Functional and behavioral restoration of vision by gene therapy in the guanylate cyclase-1 (GC1) knockout mouse

Abstract

Background: Recessive mutations in guanylate cyclase-1 (Gucy2d) are associated with severe, early onset Leber congenital amaurosis-1(LCA1). Gucy2d encodes guanylate cyclase (GC1) is expressed in photoreceptor outer segment membranes and produces cGMP in these cells. LCA1 patients present in infancy with severely impaired vision and extinguished electroretinogram (ERG) but retain some photoreceptors in both their macular and peripheral retina for years. Like LCA1 patients, loss of cone function in the GC1 knockout (GC1KO) mouse precedes cone degeneration. The purpose of this study was to test whether delivery of functional GC1 to cone cells of the postnatal GC1KO mouse could restore function to these cells.

Methodology/principal findings: Serotype 5 AAV vectors containing either a photoreceptor-specific, rhodopsin kinase (hGRK1) or ubiquitous (smCBA) promoter driving expression of wild type murine GC1 were subretinally delivered to one eye of P14 GC1KO mice. Visual function (ERG) was analyzed in treated and untreated eyes until 3 months post injection. AAV-treated, isogenic wild type and uninjected control mice were evaluated for restoration of visual behavior using optomotor testing. At 3 months post injection, all animals were sacrificed, and their treated and untreated retinas assayed for expression of GC1 and localization of cone arrestin. Cone-mediated function was restored to treated eyes of GC1KO mice (ERG amplitudes were approximately 45% of normal). Treatment effect was stable for at least 3 months. Robust improvements in cone-mediated visual behavior were also observed, with responses of treated mice being similar or identical to that of wild type mice. AAV-vectored GC1 expression was found in photoreceptors and cone cells were preserved in treated retinas.

Conclusions/significance: This is the first demonstration of gene-based restoration of both visual function/vision-elicited behavior and cone preservation in a mammalian model of GC1 deficiency. Importantly, results were obtained using a well characterized, clinically relevant AAV vector. These results lay the ground work for the development of an AAV-based gene therapy vector for the treatment of LCA1.

Figure 1. AAV-mediated GC1 expression in photoreceptors of the GC1KO mouse.

AAV5-hGRK1-mGC1 drives expression of GC1 in photoreceptor outer segments of GC1KO mice. (A). No GC1 expression is seen in the untreated, contralateral control eye (B). AAV5-smCBA-mGC1 drives expression of GC1 in photoreceptor outer segments (C) and occasionally in photoreceptor cell bodies (arrows in F). No such GC1 expression is seen in the untreated, contralateral control eye (D). Levels of therapeutic transgene expression in the AAV5-mGC1-treated eyes are only slightly less than that seen in isogenic GC1+/+ control eyes (E). All retinas were taken from mice 3 months post treatment or age matched untreated controls. Scale bars in A = 100µm, F = 25µm. OS-outer segments, IS-inner segments, ONL-outer nuclear layer.

Figure 2. AAV5-mGC1 drives expression of GC1 in both rod and cone photoreceptors.

Representative retinal section from a GC1KO eye injected with AAV5-smCBA-mGC1 stained for GC1 (red) and PNA lectin (green) reveals GC1 expression in cone outer segments (yellow overlay) as well as in rod outer segments (red alone). hGRK1-mGC1 injected eyes revealed the same pattern (data not shown).

Figure 3. AAV5-mGC1 restores retinal function to cone photoreceptors of the GC1KO mouse.

Representative cone (left column)- and rod (right column)-mediated ERG traces from GC1 +/+ (upper waveforms), untreated GC1KO (middle waveforms) and AAV5-mGC1-treated (bottom waveforms) mice. For the middle and bottom waveforms in each panel, red traces correspond to eyes injected with AAV5-smCBA-mGC1 (bottom) and their uninjected contralateral eyes (middle) and black traces correspond to eyes injected with AAV5-hGRK1-mGC1 (bottom) and uninjected contralateral eyes (middle). Cone responses in AAV5-mGC1-treated eyes are restored to approximately 45% of normal amplitude.

Figure 4. Average photopic b-wave maximum amplitudes as a function of both flash intensity and time after treatment.

Responses of GC1KO, isogenic +/+ controls, AAV5-smCBA-mGC1-treated (A) and AAV5-hGRK1-mGC1-treated (B) GC1KO mice reveal that cone responses in both smCBA-mGC1 and hGRK1-mGC1-treated mice are approximately 45% of normal for at least 3 months post injection (the longest time point evaluated in this study).

Figure 5. Optomotor analysis of visual function restoration in GC1KO mice treated with either AAV5-smCBA-mGC1 or AAV5-hGRK1-mGC1.

M1–M9 correspond to the nine mice used for testing. Photopic acuities and contrast sensitivities of GC1+/+ control mice (M1, M2), naïve GC1KO (M3, M4), smCBA-mGC1-treated (M5, M6, M7) and hGRK1-mGC1-treated GC1KO (M8, M9) mice reveal that treated mice behave like normal sighted mice (A, B). Average values for photopic visual acuity and contrast sensitivity of all GC1+/+ eyes (n = 4), untreated GC1KO eyes (n = 9) and AAV5-mGC1-treated eyes (n = 5) are shown (B) (* = P<0.0001). Cone-mediated ERG responses from each mouse (M1–M9) are shown for comparison (C).

Figure 6. Cone arrestin expression in cone photoreceptors of +/+, GC1KO, AAV5-smCBA-mGC1-treated and AAV5-hGRK1-mGC1-treated mice.

Untreated GC1KO retinas contain characteristic disorganized, detached cone outer segments (B), whereas cone outer segments were intact and cone arrestin distribution appeared normal in treated GC1KO (C,D) and +/+ (A) retinal sections. All retinas were taken from mice 3 months post treatment or age matched untreated controls. Scale bar in D = 100 µm. OS-outer segments, IS-inner segments, S-synaptic terminals.

Figure 7. AAV-mediated GC1 expression preserves cone photoreceptors in GC1KO mice.

Representative retinal whole mounts from AAV5-hGRK1-mGC1 treated (A, right), AAV5-smCBA-mGC1-treated (B-right) and contralateral, uninjected GC1KO eyes (A and B, left) stained for cone arrestin reveal that cone photoreceptors are preserved in GC1KO mice treated with AAV5-mGC1 for at least 3 months post treatment (the latest time point evaluated in this study). Cone cell densities were counted (see materials and methods) in central and inferior regions of treated and untreated GC1KO mice. Significant differences were found in both areas following treatment with either viral vector. All retinas were taken from mice 3 months post treatment.