Long-term rescue of cone photoreceptor degeneration in retinitis pigmentosa 2 (RP2)-knockout mice by gene replacement therapy

Abstract

Retinal neurodegenerative diseases are especially attractive targets for gene replacement therapy, which appears to be clinically effective for several monogenic diseases. X-linked forms of retinitis pigmentosa (XLRP) are relatively severe blinding disorders, resulting from progressive photoreceptor dysfunction primarily caused by mutations in RPGR or RP2 gene. With a goal to develop gene therapy for the XLRP-RP2 disease, we first performed detailed characterization of the Rp2-knockout (Rp2-KO) mice and observed early-onset cone dysfunction, which was followed by progressive cone degeneration, mimicking cone vision impairment in XLRP patients. The mice also exhibited distinct and significantly delayed falling phase of photopic b-wave of electroretinogram (ERG). Concurrently, we generated a self-complementary adeno-associated viral (AAV) vector carrying human RP2-coding sequence and demonstrated its ability to mediate stable RP2 protein expression in mouse photoreceptors. A long-term efficacy study was then conducted in Rp2-KO mice following AAV-RP2 vector administration. Preservation of cone function was achieved with a wide dose range over 18-month duration, as evidenced by photopic ERG and optomotor tests. The slower b-wave kinetics was also completely restored. Morphologically, the treatment preserved cone viability, corrected mis-trafficking of M-cone opsin and restored cone PDE6 expression. The therapeutic effect was achieved even in mice that received treatment at an advanced disease stage. The highest AAV-RP2 dose group demonstrated retinal toxicity, highlighting the importance of careful vector dosing in designing future human trials. The wide range of effective dose, a broad treatment window and long-lasting therapeutic effects should make the RP2 gene therapy attractive for clinical development.

Figure 1.

Progression of photoreceptor dysfunction in Rp2-KO mice. (A) ERG of Rp2-KO mice and their WT littermates over an 18-month period. The stimulus intensities for dark- and light-adapted ERGs were −4.0 to 3.0 and −1.0 to 2.0 log cd s/m², respectively. Only the amplitudes with the highest flash stimuli are shown. The ERG response with a full range of flash stimuli is shown in the Supplementary Material, Figure S2. Significantly reduced amplitudes of dark-adapted a- (left penal) and light-adapted b-wave (right panel) were observed in Rp2-KO mice, starting from 4 months of age and through the entire duration. (B) The ratio between KO and WT of dark-adapted a-wave and light-adapted b-wave over the 18-month period. The ratio for light-adapted b-wave decreased at a nearly constant rate, whereas that for dark-adapted a-wave remained stable during 4–18 months of age. (C) Representative ERG traces including 10 Hz flicker at 4 and 18 months from one KO mouse and one WT littermate. (D) Overlay of light-adapted b-waves from an Rp2-KO and a WT mouse. (E) Illustration of the parameters used to characterize the kinetics of the light-adapted b-wave. T_{50 rise} refers to the time from the stimulus onset to the 50% of the peak amplitude in the rising phase. T_max refers to the time from the stimulus onset to the peak amplitude, same as b wave implicit time. T_{50 decay} refers to the time from the peak amplitude to 50% of the peak amplitude in the falling phase. Calculations were based on filtered ERG trace shown in blue. The original ERG trace is shown in black. (F) Comparison between Rp2-KO and WT mice for light-adapted b-wave kinetics at 1 and 4 months. In addition to a significantly prolonged implicit time, the KO mice exhibit a much slower kinetics of the b-wave falling phase. In (A) and (F), the number of mice tested in each group is indicated by white lettering inside each bar. Data from WT and KO mice were compared by two-tailed unpaired t-test and represented as mean ± SEM. *P < 0.05, **P < 0.01, ****P < 0.0001.

Figure 2.

Cone degeneration in 18-month-old Rp2-KO mice. Age-matched WT C57/Bl6 mice were used as controls. (A) M- and S-opsin staining of retinal sections. Markedly reduced numbers of M- and S-opsin expressing cells in the KO retina were observed compared with those in the WT retina. (B) Rhodopsin staining of retinal sections. No obvious differences were observed between KO and WT mice in rhodopsin expression and its intracellular localization, and thickness of rod-dominant photoreceptor layer. Scale bar 20 μm. OS, outer segments; IS, inner segments; ONL, outer nuclear layer; INL, inner nuclear layer.

