Long-term rescue of retinal structure and function by rhodopsin RNA replacement with a single adeno-associated viral vector in P23H RHO transgenic mice

Abstract

Many mutations in the human rhodopsin gene (RHO) cause autosomal dominant retinitis pigmentosa (ADRP). Our previous studies with a P23H (proline-23 substituted by histidine) RHO transgenic mouse model of ADRP demonstrated significant improvement of retinal function and preservation of retinal structure after transfer of wild-type rhodopsin by AAV. In this study we demonstrate long-term rescue of retinal structure and function by a single virus expressing both RHO replacement cDNA and small interfering RNA (siRNA) to digest mouse Rho and human P23H RHO mRNA. This combination should prevent overexpression of rhodopsin, which can be deleterious to photoreceptors. On the basis of the electroretinogram (ERG) response, degeneration of retinal function was arrested at 2 months postinjection, and the response was maintained at this level until termination at 9 months. Preservation of the ERG response in P23H RHO mice reflected survival of photoreceptors: both the outer nuclear layer (ONL) and outer segments of photoreceptor cells maintained the same thickness as in nontransgenic mice, whereas the control injected P23H eyes exhibited severe thinning of the ONL and outer segments. These findings suggest that delivery of both a modified cDNA and an siRNA by a single adeno-associated viral vector provided long-term rescue of ADRP in this model. Because the siRNA targets human as well as mouse rhodopsin mRNAs, the combination vector may be useful for the treatment of human disease.

FIG. 1.

A single AAV vector expresses both RHO301 and siRNA301. (A) Map of AAV5-RHO301-siRNA301 (AAV-RS301). The expression of RHO301 was controlled by the proximal promoter region of the mouse opsin gene (MOPS500), and the small hairpin RNA containing the siRNA was controlled by the human H1 promoter. The entire coding region was contained between AAV2 terminal repeats (TR) and packaged in AAV5 capsids. (B) Various target sequences of human rhodopsin, mouse rhodopsin, and a resistant version of mouse rhodopsin (RHO301). (C) Subretinal injection of AAV2/5-MOPS-GFP leads to widespread transduction of the retina. One month after injection of our control virus, AAV-MOPS-GFP expression of GFP could be detected in a large area of the retina in a living mouse, using a Micron III fundus microscope equipped with a fluorescein filter. Color images available online at www.liebertonline.com/hum

FIG. 2.

Subretinal injection of AAV-RS301 leads to an increase in total rhodopsin RNA. (A) Reverse transcription (RT)-PCR products of mouse and human rhodopsin mRNAs were prepared with a common set of primers in the same reaction (Mao et al., 2011). Subsequent incubation with restriction enzyme MseI digested RHO301 into bands of 283 and 70 bp, and undigested human P23H RHO and mouse Rho had the same 353-bp band. RT-PCR products from uninjected eyes contained only the uncut 353-bp product after MseI treatment. Amplification products of β-actin mRNA from the same retinas (without MseI digestion) were used for normalization. (B) Endogenous rhodopsin mRNA levels (human P23H RHO and mouse Rho) were reduced in RS301-treated eyes (open column) (p<0.05; n=3). (C) Total rhodopsin mRNA (human P23H RHO, mouse Rho, and RHO301) increased in RS301-injected eyes (open column) (p<0.05; n=3). In (B) and (C) the ratio of rhodopsin to actin PCR products in untreated eyes was set to 100%.

FIG. 3.

Increased rhodopsin protein in P23H transgenic eyes injected with AAV-RS301. Four P23H transgenic mice injected with AAV-RS301 in right (R) eyes and uninjected in their left (L) eyes were analyzed by immunoblot. (A) Image of anti-rhodopsin staining using C-terminal specific antibody 1D4. Protein levels were normalized to β-actin protein from the same eye. (B) Quantification of image (A), demonstrating increased RHO protein content in injected right eyes (open columns) compared with uninjected left eyes (solid columns) (n=4; *p<0.05).

FIG. 4.

