The dose-response relationship of subretinal gene therapy with rAAV2tYF-CB-h RS1 in a mouse model of X-linked retinoschisis

Abstract

Purpose: X-linked retinoschisis (XLRS), due to loss-of-function mutations in the retinoschisin (RS1) gene, is characterized by a modest to severe decrease in visual acuity. Clinical trials for XLRS utilizing intravitreal (IVT) gene therapy showed ocular inflammation. We conducted a subretinal dose-response preclinical study using rAAV2tYF-CB-hRS1 utilizing the Rs1 knockout (Rs1-KO) mouse to investigate short- and long-term retinal rescue after subretinal gene delivery.

Methods: Rs1-KO mice were subretinally injected with 2 μL of rAAV2tYF-CB-hRS1 vector with 8E9 viral genomes (vg)/eye, 8E8 vg/eye, 8E7 vg/eye, or sham injection, and compared to untreated eyes. Reconstitution of human RS1 protein was detected using western blotting. Analysis of retinal function by electroretinography (ERG) and structural analysis by optical coherence tomography (OCT) were performed at 1, 2, 3, 5, 7, and 12 months post injection (MPI). Immunohistochemistry (IHC) was performed to evaluate cone rescue on the cellular level. Functional vision was evaluated using a visually guided swim assay (VGSA).

Results: Western blotting analysis showed human RS1 protein expression in a dose-dependent manner. Quantification of western blotting showed that the RS1 protein expression in mice treated with the 8E8 vg dose was near the wild-type (WT) expression levels. ERG demonstrated dose-dependent effects: At 1 MPI the 8E8 vg dose treated eyes had higher light-adapted (LA) ERG amplitudes in 3.0 flash and 5 Hz flicker compared to untreated (p < 0.0001) and sham-treated eyes (p < 0.0001) which persisted until the 12 MPI endpoint, consistent with improved cone function. ERG b-wave amplitudes were higher in response to dark-adapted (DA) 0.01 dim flash and 3.0 standard combined response (SCR) compared to sham-treated (p < 0.01) and untreated eyes (p < 0.001) which persisted until 3 MPI, suggesting short-term improvement of the rod photoreceptors. All injections, including sham-treated, resulted in a cyst severity score of 1 (no cavities), with significant reductions compared to untreated eyes up to 3 MPI (p < 0.05). The high and low dose groups showed inconsistent ERG improvements, despite reduced cyst severity, emphasizing the dose-dependent nature of gene augmentation's efficacy and the tenuous connection between cyst reduction and ERG improvement. IHC data showed a significant cone rescue in eyes treated with the 8E8 vg dose compared to sham-treated and untreated eyes. VGSA showed better functional vision in 8E8 vg dose treated mice. Eyes treated with the highest dose showed occasional localized degeneration in the outer nuclear layer.

Conclusion: Our data suggest that a dose of 8E8 vg/eye subretinally improves retinal function and structure in the Rs1-KO mouse. It improves cone function, rod function, and reduces cyst severity. Sham treatment resolves schisis cysts, but 8E8 vg/eye is needed for optimal retinal electrical function rescue. These findings offer a promising path for clinical translation to human trials.

Keywords: Rs1 knockout mouse; X-linked retinoschisis; dose-response; electroretinogram; functional vision; subretinal gene therapy; visually guided swim assay.

Figure 1

Expression cassette of the rAAV2tYF-CB-hRS1 vector. Components of rAAV2tYF-CB-hRS1, the vector contains AAV serotype 2 inverted terminal repeats (ITRs) and an expression cassette consisting of a cytomegalovirus enhancer (CMV) and chicken β-actin promoter (CB), human RS1 complementary DNA, and a simian vacuolating virus 40 polyadenylation (SV40 PolyA) sequence and is packaged in an AAV2 capsid containing tyrosine-to-phenylalanine mutations in 3 surface-exposed tyrosine residues (Y444F, Y500F, Y730F) in the VP3 protein capsid (rAAV2tYF).