Figure 3.

Human RP2 AAV vector and its expression in Rp2-KO retina. (A) Schematic representation of the vector. (B) Immunoblot analysis using an RP2 antibody that recognizes both mouse and human RP2 proteins. The retinal lysate from an Rp2-KO mouse receiving subretinal administration of the AAV8-scRK-hRP2 vector for 1 month revealed a ∼40 kDa protein band corresponding to the human RP2 protein, which is similar to that detectable in human retinal lysate. The retinal lysate from a WT mouse revealed a band that migrates slightly faster than that from the human. β-Actin was used as loading controls. (C) Immunostaining of retinal sections from Rp2-KO mice 1 month after they received subretinal injection of vector or vehicle, using an antibody against both mouse and human RP2 proteins. A WT C57/Bl6 retinal section was used as positive control. Endogenous mouse RP2 protein was detected in multiple layers in WT retina, including the IS, OPL and IPL, which was not seen in the Rp2-KO retina. The vector-expressed human RP2 protein was primarily observed at the IS and nuclei of photoreceptors. Scale bar 50 μm. OS, outer segments; IS, inner segments; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer.

Figure 4.

Preservation of ERG response in Rp2-KO mice that received vector treatment. Mice received unilateral injections of the AAV8-scRK-hRP2 vector and contralateral injections of vehicle at 4–6 weeks of age. (A) Light-adapted b-wave amplitude of Rp2-KO mice injected with 1 × 10⁸ (left panel) or 3 × 10⁸ vg/eye (right panel) of the vector. The stimulus intensities were from −1.0 to 2.0 log cd s/m². Only the amplitude with the highest flash stimulus is shown. The ERG response with a full range of flash stimuli is shown in the Supplementary Material, Figure S4 and S5. Sustained higher amplitude was observed in vector-treated eyes than control eyes from 4 to 18 months of age. The number of mice tested in each group is indicated by white lettering inside each bar. The ages of the tested mice are indicated in the X-axis. ERG amplitudes from vector- and vehicle-injected eyes were compared by two-tailed paired t-test and represented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. ERG of age-matched WT mice shown in Figure 1 was used as reference. (B) Overlay of light-adapted ERG traces from vehicle- and 3 × 10⁸ vg vector-injected eyes of an Rp2-KO mouse at different time points. ERG trace from an age-matched WT littermate is shown as control. While slower-b wave kinetics was observed in the vehicle-treated eye, the vector-treated eye displays a similar kinetics to that of WT. (C) Comparison between vehicle- and vector-treated eyes of Rp2-KO for their photopic b wave kinetics at 4 months. Age-matched WT littermates were used as references. While the vehicle-injected eyes exhibit a much slower kinetics of the b wave especially at the falling phase as indicated by the T_{50 decay}, the vector-injected eyes display normal kinetics as the WT. Data from vector- and vehicle-injected eyes were compared by two-tailed paired t-test and represented as mean ± SEM. ***P < 0.001. In (A) and (C), the number of mice tested in each group is indicated by white lettering inside each bar.

Figure 5.

Preservation of optomotor response in Rp2-KO mice following vector treatment. Optomotor test was performed on 19-month-old Rp2-KO mice that received subretinal injections of AAV8-scRK-hRP2 vector at 4–6 weeks of age. Mice were injected unilaterally with a dose of either 1 × 10⁸ or 3 × 10⁸ vg/eye, with vehicle injected to the contralateral eyes as controls. Age-matched WT littermates were used as references. The number of mice used in each group is indicated by white lettering inside each bar. Spatial resolution (expressed as cycles/degree) from vector- and vehicle- treated eyes were compared by paired two-tailed t-test. All the values in the bar graph were represented as mean ± SEM. *P < 0.05, **P < 0.01.

Figure 6.

