Figure 1.
Recombinant AAV-mediated GFP expression in the canine eyes by fundus autofluorescence imaging.
A representative summary of findings from rAAV2/2-hVMD2-GFP injected canine eyes. A normalized autofluorescence mode was used to obtain a baseline photograph for the native eye before vector administration (A). GFP fluorescence was first detectable at 2 weeks p.i. (B, arrowheads) and the outline of the injection area corresponded to the subretinal bleb documented immediately after vector administration (A, inset). An evident increase in GFP expression was observed at 4 to 6 weeks p.i. (C, D), when the peak of rAAV vector expression is expected. The boundaries of the autofluorescent areas (arrowheads), as well as retinotomy scar (arrows) remained unchanged throughout the observation period. High-resolution images were captured in AF mode at 55°; p.i.: post injection.
Figure 2.
Specificity and stability of rAAV-mediated RPE transduction regulated by hVMD2 in the canine retina.
The rAAV2/2 vector construct carrying GFP reporter gene under control of human VMD2 promoter specifically and exclusively target transgene expression to the RPE cells. Native (green) or anti-GFP probed (red) GFP expression was analyzed on frozen retinal cross-sections 2- (9.11×1010 vg), 4- (1.21×1011 vg), 6- (9.11×1010 vg) weeks, and 6 months (9.11×1010 vg) post injection. Immunohistochemical staining confirmed the gradual increase of the transgene expression level over the first 6 weeks p.i. that remained stable up to 6 months after vector administration. DAPI stain was used to detect cell nuclei; vg: vector genomes injected; p.i.: post injection; scale bar: 40 µm.
Figure 3.
Comparison of rAAV2/1- and rAAV2/2-mediated GFP expression in the wild-type canine retina.
Immunohistochemical assessment of rAAV2/1-hVMD2-GFP (2.63×1011 vg) (A) and rAAV2/2-hVMD2-GFP (9.11×1010 vg) (B) injected retinas 6 weeks post injection. GFP expression (native expression = green; anti-GFP antibody = red) is shown only in the first row of images; selected retinal and RPE proteins were evaluated by antibody labeling. RPE cells expressed Best1 and RPE65 proteins; the structure of cone photoreceptors was demonstrated by hCAR and L/M & S opsin labeling, while rods were assessed based on Rho localization. In all cases, protein expression was normal, specific and comparable to the non-injected eyes (data not shown), irrespective of the recombinant vector serotype used. Preservation of the retinal structure is demonstrated by H&E. RPE: retinal pigment epithelium, OS: photoreceptor outer segments; IS: photoreceptor inner segments; ONL: outer nuclear layer; INL: inner nuclear layer; cell nuclei were stained with DAPI; vg: vector genomes injected; scale bar: 40 µm.
Figure 4.
Evaluation of rAAV2/1- and rAAV2/2-mediated BEST1 transgene expression in the canine retina.
(A) Comparison of two normal canine eyes that received subretinal injection of rAAV2/1 (1.94×1011 vg) and a spike-in of corresponding vector expressing GFP (3.81×109 vg) or rAAV2/2 (3.92×1011 vg) expressing canine BEST1 under control of human VMD2 promoter. The outlines of the injected areas detectable in NIR mode and more evident in AF mode for rAAV2/1 (arrowheads) corresponded to the bleb formed immediately after injection (insets). The arrows indicate retinotomy sites. (B) Retinal thickness profiling done by manual segmentation across the bleb boundaries revealed no significant changes 4 weeks p.i. with either vector construct. High-resolution OCT images were obtained using a 30° lens; NIR and AF images were captured using a 55° lens; vg: vector genomes injected; p.i.: post injection.
Figure 5.
Bestrophin1 overexpression induced by rAAV2/2 in the wild-type canine retina.
