. The ONL was approximately 14 cells thick, and there was no obvious change over the period of 1 year. (C–F) Retinal sections from Ush2a mice showing the slow photoreceptor degeneration. At P80, the ONL was comparable to that of age-matched WT mice (A). Even at P360, the ONL was hardly thinner (12 cells thick compared with 14 cells thick at P80). At more advanced ages—P623 (E) and P727 (F)—the ONL was reduced to 8 cells thick. (C, D) Both OPL and IPL (double arrows) were also considerably thinner than at early time points. Scale bar, 100 μm.

There was a clear difference in the distribution of red/green cone opsin as revealed by antibody staining. In WT mice, staining was confined to the cone outer segments (Fig. 2A); even at P360, no other part of the cone was stained (Fig. 2B). In the Ush2a retinas, however, the staining was no longer confined to the outer segments; mislocalization of red/green cone opsin was observed in cone pedicles, a few cone cell bodies, and axons as early as P80 (Fig. 2C). With time staining was seen in the cell body (Fig. 2D) and eventually the whole cone profile, including segments, cell body, axon, and pedicle (Figs. 2E and 2F at P623 and P727, respectively). 

Figure 2.

 

Confocal images showing red/green cone opsin (C–F) and rhodopsin (G–H) staining in WT and Ush2a mice. Red/green cone opsin staining was confined to the cone outer segments (A), even at P360 (B); no other part of the cone was stained. In the Ush2a mice, as early as P80 (C), the staining was not limited to the outer segments, and the cone pedicles were heavily stained. Some cone cell bodies and axons were also stained. By P360, more cell bodies were positively stained (D) and the whole cone profile was stained consistently at advanced ages (E, P623; F, P727). (G, H) Rhodopsin staining of WT and Ush2a mice at P80 (arrows) show a similar staining pattern. Scale bar, 100 μm.

No differences were seen between Ush2a and WT mice using a number of retinal markers in retinas up to at least P360. These included rhodopsin (Figs. 2G, 2H), cone-specific markers, cone arrestin, γ-transducin, and blue cone opsin (data not shown). 

The other marker used here that showed differences between Ush2a and WT was glial fibrillary acidic protein (GFAP). Retinal supporting cells-Müller glia overexpress GFAP in response to all kinds of retinal perturbations, including those associated with macular degeneration, ^18,19 retinitis pigmentosa, ^20,21 diabetic retinopathy, ^22,23 aging, ²⁴ and retinal detachment. ^25,26 GFAP is expressed mainly in astrocytes in the nerve fiber layer of RGC axons (Figs. 3A, 3B) in WT mice: no staining was seen within the retina even at P360 (Fig. 3B). However, as early as P80, GFAP staining was seen in the processes of Müller glia and within the optic nerve fiber layer of Ush2a mice (Fig. 3C). The processes of Müller glia span from the inner limiting membrane to the photoreceptor layers (Figs. 3D–F). The outer limiting membrane was not stained even at late stages of degeneration, in contrast to our previous study in the RCS rat, which showed that both inner and outer limiting membranes were strongly positive for GFAP; and Müller glia processes sprouted into the debris zone (containing an accumulation of undigested outer segment tips). 

Figure 3.

 

Confocal images showing GFAP staining (red) in WT (A, B) and Ush2a (C–F) mice. Sections were counterstained with DAPI (blue). GFAP antibody stained the nerve fiber layer in WT mice at P80 (A, upward-pointing arrows). No inner retinal staining was evident even at P360 (B, upward-pointing arrows). In Ush2a mice, even at P80 (C), stronger staining was seen in the nerve fiber layer (upward-pointing arrows) and processes of Müller cells (right-pointing arrows). Inner retinal staining was even stronger at P360 (D, right-pointing arrows). With age, cell processes were observed in the OPL and ONL (E, F, right-pointing arrows). Scale bar, 100 μm.

Course of Functional Deterioration

Unlike the course of photoreceptor degeneration, functional deterioration could be recognized much earlier in Ush2a mice. Visual acuity at P70 was 0.36 ± 0.02 cyc/deg, slightly less than, but not significantly different from, the 0.38 ± 0.00 cyc/deg recorded in WT mice (Fig. 4A). Visual acuity was hardly changed in WT up to P360 (the latest time point tested here). From P80, it gradually decreased in Ush2a animals, and by P360 it was reduced to 0.29 ± 0.02 cyc/deg; at P727, it was reduced to 0.12 ± 0.00 cyc/deg. This indicated that even at P360, visual acuity was still approximately 70% of normal WT value. The contrast sensitivity was clearly elevated in Ush2a mice, as early as P80 (Fig. 4B), giving a figure of 7.9% compared with 3.6% in WT mice. As did acuity, contrast sensitivity remained unchanged up to P360 in WT mice, but it gradually increased in the Ush2a mice with age, to 13% at P360 and 34% at P727. 

Figure 4.