Correction of cone opsin mis-localization and restoration of cone PDE6 expression in Rp2-KO retina following vector treatment. Immunostaining of M-opsin (A), S-opsin (B) and cone PDE6 (C) of retinal sections from a 6.5-month-old Rp2-KO mouse that received vehicle or 5 × 10⁷ vg AAV8-scRK-hRP2 vector administration at 4 weeks of age. Retinal sections from an age-matched WT C57/Bl6 mouse were used as controls. While M-opsin is mis-trafficked to OS, IS, nuclei and synaptic terminals in vehicle-treated retina, it is only localized to OS in vector-treated retina, similar to the WT mouse. Diminished cone PDE6 was observed in the vehicle-treated retina, whereas in the vector-treated retina, cone PDE6 restored and is localized to OS. Scale bar 50 μm. OS, outer segments; IS, inner segments; ONL, outer nuclear layer; INL, inner nuclear layer.

Figure 7.

Long-term preservation of cone photoreceptors in Rp2-KO mice following vector treatment. (A) Immunofluorescence analysis of retinal whole mounts of an 18-month-old Rp2-KO mouse that received subretinal administration of vehicle or 1 × 10⁸ vg AAV8-scRK-hRP2 vector at 4 weeks of age. Quantification of cone cells in retinal whole mounts (B) revealed a significantly higher number of cone cells in vector-treated eyes compared with vehicle-treated control eyes. The number of mice tested in each group is indicated by white lettering inside each bar. Data from vector- and vehicle-injected eyes were compared by two-tailed paired t-test and represented as mean ± SEM. **P < 0.01. (C) Retinal sections showing a higher number of S- and M-opsin expressing cells in vector-treated eye than in vehicle-treated eye. Scale bar 50 μm. OS, outer segments; IS, inner segments; ONL, outer nuclear layer; INL, inner nuclear layer.

Figure 8.

Retinal toxicity at high vector dose. (A) Dark- and light-adapted ERG responses at different time points in Rp2-KO mice treated with AAV8-scRK-hRP2 vector. Mice were injected unilaterally with 1 × 10⁹ vg/eye of the vector when they were 4–6 weeks old, with the vehicle injected to the contralateral eyes as controls. The number of mice tested in each group is indicated by white lettering inside each bar. The stimulus intensities for dark- and light-adapted ERGs were −4.0 to 3.0 and −1.0 to 2.0 log cd s/m², respectively. Only the amplitudes with the highest flash stimuli are shown. The ERG response with a full range of flash stimuli is shown in the Supplementary Material, Figure S11. ERG amplitudes from vector- and vehicle- injected eyes were compared by two-tailed paired t-test and represented as mean ± SEM. *P < 0.05, **P < 0.01, ****P < 0.0001. Lower dark-adapted ERG response in the vector-treated eyes indicates retinal toxicity likely caused by high RP2 expression. (B) RP2 staining of retinal sections from 18-month-old Rp2-KO mice that received injections of 1 × 10⁹ or 1 × 10⁸ vg/eye of the vector. Thinning of ONL was observed at multiple regions in the retina treated with 1 × 10⁹ vg/eye, while relatively normal ONL thickness was maintained in the 1 × 10⁸ vg-treated retina. The magnified images of the marked areas are shown. Scale bar 50 μm. IS, inner segments; ONL, outer nuclear layer; INL, inner nuclear layer.

Figure 9.

Preservation of cone function and cone proteins in Rp2-KO mice following vector treatment at an older age. (A) Light-adapted ERG b-wave amplitude of 18-month-old Rp2-KO mice that received subretinal injection of AAV8-scRK-hRP2 vector at 10 months of age. Mice were injected unilaterally with the vector at a dose of 3 × 10⁸ vg/eye, with the vehicle injected to the contralateral eyes as controls. The stimulus intensities were from −1.0 to 2.0 log cd s/m². Only the amplitude with the highest flash stimulus is shown. The ERG response with a full range of flash stimuli is shown in the Supplementary Material, Figure S10. The number of mice tested in each group is indicated by white lettering inside each bar. ERG amplitudes from vector- and vehicle-injected eyes were compared by two-tailed paired t-test and represented as mean ± SEM. **P < 0.01. The vector-injected eyes exhibited significantly higher amplitude than control eyes. (B) Representative light-adapted and 10 Hz flicker ERG traces. (C) Immunofluorescence of retinal sections. Higher numbers of S- and M-opsin expressing cells were maintained in the vector-treated retinal sections. Cone PDE6 (cPDE6) expression was also restored in the OS of the vector-treated retina. Scale bar 50 μm. OS, outer segments; IS, inner segments; ONL, outer nuclear layer; INL, inner nuclear layer.