Figure 2

Characterization of retinal phenotypes in retinoschisin-1 knockout mice: rod and cone photoreceptors functional loss and cyst formation. (A) OCT images of a Rs1-KO mouse (1) and a wild-type (WT) control littermate (2) are shown at 1 month of age. The outer nuclear layer (ONL) thicknesses are indicated by a white solid bar, and cysts are indicated by white stars on the OCT image. (B) Quantification of ONL thicknesses in Rs1-KO and WT or heterozygous control mice at 1 month of age. Rs1-KO mice have thinner ONLs. (C) Quantification of the cyst severity in Rs1-KO and WT or heterozygous control mice. Cyst severity score is modified from Bush et al. (30) where (1) no cavities; (2) <30 μm; (3) 30–49 μm; (4) 50–69 μm; (5) 70–99 μm; (6) ≥100 μm. Rs1-KO mice have cysts whereas WT or heterozygous do not. (D) (1) Waveforms in dark-adapted (DA) standard combined response (SCR) ERG using 3.0 cd·s/m² stimuli in Rs1-KO and WT or heterozygous control mice. (2) b/a wave ratio is reduced in Rs1-KO compared to WT or heterozygous controls. (E) At 1 month of age, Rs1-KO mice have greater a-wave amplitudes compared to WT or heterozygous control mice after 0.01 cd·s/m² stimuli in DA conditions. Representative waveforms are shown in (1), and the statistical comparison of their a-wave amplitudes is shown in (2). (F) Cone function on ERG measured after light adaptation using 3.0 flash (1) and 5 Hz flicker (2). Amplitudes are reduced in Rs1-KO compared to WT or heterozygous control mice. MO: month old, Rs1-KO: retinoschisin-1 knockout, WT: wild-type or heterozygous mice, μV: Microvolt, *p < 0.05, **p < 0.01, and ****p < 0.0001. Scale bar, 100 μm.

Figure 3

Expression of human RS1 protein in treated retinas of Rs1-KO mice reveals a dose-dependent response. (A) Immunoblot analysis of eyes treated with 8E9 vg, 8E8 vg, 8E7 vg doses of the AAV2tYF vector, and diluent-treated samples (Dil) retinal protein extracts. Bleb qualities were excellent for all treatment groups except for one retina treated with 8E7 dose that had a very good bleb (arrowhead). Completely untreated Rs1-KO and WT retinas served as negative and positive controls. Each lane represents an individual eye, with three eyes analyzed per dose, 2 eyes for sham treatment, 2 eyes for WT, and 1 eye for untreated KO. Eighteen μg of protein were loaded per lane. The expected molecular weights for proteins are indicated on the left, and the antibodies used to visualize the proteins are indicated on the right. (B) RS1 expression levels relative to completely WT eyes. Quantification involved normalizing the RS1 band intensity to the actin band intensity for each sample, with the expression level presented as a percentage of WT retinal samples which is set to 100%. The results illustrate a dose-dependent increase in RS1 protein expression in Rs1-KO mouse retinas treated with 8E7, 8E8, and 8E9 vg of rAAV2-tYF-CB-hRS1.

