Improvements induced by retinal gene therapy with voretigene neparvovec depend on visual cortical hemispheric dominance mechanisms

Abstract

Background: RPE65-associated retinal degeneration (RPE65-RD) causes severe visual deficits. Gene therapy with AAV2-hRPE65v2 is a breakthrough but it is currently unknown which visual pathways benefit from treatment and if cortical mechanisms can amplify retinal improvements.

Methods: In this within-subject design, ten patients with biallelic RPE65-RD underwent sub-retinal injection of AAV2-hRPE65v2. Psychophysical full-field stimulus threshold determination and functional magnetic resonance imaging were performed before and 12 months after treatment. Population receptive fields (pRF) were computed in V1 and visual responses assessed using contrast-reversed checkerboards (3 contrast levels).

Results: Here we show significant improvement in light sensitivity at low-luminance and neural response enhancements under low-luminance conditions specifically in the right hemisphere, which is known to show dominance in attentional and visual pooling of spatial information. Changes in pRF size also reflect known hemispheric spatial asymmetries (left/right biased for local/global analysis, respectively).

Conclusions: Our findings show a contribution of known early and high-level cortical dominance mechanisms on improvement, which constrain the effects of therapy and are therefore a target for neurorehabilitation. These findings provide insight into the limits of clinical benefits of gene therapy and suggest that neurorehabilitation approaches may be needed to enhance improvements, similarly to cochlear implants.

Fig. 1. Therapeutic intervention.

a Representative image of subretinal injection in AAV2-hRPE65v2 gene therapy (Luxturna®) in the superotemporal region. b Visualization of the injection site in the macular area (superotemporal retina) in 6 eyes from our cohort (intraoperative, surgeon's view). The therapeutic solution bubble is delimited by the dashed yellow line. The location of this bubble cannot be fully predicted. As reference points, the fovea and the optic disc are depicted. Created with BioRender.

Fig. 2. Representation of the stimulus paradigm used in functional magnetic resonance imaging (fMRI).

a Illustration of the checkerboard stimuli with three luminance levels in the white squares (high, H; medium, M; low, L). The paradigm consisted of 15 s active blocks of chessboards interspersed with a 15 s rest period (black screen) as control blocks. Each luminance level was presented three times, and the blocks of each luminance were randomly interspersed with ten rest blocks. b Illustration of simultaneous visual stimuli with orthogonal bars. The arrows in the image indicate the direction of bar movement. For the initial 144 s, the horizontal bars moved downward while the vertical bars moved to the left. Following the isoluminant period, the horizontal and vertical bars traveled in the opposite direction.

Fig. 3. Relationship between changes in V1 activation and best corrected visual acuity (BCVA).

The magnitude of the activation variation in the V1 visual cortex in response to a checkerboard pattern stimulus with a low luminance, b medium luminance, and (c) high luminance in the white squares, presented to both eyes, is correlated with the BCVA (N = 9). The neural response is an averaged combination of the left and right hemisphere responses under bilateral eye stimulation. BCVA is a combined measure of both eyes. The variation (Δ) is the percentage difference between measurements taken before and after 12 months of GT treatment.

Fig. 4. Changes in cortical responses in V1 across luminance levels.

BOLD signal in the V1 region percentage variation across the three luminance conditions of the stimulus: low-luminance (L Level); medium-luminance (M Level); and high-luminance (H Level). The variation in the BOLD signal was derived from the variation in the combined measure of the left hemisphere response during left eye stimulation and the right hemisphere response during right eye stimulation. This bar graph illustrates the mean and standard error between subjects for each luminance level, for N = 9. Additionally, individual data points for each luminance condition are plotted. In the M Level condition, two participants presented outlier values (not shown).

Fig. 5. pRF mapping before and after gene therapy.

Representative maps obtained using the pRF technique: size, polar angle (PA) and eccentricity (ECC) maps of a participant before (1^st visit) and after (2^nd visit) gene therapy for both left (LH) and right (RH) hemispheres. The maps are presented in pseudocolor code: size, 0.00 (red) to 10.00 (light blue) degrees of visual angle; ECC, 0.00 (red) to 12.00 (dark blue) degrees of visual angle. The region of interest is the primary visual cortex (V1) and is represented in purple.

Fig. 6. Representation of mean pRFs size as a function of eccentricity (ECC) in the primary visual cortex (V1).

a Left hemisphere (LH) and b the right hemisphere (RH) before (S1) and after (S2) gene therapy (GT). Note that for the left hemisphere, the pRFs show a consistent reduction virtually across all eccentricities.