Improved retinal function in a mouse model of dominant retinitis pigmentosa following AAV-delivered gene therapy

Abstract

Mutational heterogeneity represents one of the greatest barriers impeding the progress toward the clinic of gene therapies for many dominantly inherited disorders. A general strategy of gene suppression in conjunction with replacement has been proposed to overcome this mutational heterogeneity. In the current study, various aspects of this strategy are explored for a dominant form of the retinal degeneration, retinitis pigmentosa (RP), caused by mutations in the rhodopsin gene (RHO-adRP). While > 200 mutations have been identified in rhodopsin (RHO), in principle, suppression and replacement may be employed to provide a single mutation-independent therapeutic for this form of the disorder. In the study we demonstrate in a transgenic mouse simulating human RHO-adRP that RNA interference-based suppression, together with gene replacement utilizing the endogenous mouse gene as the replacement, provides significant benefit as evaluated by electroretinography (ERG). Moreover, this is mirrored histologically by preservation of photoreceptors. AAV-based vectors were utilized for in vivo delivery of the therapy to the target cell type, the photoreceptors. The results demonstrate that RNAi-based mutation-independent suppression and replacement can provide benefit for RHO-adRP and promote the therapeutic approach as potentially beneficial for other autosomal dominantly inherited disorders.

Figure 1

Schematic representation of the AAVshQ1 suppression construct. shQ1 RNA was expressed from the H1 promoter (H1shQ1). The H1shQ1 cassette was flanked up- and down-stream by spacer DNA fragments. EGFP was coexpressed from the CMV immediate-early promoter (CMVP). The SV40 polyadenylation signal (PolyA) was located at the 3′ end of the EGFP gene. Restriction enzyme sites used for cloning are given. Numbers indicate molecular sizes in base pairs, and arrows indicate the direction of transcription. L-ITR and R-ITR: left and right inverted repeats of AAV.

Figure 2

Suppression of human and mouse rhodopsin in vivo. Fellow eyes of NHR (RHO^+/−Rho^−/−) and wild-type (Rho^+/+) mice were subretinally injected with 2 µl of 2 × 10¹² vp/ml AAVshQ1, which enables coexpression of shQ1 and EGFP in transduced retinal cells. (a) Expression of the 21-nucleotide (nt) shQ1 was confirmed by RNase protection assay in adult mice 10 days postinjection (n = 2). Protected RNA was separated on 15% denaturing polyacrylamide gels and detected using an shQ1 RNA probe, labeled with P³²-γATP (lane Q1). In lane M, size markers indicate 10, 20, and 30 nt. Four weeks after AAVshNT (NT) and AAVshQ1 (Q1) administration at postnatal day 10, retinas were dissociated with trypsin and retinal cells sorted and analyzed by FACS (n = 4). (b) Representative plots of forward versus side scatter and histograms of EGFP fluorescence of the gated population (red dots on scatter plots) of NHR retinas are given for both AAVshNT (NT) and AAVshQ1 (Q1). (c) The bar chart indicates RHO mRNA expression from NHR mice in AAV-transduced (EGFP-positive) cells expressing shNT (NT) and shQ1 (Q1), isolated by FACS and quantified by qRT-PCR. (d) The bar chart indicates Rho mRNA expression from Rho^+/+ mice in AAV-transduced (EGFP-positive) cells expressing shNT (NT) and shQ1 (Q1), isolated by FACS and quantified by qRT-PCR. Error bars represent SD values. ^*EGFP-positive fraction of cells; ^***P < 0.001.

Figure 3

Retinal histology and ERG of NHR mice following suppression of RHO. Fellow eyes from NHR (RHO^+/−Rho^−/−) mice were injected with 2 µl 2 × 10¹² of vp/ml AAVshQ1 and AAVshNT at postnatal day 10 and analyzed 4 weeks postinjection. For histology, eyes (n = 3) were fixed in 4% paraformaldehyde and cryosectioned (12 µm). Cy3 label RHO immunocytochemistry was carried out and nuclei counterstained with DAPI. Representative microscopic images are provided from eyes injected with AAVshNT (a, d, and g), AAVshQ1 (b, e, and h), and from uninjected eyes (c, f, and i). Red, green, and blue fluorescence corresponds to signals of RHO, EGFP, and cell nuclei, respectively. For ERG analysis (n = 6), mice were dark adapted over night and ERG responses recorded from both eyes. (j) Means of relative amplitudes of ERGs are given for rod-isolated (rod), mixed rod and cone (mixed), and cone-isolated (cone) responses corresponding to AAVshNT (NT) and AAVshQ1 (Q1) injections. OS, photoreceptor outer segments; IS, photoreceptor inner segments; ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer; Bar = 25 µm. Error bars correspond to SD values; ^*P < 0.05.