 

Optomotor response was measured in WT and Ush2a mice at different time points. (A) Visual acuity was approximately 0.38 cyc/deg in WT mice. There was slow deterioration with time in Ush2a mice (from 0.36 cyc/deg at P70 to 0.29 cyc/deg at P360 and then to 0.12 cyc/deg at P727. (B) Contrast sensitivity showed clear deterioration in Ush2a mice compared with WT mice. Contrast sensitivity was below 5% in WT mice; however, in Ush2a mice, contrast sensitivity was elevated even at P80 (approximately 8%); with time it continued to increase; 13% was recorded by P360, and 34% at P727. (C, D) Both visual acuity and contrast sensitivity at 4 and 10 weeks after surgery were significantly better in Ush2a mice receiving subretinal injection of human forebrain-derived progenitor cells at P80 than in sham-injected and untreated controls (P < 0.05, t-test).

Effects of Cell Transplantation

Functional Evaluation.

To examine whether subretinal injection of forebrain-derived progenitor cells would prevent functional deterioration in the Ush2a mice, animals were tested at 4 weeks and 10 weeks after transplantation at P80. We found that animals that received grafts had higher visual acuity and lower contrast sensitivity than controls (medium injection alone and untreated; Figs. 4C, 4D). Further analysis indicated that the difference between grafted and control animals was significant (P < 0.05, t-test) in visual acuity and contrast sensitivity tests. 

Donor Cell Distribution.

Eyes were harvested at 4 and 10 weeks after surgery and were processed for histology. There was an extra layer of cells in the subretinal space only in grafted eyes on cresyl violet–stained sections (Fig. 5A). Human-specific antibody (anti-human nuclear marker) staining revealed that these were donor cells. Antibody-positive cells were confined in the subretinal space, and no donor cells were found in any other part of the retina (our previous studies in RCS rats showed that the donor cells were dispersed within retina and that they formed a layer between photoreceptors and the RPE ^11,12 ). The retina maintained an orderly lamination, and no untoward or pathologic manifestations, such as uncontrolled cell growth, tumor formation, or invasive cells, were observed on cresyl violet–stained sections. 

Figure 5.

 

Retinal sections at P150 from Ush2a mice that had received subretinal injections (at P80) of human forebrain-derived progenitor cells (A–D). Distance from the graft (E, F) and sham-injected controls (G, H). (A) Cresyl violet–stained retinal section; upward-pointing arrows indicate an extra layer of cells in the subretinal space. (B) Confocal image taken from a section close to (A) that was double stained with human-specific nuclear marker (red) and human-specific nestin (green) and was counterstained with DAPI (blue). Upward-pointing arrows: positive human-specific antibody staining. (C, D) Confocal images taken from sections close to (A). Sections were double stained with human-specific nuclear marker (red; C, D, upward-pointing arrows) and red/green cone opsin antibody (green). Section was counterstained with DAPI (C). The red/green cone opsin staining was confined within cone outer segments. (E–H) Confocal images of the same section taken from an area distant from the graft (E, F) and sham-injected control (G, H) showing the abnormal distribution of red/green cone opsin in the cell bodies (F, H, double arrows) and cone pedicles (H, arrows). Scale bars, 50 μm (B); 100 μm (A, C–F).

Further study showed that donor cells expressed human nestin, as revealed by double staining human nuclear marker with nestin antibodies (Fig. 5B). Our previous study showed that these progenitors expressed nestin for more than 8 months after they were injected into the subretinal space of RCS rats. ¹² This result suggested that human progenitor cells also remained undifferentiated when injected into Ush2a mice. 

To examine whether these progenitor cells proliferated after injection into the subretinal space, as shown in previous studies in both retina and central nervous system, ^11,27 we applied an antibody against human proliferating cell nuclear antigen to retinal sections. There was no positive staining (data not shown). 

Reverse Mislocalization of Cone Pigment.

To examine whether there was morphologic support for the improvement of visual function, in particular the distribution of red/green opsin, retinal sections were double stained with human nuclear marker (labeling donor cells) and red/green opsin. To our surprise, the mislocalization of cone pigment was no longer evident in the area with donor cells (Figs. 5C, 5D); in areas distant from donor cells, clear staining was seen in the outer segments, cell body, axons, and pedicles (Figs. 5E, 5F); in sham-injected control, abnormal distribution of red/green opsin was evident (Figs. 5G, 5H), and this matched findings in unoperated eyes of the same age. Because the abnormality was already seen in unoperated Ush2a mice at P80, at the time of cell injection, this result indicated that subretinal injection of forebrain-derived progenitor cells actually reversed the abnormal distribution of cone pigment. 

Discussion

This study demonstrated that though the outer nuclear layer is intact on gross examination until more than 12 months of age, there are subtle changes in the intracellular distribution of red/green opsin and slow deterioration in visual function from as early as P70, which becomes more pronounced with age in Ush2a mice. Introduction of forebrain-derived progenitor cells to the subretinal space around the time of inception of functional deterioration delays the decline in visual performance for at least 10 weeks after surgery; it also corrects the mislocalization of red/green cone opsin. 