Figure 4

rAAV2tYF subretinal gene therapy improves and sustains cone photoreceptor function in Rs1-KO mice treated with the medium dose of 8E8 vg/eye over the course of 12 months. (A) Cone-dependent retinal function, assessed through light-adapted (LA) electroretinography after subjecting the eyes to a 3.0 cd·s/m² bright flash. Eyes treated with the 8E8 vg dose consistently displayed significantly greater b-wave amplitudes compared to sham-treated eyes and untreated eyes over 1 year. Eyes treated with 8E9 showed significance compared to the sham-treated eyes only at the experimental endpoint. (B) Representative LA waveforms were shown for eyes treated with different doses, as well as for sham-treated, and untreated eyes at 3 MPI and 7 MPI, respectively. (C) Another metric used to evaluate cone-dependent retinal function was the 5 Hz flicker test, which consistently showed that 8E8 vg dose-treated eyes outperformed sham-treated and untreated eyes throughout the treatment duration. Neither 8E9 nor 8E7 vg doses showed significance compared to the sham-treated eyes over the treatment course. (D) When subjected to a 5 Hz flicker, cones in treated eyes elicited higher amplitudes at 3 MPI, which was not observed in the sham-treated eyes or the untreated contralateral eyes. Over the course of 7 months, treated eyes retained significantly better amplitudes in cone ERGs. (E) Comparison of ERG flicker amplitudes between mice that received the medium 8E8 vg dose and WT or heterozygous control mice. At 1 MPI (2 MO), treatment with the 8E8 vg dose restored ERGs to near-WT amplitudes. (F) Waveforms from an eye that received the medium 8E8 dose and a WT eye. MO, months old, WT: wild-type or heterozygous mice, µV: Microvolt, *p < 0.05, **p < 0.01, ***p < 0.005, and ****p < 0.0001.

Figure 5

Subretinal gene therapy of Rs1-KO delays loss of retinal function in standard combined response ERG. (A) Combined rod-cone function was measured by 3.0 cd·s/m² bright flash under dark-adapted conditions, at 1, 2, 3, 5, 7 and 12 MPI. The 8E8 vg dose treated eyes had higher b-wave amplitudes compared to diluent-injected eyes and untreated contralateral eyes. (B,C) Representative waveforms of the 3.0 cd·s/m² bright flash SCR ERG test are shown for all vector doses, diluent-injected eyes, and untreated eyes at 3 MPI and 7 MPI, respectively. The 8E8 dose had higher amplitudes at 3 MPI and at 7MPI. Waveforms from the same mice were used at each time point. (D) An elevated b/a ratio was observed in the 8E8 vg dose-treated eyes until 2 MPI time point and in 8E9 vg dose treated eyes until 5 MPI. (E) SCR a-wave amplitudes of eyes treated with different doses of rAAV2tYF gene therapy vector, sham-treated and untreated control mice. (F) Rod and cone function comparison between the medium 8E8 vg dose and WT or heterozygous control mice at 2 and 6 MO. MO, months old, WT: wild-type or heterozygous mice, µV: Microvolt, *p < 0.05, **p < 0.01, and ****p < 0.0001.

Figure 6

Subretinal gene therapy of Rs1-KO with a dose of 8E8 vg/eye delays loss of rod photoreceptor function. (A) Rod function was measured by 0.01 cd·s/m² dim flash after DA. Notably, at 1, 2, and 3 MPI, some vector-treated eyes demonstrated a normalization of the hyper-normal a-wave seen in the Rs1-KO mouse model (as shown in Figure 2). (B) B-wave amplitudes of the 0.01 cd·s/m² dim flash after DA showing the significantly increased b-wave amplitudes of the 8E8 vg dose compared to the sham-treated and untreated eyes. (C,D) Waveform comparisons of the 0.01 cd·s/m² dim flash test were made between all vector doses, diluent-injected eyes, and untreated eyes at 3 MPI and 7 MPI, respectively. (E) Rod function comparison between the medium 8E8 vg dose and WT or heterozygous control mice (WT) at 2 and 6 MO. MO: months old, WT: wild-type or heterozygous mice, µV: Microvolt, *p < 0.05, **p ≤ 0.005, ***p ≤ 0.0005, and ****p < 0.0001.