Confocal photomicrographs illustrating Best1 expression (red) in the wild-type canine RPE six months p.i. The endogenous expression of Best1 (boxed area left and corresponding magnification) was limited to the basolateral plasma membrane while the transgene protein was also observed in the cell cytoplasm as a result of overexpression mediated by rAAV2/2-hVMD2-cBEST1 (3.92×1011 vg) (boxed area right and corresponding magnification). Cell nuclei were stained with DAPI; vg: vector genomes injected; p.i.: post injection; scale bar: 40 µm.
Figure 6.
Consequences of rAAV2/1- and rAAV2/2-induced BEST1 transgene expression in vivo.
Histological and immunohistochemical evaluation of wild-type canine retinae injected with rAAV2/1-hVMD2-cBEST1 (2.63×1011 vg) and a spike-in of corresponding vector expressing GFP (2.5×109 vg) or rAAV2/2-hVMD2-cBEST1 (4.44×1011 vg) in comparison to the non-injected control. H&E staining did not reveal any histological changes with either vector serotype. Both vectors induced bestrophin1 overexpression in the RPE cells 4 weeks post injection (Best1, red). While no abnormalities were observed in rAAV2/2-transduced retina, the rAAV2/1 serotype caused fluorescence in individual photoreceptor cells (green), occasional mislocalization of cone and rod opsins (arrowheads) and patchy loss of cone photoreceptors (arrows) in the rAAV2/1-hVMD2-cBEST1-injected area. RPE: retinal pigment epithelium, OS: photoreceptor outer segments; IS: photoreceptor inner segments; ONL: outer nuclear layer; INL: inner nuclear layer. Cell nuclei were stained with DAPI; vg: vector genomes injected; scale bar: 40 µm and applies to all panels.
Figure 7.
rAAV-mediated BEST1 transfer to the retinal pigment epithelium and cone toxicity associated with rAAV2/1 vector serotype. (A-B)
Co-expression of endogenous canine and human Best1 transgene in vivo. Histological and immunohistochemical analysis of wild-type canine retinae injected with rAAV2/1-hVMD2-hBEST1 (1.16×1012 vg) and a spike-in of corresponding vector expressing GFP (1.04×109) (A) and rAAV2/2-hVMD2-hBEST1 (8.82×1011 vg) (B) at 4 weeks post injection. No structural abnormalities were seen by H&E staining in any of the samples. In the retina transduced with rAAV2/1 capsid serotype (A), individual photoreceptor cells emitted green autofluorescence (arrowheads) as shown on the photomicrographs probed with anti-Best1, anti-L/M & S opsin and anti-hCAR (red) (A). Cone-specific labeling revealed loss of cone photoreceptors (arrows) only in the areas transduced with rAAV2/1 vector serotype (A). Co-expression of endogenous canine bestrophin1 and human BEST1 transgene was well tolerated when injected with rAAV2/2 serotype and no abnormalities were noted (B). (C–D) Comparison of rAAV2/1- and rAAV2/2-mediated cBEST1 transgene expression in the cmr1 (C73T/R25X) carrier retina. Histological and immunohistochemical evaluation of cmr1+/− retinae injected with rAAV2/1-hVMD2-cBEST1 (1.92×1011 vg) and a spike-in of corresponding vector expressing GFP (1.74×109) (C) and rAAV2/2-hVMD2-cBEST1 (4.44×1011 vg) (D) at 4 weeks post injection. No morphological abnormalities were detected by H&E staining between the two capsid serotypes (C–D). Sporadic autofluorescent cells (green) were observed in IS and ONL layers (C, arrowheads), and cone-specific immunolabeling (L/M & S opsin and anti-hCAR in red) revealed focal loss of cone photoreceptors (arrows) within the rAAV2/1-hVMD2-cBEST1-injected area (C). rAAV2/2-mediated cBEST1 transfer to the cmr1+/− retina results in bestrophin1 overexpression in the RPE (D, Best1 in red) with no adverse effects in the retinal tissue (D). RPE: retinal pigment epithelium, OS: photoreceptor outer segments; IS: photoreceptor inner segments; ONL: outer nuclear layer; INL: inner nuclear layer. Cell nuclei were stained with DAPI; vg: vector genomes injected; scale bar: 40 µm and applies to all panels.