Figure 4

Viral spread in Pro347Ser mice. Fellow eyes from Pro347Ser (RHO347^+/−Rho^+/+) mice were injected with 1 µl of 2 × 10¹² vp/ml AAVshQ1 and AAVshNT at postnatal day 10. Five weeks postinjection, eyes were fixed in 4% paraformaldehyde and retinal whole mounts prepared. Representative microscope images are depicted from eyes injected with AAVshNT (shNT; a) and AAVshQ1 (shQ1; b). Green label corresponds to EGFP fluorescence signal in the AAV-targeted region of the retina. Red dots border the extent of the retinas. Bar = 1 mm.

Figure 5

Retinal histology of Pro347Ser mice following subretinal administration of AAVshQ1 and AAVshNT. Fellow eyes from Pro347Ser (RHO347^+/−Rho^+/+) mice were injected with 1 µl of 2 × 10¹² vp/ml AAVshQ1 and AAVshNT at postnatal day 10. Five and 10 weeks postinjection, eyes were fixed in 4% paraformaldehyde, cryosectioned (12 µm) and nuclei counterstained with DAPI. Representative microscope images are shown from eyes injected with (a,d) AAVshNT, (b,e) AAVshQ1, and (c,f) from uninjected eyes. Green label corresponds to EGFP fluorescence signal in the AAV-targeted region of the retina. Three measurements of ONL thickness were taken in each of two or three microscope images, 200 µm apart in the AAV-targeted area of each eye (n = 3–5). (g) Bars represent relative ONL thickness (normalized to 5-week AAVshNT-control eyes) at 5 or 10 weeks postinjection. PS, photoreceptor segments; INL, inner nuclear layer; GCL, ganglion cell layer. Bar = 25 µm. Error bars correspond to SD values; ^*P < 0.05; ^***P < 0.001.

Figure 6

Selective suppression of RHO versus Rho expression in Pro347Ser retina. Fellow eyes of Pro347Ser (RHO347^+/−Rho^+/+) mice were injected with 1 µl of 2 × 10¹² vp/ml AAVshQ1 and AAVshNT at postnatal day 10. Five weeks postinjection, eyes were fixed in 4% paraformaldehyde and cryosectioned (12 µm). Immunocytochemistry using human RHO-specific or RHO- and Rho-specific rhodopsin antibodies was carried out (using Cy3-conjugated secondary antibody) and nuclei counterstained with DAPI. Representative microscopic images are presented from eyes injected with (a, b, c, and d) AAVshNT, and (e, f, g, and h) AAVshQ1, or (i, j, k, and l) uninjected eyes. Green label corresponds to EGFP fluorescence signal in the AAV-targeted region of the retinas. The top row depicts rhodopsin signals (RHO and Rho) and the bottom row illustrates overlaid signals of rhodopsin, EGFP, and DAPI. Sections used for immunocytochemistries were from the same eye and were no >50 µm apart. PS, photoreceptor segments; ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. Bar = 25 µm.

Figure 7

ERG of Pro347Ser mice following subretinal administration of AAVshQ1 and AAVshNT. Fellow eyes from Pro347Ser (RHO347^+/−Rho^+/+) mice were injected with 1 µl of 2 × 10¹² vp/ml AAVshQ1 and AAVshNT at postnatal day 10 (n = 15). Ten weeks postinjection, mice were dark adapted overnight and (a,b) rod-isolated, (c,d) mixed rod and cone, and (e,f) cone-isolated ERG responses were recorded from both eyes. a, c, and e: ERGs, from eyes injected with AAVshQ1 (Q1: green) and AAVshNT (NT: blue). b, d, and f: mean ERG amplitudes. Error bars represent SD values; ^**P < 0.01. Note that amplitudes on a, c, and e, and Y axes on b, d, and e are set to different scales.