Figure 7

OCT analysis unveils positive outcomes of rAAV2tYF gene therapy over 12 Months. (A) OCT images from 1 MPI to 12 MPI showing the cyst reduction and the ONL thickness of all the gene therapy vector treated eyes compared to sham-treated eyes and untreated eyes. Each column of images represents the same mouse, column 3 at 3 MPI timepoint is a different mouse that was treated with the same dose; each column is labeled with the gene therapy dose, the diluent or untreated status. ONL thicknesses are indicated by a white solid bar, and cysts are indicated by white stars on the OCT images. (B) Comparison of cyst severity between eyes treated with different doses, diluent treated eyes and untreated contralateral counterparts. Cyst severity was scored on a scale from 1 (least severe, no schisis) to 6 (most severe, schisis size >100 μm) from 1 MPI to 7 MPI. Diluent-treated eyes show a mild reduction in cyst severity, whereas vector-treated eyes show a further reduction in cyst severity compared to the diluent (sham)-treated eyes The results illustrate that eyes treated with the vector demonstrated a more substantial reduction in cyst severity compared to eyes treated with the diluent. (C) Comparison of ONL thicknesses between vector treated, sham-treated and untreated contralateral eyes from 1 MPI to 12 MPI. At 12 MPI, eyes treated with 8E8 vg of the vector had significantly thicker ONLs compared to the untreated eyes. *p < 0.05, **p < 0.01, ***p = 0.0001, and ****p < 0.0001. Scale bar, 100 μm.

Figure 8

Eyes treated with the 8E8 vg dose exhibit superior cone rescue at 14 months old compared to sham-treated and untreated eyes. (A–E) Visualization of cone outer segments using peanut agglutinin. Retinal sections collected from all vector treated eyes, diluent-treated eyes, and untreated Rs1-KO eyes at 14 months of age were processed and stained with DAPI (blue) and PNA (red) to visualize cone outer segments. (F) Quantification of cone outer segments per 100 μm of the retina indicates that there is a significant preservation of cones in eyes treated with the 8E8 vg dose compared to eyes that received sham treatment and untreated Rs1-KO eyes. Eyes treated with all doses of the vector, as well as diluent-treated eyes, had significantly more cones than untreated eyes. Quantification was performed by three individuals masked to treatment groups. OS, outer segment; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer. *p < 0.05, **p < 0.005. Scale bar, 50 μm.

Figure 9

Mice treated with the highest dose (8E9 vg) exhibit signs of toxicity. (A) Two out of 6 mice in the 8E9 vg treatment group had localized ONL degeneration in their treated eyes, which was not observed in mice from any other treatment groups nor in sham-treated eyes. OCT images of one mouse (R754) at 1 MPI and 5 MPI reveal ONL degeneration at 5 MPI. The ONL thicknesses are indicated by a white solid bar. Scale bar, 100 µm. (B) Analysis of ONL degeneration from 2 MPI to 7 MPI for 6 mice treated with the highest dose (8E9 vg) at four points equidistant to the optic nerve (temporal, nasal, superior, and inferior). Pink lines represent mice with localized ONL degeneration, whereas grey lines represent mice without observed abnormal ONL degeneration. (C) Immunohistochemistry (IHC) of a whole eye of the R754 mouse with 4x magnification, stained with DAPI for nucleus visualization and PNA for cone outer segment visualization. The yellow box indicates a 40× magnification of the specified area (inferior to the optic nerve) demonstrating severe loss of the ONL. Scale bar, 50 µm. Center: center scan, Superior: superior side scan, Inferior: inferior side scan.

Figure 10

Medium dose (8E8 vg) of subretinal rAAV2tYF gene therapy shows improvement in functional vision. (A) Under normal fluorescent ceiling light, Rs1-KO mice treated with 8E8 vg dosage of the rAAV2tYF subretinal gene therapy took significantly less time to locate the platform compared to untreated Rs1-KO mice but not compared to the sham-treated mice at both time points (4–6 and 9–11 MO). (B) In dark testing conditions (dim red lighting), mice treated with the 8E8 vg dose of the gene therapy took significantly less time to locate the platform compared to untreated Rs1-KO mice at younger ages but not at older ages. No statistical difference was observed for any of the vector treated eyes compared to the sham-treated eyes. (C) Confidence intervals were employed to show the data spread. No statistically significant difference or clear trend towards improvement at older age compared to the untreated eyes was observed. MO: months old, *p < 0.05.