RHO301–siRNA301 gene transfer rescued retinal function in P23H transgenic mice. (A) a-wave amplitudes were measured 1, 2, 3, 6, and 9 months postinjection. The data presented here are from high-intensity flashes (0 dB) and represent both rod and cone responses. AAV-RS301-injected eyes exhibited a significantly increased a-wave response at all time points compared with control injected eyes (AAV-GFP) (n=9, **p<0.005 at 1, 3, 6, and 9 months; p<0.05 at 2 months). (B) b-wave amplitudes are presented from 1, 2, 3, 6, and 9 months postinjection. Right AAV-RS301-injected eyes showed a significantly increased b-wave response compared with left AAV5-GFP-treated eyes (n=9 at 1, 2, 6, 9 months; **p<0.005 at 3 months, *p<0.05). To determine the extent of rescue of ERG amplitudes in P23H eyes relative to untreated wild-type eyes, we compared a-wave and b-wave amplitudes in P23H transgenic mice with those of nontransgenic mice at 1 month postinjection (C) and 9 months postinjection (D). Left eyes of both strains were untreated and right eyes were injected with AAV-RS301. ERG amplitudes of the untreated nontransgenic eyes were set at 100% for both time points. At 1 month, both the uninjected transgenic and injected nontransgenic eyes showed a statistically significant (*p<0.05) decrease in a-wave amplitude. However, by 9 months only the untreated P23H transgenic eyes showed a significantly diminished response in both a- and b-wave amplitudes,<20% that of the untreated, nontransgenic eyes (n=7; **p<0.005). P23H eyes injected with AAV-RS301 maintained 80% of the a-wave amplitude of untreated nontransgenic eyes (p>0.05). For the b wave, P23H eyes injected with AAV-RS301 maintained almost 95% of the amplitude of the uninjected, nontransgenic eyes.

FIG. 5.

Injection of AAV-RS301 preserved retinal structure and morphology in P23H transgenic mice. (A) Spectral domain optical coherence tomography (SD-OCT) images were captured in live mice at the 3-, 6-, and 9-month postinjection time points. Scale bars: 30 μm. (B) The thickness of the outer nuclear layer (ONL) was measured from SD-OCT images at four standard locations relative to the optic nerve head at 3, 6, and 9 months postinjection. The thickness of the ONL from right AAV-RS301-injected eyes was significantly greater than that from left AAV-GFP-injected eyes at corresponding time points (n=9; **p<0.005 at 3, 6, and 9 months). (C) Fundus images were captured at 1 and 9 months postinjection. Top left: A P23H transgenic mouse 1 month postinjection of AAV-GFP. Pigmentary changes were observed in the area near the optic nerve head (at 7 o'clock in this image). Top right: Normal appearance of the fundus in the AAV-RS301 contralateral eye. Bottom left: Significant bleaching and pigmentary irregularities in the retina of a control P23H eye treated with the control virus. Many white spots were present in the retina, with one hyperpigmented region near the optic nerve head. The optic nerve head was also pale in the control treated eyes. Bottom right: An image of an AAV-RS301-treated right eye. The fundus picture appears similar to the 1-month postinjection picture. IPL, inner plexiform layer; IS/OS, inner segment/outer segment interface.

FIG. 6.

Histological preservation of retinas after 9 months of treatment with AAV-RS301. (A) Semithin sections were imaged 9 months postinjection. Left: Image of a retina from a P23H transgenic eye treated with AAV-GFP. In accordance with the in vivo image from SD-OCT, the image reveals a thin ONL (five or six rows of nuclei, with numerous gaps). This image also reveals shortened rod outer segments (OS). Scale bar: 30 μm. Right: In eyes treated with AAV-RS301, retinal structure was well preserved with a thicker whole retina, ONL, and OS. (B) The contour lengths of outer segments analyzed at 10 positions across the retina revealed longer outer segments at each position analyzed (S, superior hemisphere; I, inferior hemisphere). In five superior locations near the optic nerve head and five inferior locations, the differences between control treated and RS301-treated eyes were significantly different (n=4; *p<0.05, **p<0.005). (C) OS length was analyzed in whole superior area, whole inferior area, and whole retina including both superior and inferior areas to determine the overall length of OSs (n=4; **p<0.005). ONH, optic nerve head.