Long-term restoration of rod and cone vision by single dose rAAV-mediated gene transfer to the retina in a canine model of childhood blindness

Abstract

The short- and long-term effects of gene therapy using AAV-mediated RPE65 transfer to canine retinal pigment epithelium were investigated in dogs affected with disease caused by RPE65 deficiency. Results with AAV 2/2, 2/1, and 2/5 vector pseudotypes, human or canine RPE65 cDNA, and constitutive or tissue-specific promoters were similar. Subretinally administered vectors restored retinal function in 23 of 26 eyes, but intravitreal injections consistently did not. Photoreceptoral and postreceptoral function in both rod and cone systems improved with therapy. In dogs followed electroretinographically for 3 years, responses remained stable. Biochemical analysis of retinal retinoids indicates that mutant dogs have no detectable 11-cis-retinal, but markedly elevated retinyl esters. Subretinal AAV-RPE65 treatment resulted in detectable 11-cis-retinal expression, limited to treated areas. RPE65 protein expression was limited to retinal pigment epithelium of treated areas. Subretinal AAV-RPE65 vector is well tolerated and does not elicit high antibody levels to the vector or the protein in ocular fluids or serum. In long-term studies, wild-type cDNA is expressed only in target cells. Successful, stable restoration of rod and cone photoreceptor function in these dogs has important implications for treatment of human patients affected with Leber congenital amaurosis caused by RPE65 mutations.

FIG. 1

Short-term (A–C, 1–3 months) and long-term (D–F) restoration of rod and cone retinal function after a single subretinal treatment of AAV-RPE65. (A) Representative ERGs evoked by standard white flashes (0.4 log scot-cd s m⁻²) presented under dark-adapted (DA) and light-adapted (LA) conditions. DA traces are single flashes, LA traces are averages obtained at repetition frequencies of 1 (top) and 29 Hz (bottom). Black vertical lines show the timing of the flashes. Identities of the dogs (BR74, BR61, BR164) refer to Table 1; R, right eye, L, left eye. ERGs were performed at 3 and 1 months posttreatment in BR61 and BR164, respectively. (B) ERG photoresponses evoked by white flashes of high energy (3.7 log scot-cd s m⁻²) under DA and LA conditions, same data are shown on slow (top) and fast (bottom) time scales to allow interpretation of late and early components, respectively. Gray lines show the baseline and the 4-ms time point at which rod and cone photoreceptor responses were measured. (C) Comparison of rod and cone function in the control eyes to that of the two treatment groups. Rod function shown refers to the DA ERG photoresponse amplitude at 4 ms and cone function refers to the peak amplitude of the LA 29-Hz waveform. Each triangle represents an eye. Horizontal dashed lines represent the upper limit (mean + 3 SD) of the respective measurement in the group of control RPE65^−/− affected eyes (n = 47), which had not received treatment (No Tx). Successful recovery of rod and cone function is demonstrable in 23/26 (88%; green triangles) of the eyes receiving subretinal AAV-RPE65 but 0/11 (0%) of the eyes receiving intravitreal AAV-RPE65. Symbols with error bars show the statistics (means ± SD) for the two control groups. (D) ERGs evoked by standard white flashes in the right eye of an RPE65 mutant dog (BR33) before treatment (Pre-Tx) and over a 3-year interval after treatment. Color coding as in B. (E) ERG photoresponses evoked with white flashes of high energy over the same 3-year interval in the same eye as in D. Waveforms displayed as in A and B. (F) Two eyes with subretinal AAV-RPE65 show stable level of partial restoration of retinal rod and cone function, whereas two eyes with intravitreal AAV-RPE65 show amplitudes similar to those of untreated eyes. Horizontal dashed lines represent the upper limit (mean + 3 SD) of the respective measurement in the group of control RPE65^−/− affected eyes, which had not received treatment.

FIG. 2

Chromatographic separation of nonpolar retinoids in retina/RPE–choroid in normal and untreated and treated RPE65^−/− dogs. (A) RPE65^+/+ dog (BR140, OS-SC, see Supplementary Table 1). (B) RPE65^−/− affected dog, untreated (BR122 OD-SC, see Supplementary Table 1). (C) RPE65^−/− affected dog treated with AAV-RPE65 (BR74 OD; C = SN region). Retinoids were extracted from the indicated regions of the eye and separated on normal-phase HPLC as described under Materials and Methods. The peaks correspond to the following retinoids: 1, all-trans-retinyl esters; 2 and 2′, syn- and anti-11-cis-retinal oximes; 3 and 3′, syn-all-trans-retinal oximes; 4, 11-cis-retinol; 5, all-trans-retinol. *Artifact related to a change in the solvent composition. The syn isomer of syn-11-cis-retinal oxime is shown in the red box; the expected elution time for syn-11-cis-retinal oxime is shown by the red arrow; blue arrow indicates expected elution time of syn-9-cis-retinal oxime. 11-cis-Retinal oximes are present at high level in A, and low level in C, but not in B. There is dramatic accumulation of retinyl esters in the RPE65^−/−affected dog (C) with or (B) without treatment. (D) Regional variation in nonpolar retinoids in six retinal sectors from an eye treated in the superior central retina with rAAV-RPE65. Note the limited diffusion of 11-cis-retinal in the treated eye. Its production is restricted to the site of subretinal injection (SC). Inset: The spectrum of peak 2 (syn-11-cis-retinal oxime). SC, superior central; SN, superior nasal; ST, superior temporal; I, inferior.

FIG. 3

Retinal photomicrographs of (A, B, G, H) normal, (C, D, I, J) RPE65^−/−, and (E, F, K–V) vector-treated RPE65^−/− dogs. Vector-treated animals were (E, F) BR119, right eye, injected with AAV2/2-CBA-cRPE65 at 3.5 months and sampled at 10.5 months; (K, O–R) BR53, right eye, injected with AAV2/2-CBA-cRPE65 at 10.5 months and sampled at 17.5 months; (L, M, N) BR117, left eye, injected with AAV2/5-CBA-hRPE65 at 10 months and sampled at 2 years; and (S–V) BR53, left eye, injected with AAV2/2-CBA-cRPE65 at 10 months and sampled at 18 months. As in normal (A, B; 5.5 months), the photoreceptor and outer nuclear layers are intact in affected (C, D; 3 months) and in vector-treated animals (E, F). Lipoidal inclusions in the retinal pigment epithelium are prominent in both untreated and treated affected animals (small oblique arrows in C, E). Normal dogs show intense RPE65 immunolabeling (G, H, green), and peanut agglutinin (PNA; G) and opsin (H) label, respectively, the extracellular insoluble cone domains and the rod outer segments. The RPE65^−/− dog shows the same pattern of labeling with PNA (I) and opsin (J), but RPE65 immunolabeling is absent. In treated animals, opsin immunolabeling (red) remains unchanged, but RPE65 labeling is restored (green) and limited to the RPE layer (K, L). (M) and (N) show low- and high-power views of RPE65 expression in the RPE cell layer and in one rod photoreceptor (green). (O–R) and (S–V) show that, although the distribution of expression is more limited in the right retina (O, R) than in the left (S, V), RPE65 labeling is present in RPE cells but no other cells. Even in areas with intense RPE65 expression, lipoidal inclusions in the RPE are still present (Q, U, small oblique arrows). All calibration markers represent 50 Am, marker in F applies to all except N, R, and V.