With respect to slow degeneration, the Ush2a mouse has a mutation similar to that of the RPE65 mouse ²⁸ in that though gross morphologic integrity of the outer nuclear layer is sustained, functional changes are evident from an early age. This makes it an ideal animal model in which to examine experimental intervention that may be relevant to a preventive therapeutic approach, using functional indices as the primary readout of success. In contrast to the RPE65 model, however, functional deterioration is not an acute event, and, even at P180, visual performance lies well above baseline levels. This compares with observations on USH2A patients in whom deterioration of vision is a slow event and mean onset of detection of night blindness at approximately 15 years of age is slightly ahead of the onset of acuity deterioration. ²⁹ Although there is considerable variation among families, gradual progressive peripheral vision loss generally results in constriction of the visual field to 5° to 10° by the fourth decade of life. Furthermore, cone photoreceptor dysfunction is detectable later in USH2A than in many other forms of retinitis pigmentosa. ³⁰ This may provide a longer time window for preventive therapy than is the case for most other inherited photoreceptor diseases. 

In the mouse, the lack of gross morphologic changes is accompanied by subtle changes in the distribution of red/green opsin, which, instead of being sequestered only in the outer segment, is now seen throughout all compartments of these cone photoreceptors. Surprisingly, neither rods nor blue cones are affected. Previous studies have shown that when outer segments were compromised both in patients and in animal models, ^16,20 rhodopsin staining was seen in compartments of the photoreceptors other than the outer segments. Closer comparison to the present observations is some recent work showing that in a retinitis pigmentosa condition, red/green cone opsin was also found throughout the specific class of cone photoreceptors, which in humans constitutes the predominant cone type. ³¹ Other cone types appeared normal. The absence of a clear rod abnormality—together with the ERG observation of Liu et al. ⁵ showing that the a-wave, which primarily reflects rod activity in mice, diminishes only when there is clear evidence of photoreceptor degeneration and not before—appears different from what is seen in the USH2A clinical phenotype and raises the possibility that this may not be a complete model of the clinical condition. 

In addition to the changes within the red/green cones, upregulation of GFAP expression occurred in Müller glia with time, suggesting the retina was under stress. However, there was no indication of major infiltration with invasive cells, such as inflammatory cells, that would be evident in cresyl violet–stained material. 

We have shown that, in the RCS rat, forebrain-derived progenitor cells can sustain photoreceptors for prolonged periods and can dramatically slow the deterioration of visual function. This animal has a defect in the ability of the RPE to phagocytose shed outer segment material. The hypothesized effect of the cells in this mutation centers around the potential ability to provide a trophic role, sustaining rods from degeneration and, as a result, preserving cones or possibly assuming a phagocytic role and removing shed outer segment material. ¹¹ Dark/light adaptation studies suggested that although the rods were preserved, they were unable to function normally because they were unresponsive at low luminance levels. ³² The Ush2a mouse provides a very different animal model of photoreceptor dysfunction. Here the usherin gene knockout affects a molecule involved in the transport of proteins between the cell body and the outer segment. ⁵ The opsin-labeling studies performed here show that opsin was still transported to the outer segments, and this would be consistent with a continued level of visual function albeit progressively less effective than normal, but that in the case of red/green opsin, there is mislocalization of the opsin. The functional rescue effect of the forebrain cells in this very different primary defect suggests that forebrain-derived cells are effective beyond the potential role of phagocytosis and that the trophic effect is particularly important. Here, too, it is not a simple question of photoreceptor survival but, rather, of a molecular defect within the cells, such as incomplete photopigment transport, which can be reversed or overridden by the presence of the cells. The cells are clearly not likely to correct the gene defect but rather to neutralize its deleterious effect. Whether this is an additional consequence of the production of trophic factors such as IGF-1 and FGF-2, which they are known to produce in vitro, or whether there is a presently unrecognized effect has yet to be determined. At the least they sustain two important visual functions, acuity and contrast sensitivity, and reverse the mistrafficking of cone opsin. 

The ability of the cells to be effective in this situation has two consequences. One is that they may be valuable in a much broader range of mutations affecting photoreceptor function than associated with RPE dysfunction. Another is that they provide a potential therapy route for treating USH2A patients. Both clearly deserve further attention. 

In summary, we have shown that functional monitoring of an Usher syndrome animal model exposes early changes, occurring before the first indications of photoreceptor loss, making the animal amenable to evaluation of the effects of experimental interventions. Furthermore, successful use of a cell known to protect photoreceptors from degeneration in a different animal model suggests that application of this cell in potential clinical treatment of at least one cohort of Usher's patients should be considered. 

Footnotes

 Supported by Hear See Hope, Foundation Fighting Blindness, Research to Prevent Blindness, and Walsh Foundation.

Footnotes

 Disclosure: B. Lu, None; S. Wang, None; P.J. Francis, None; T. Li, None; D.M. Gamm, None; E.E. Capowski, None; R.D. Lund, None

The authors thank Yuan Zhang, Jie Duan, and Benjamin Cottam for their assistance with histology and Clive Svendsen for being the original source for the cells used in this study.