Intravitreal injection of a rationally designed AAV capsid library in non-human primate identifies variants with enhanced retinal transduction and neutralizing antibody evasion

Abstract

Adeno-associated virus (AAV) continues to be the gold standard vector for therapeutic gene delivery and has proven especially useful for treating ocular disease. Intravitreal injection (IVtI) is a promising delivery route because it increases accessibility of gene therapies to larger patient populations. However, data from clinical and non-human primate (NHP) studies utilizing currently available capsids indicate that anatomical barriers to AAV and pre-existing neutralizing antibodies can restrict gene expression to levels that are "sub-therapeutic" in a substantial proportion of patients. Here, we performed a combination of directed evolution in NHPs of an AAV2-based capsid library with simultaneous mutations across six surface-exposed variable regions and rational design to identify novel capsid variants with improved retinal transduction following IVtI. Following two rounds of screening in NHP, enriched variants were characterized in intravitreally injected mice and NHPs and shown to have increased transduction relative to AAV2. Lead capsid variant, P2-V1, demonstrated an increased ability to evade neutralizing antibodies in human vitreous samples relative to AAV2 and AAV2.7m8. Taken together, this study further contributed to our understanding of the selective pressures associated with retinal transduction via the vitreous and identified promising novel AAV capsid variants for clinical consideration.

Keywords: AAV; adeno-associated virus; capsid library; directed evolution; intravitreal; neutralizing antibody; retina.

Graphical abstract

Figure 1

Library construction and experimental design Schematic of primary library construction (A). Four precursor libraries containing mutations in specific AAV2 variable regions were created as follows: CL-A (VR-I, VR-V, and VR-6), CL-B (VR-IV), CL-C (VR-VII), and CL-D (VR-VIII). Variable regions from viral libraries CL-A, CL-B, and CL-C were then combined into a second-level precursor library CL-ABC. The final library consisted of 90% of a combined CL-ABC precursor library and 10% of the CL-D precursor library. Outline of directed evolution in primate retina (B). NHP eyes with GFP-labeled PRs and Micro-ruby-labeled RGCs were injected intravitreally with the primary library 5–7 days prior to enucleation. From flow-sorted subpopulations of the round 1 eyes, RGC- and PR-specific sub-libraries were generated and used for a subsequent round of selection in NHP eyes. After round 2, variants present in the macula, peripheral retina, and RPE tissue were analyzed bioinformatically. PR, photoreceptors; RGC, retinal ganglion cells; RPE, retinal pigmented epithelium; VR-I, magenta; VR-IV, red; VR-V, black; VR-VI, blue; VR-VII, cyan; VR-VIII, lime.

Figure 2

Results of capsid library screening in NHP Distribution of top variants from NHP selection round 2 for the PR sub-library and RGC sub-library (A). Distribution of variants with a prevalence >4% from the central retina including the macula (Mac), the peripheral retina (Per), or the RPE for the PR sub-library NHPs (M2/M3) and the RGC sub-library NHPs (M4/M5). Variable region alignment of the top variants with wtAAV2 (B). S264A (yellow), A493G (red), and N551D (blue) have been previously associated with avoidance of neutralizing antibodies. Residues Y444F (green) and Y500F (white) have been previously associated with proteasomal avoidance and enhancement of transduction. The Y500F mutation was incorporated into the initial CL-B pre-library. Underlined, prevalent motif “DGEDF” is present in P2-V1, P2-V4, and P2-V9. A separate variant containing only this DGEDF motif was also generated for subsequent characterization. Changes in variable regions for top variants from wtAV2 are represented by structural modeling (C). PR, photoreceptors; RGC, retinal ganglion cells; RPE, retinal pigmented epithelium; VR-I, magenta; VR-IV, red; VR-V, black; VR-VI, blue; VR-VII, cyan; VR-VIII, lime.

Figure 3

Characterization of select capsid variants in Nrl-GFP mice following intravitreal injection mCherry fluorescence from representative eyes 8 weeks p.i. was detected by fundoscopy and IHC (A). mCherry fluorescence is shown in red and DAPI is shown in blue. Scale bar, 100 μm (white, bottom right). Transduction of rod PRs and non-rod neural retina was quantified by flow cytometry (B). All vectors except P2-V4 significantly outperformed benchmark vector AAV2(QuadYF+TV) (p < 0.005) in both rods and non-rod neural retina. Variants P2-V2, P2-V3, and DGEDF outperformed 7m8 in rods (p < 0.0001), while P2-V2 and DGEDF outperformed 7m8 in the non-rod neural retina (p < 0.0001). OS, outer segments; ONL, outer nuclear layer; INL, inner nuclear layer; GC, ganglion cells. Vectors significantly outperforming AAV2(QuadYF+TV) indicated with an asterisk (∗) and vectors significantly outperforming 7m8 indicated with hashmark (#). Statistical analysis for rods and non-rod neural retina performed by one-way ANOVA followed by a Holm-Sidak post hoc test (α = 0.05).

Figure 4

Retinal transduction is improved by rational design of some, but not all, variants following intravitreal injection mCherry fluorescence 8 weeks p.i. was measured by fundoscopy (A) for select rationally designed capsid variants. In vivo transduction was quantified by flow cytometry (B). P2-V1(trpYF+TV) and P2-V4(trpYF+TV) efficiently transduce non-rod retina (∼15%), while P2-V1(trpYF+TV) was significantly better than the other variants at transducing rod PRs (∼20%) (p < 0.0001). Representative IHC images (C) of retinas from Nrl-GFP mice intravitreally injected with top-performing variant P2-V1(trpYF+TV) containing either the ubiquitous smCBA (left) or photoreceptor-specific hGRK1 (right) promoter show robust transduction of both rod PRs and non-rod retinal cells. Off target expression in GCs was observed with the hGRK1 promoter as has been reported previously. Rationally designed variant P2-V1(trpYF+TV) has reduced heparin affinity relative to base vector P2-V1 (D). Scale bar, 100 μm (white). OS, outer segments; ONL, outer nuclear layer; INL, inner nuclear layer; GC, ganglion cell layer. Statistical analysis for rods and non-rod neural retina performed by one-way ANOVA followed by a Holm-Sidak post hoc test (α = 0.05).

Figure 5

Select variants effectively transduce NHP retina via intravitreal injection Vectors were intravitreally injected at a dose of either 1E−10 or 1E−11 vg. GFP fluorescence was measured by fundoscopy 4 weeks p.i. (A) and, shortly after, retinas were processed for immunohistochemical staining. Vector-mediated GFP expression is largely confined to the GCs adjacent to the fovea with some foveal cone transduction. Pixel integrated density of the foveal ring was quantified for each fundus image (B) with DGEDF displaying the highest pixel saturation. The area of transduction was quantified for each fundus image (C) with P2-V3 having the largest foveal ring transduction. ∗Fundus image corresponds with adjacent IHC section. OS, outer segments; ONL, outer nuclear layer; INL, inner nuclear layer; GC, ganglion cell layer.

Figure 6

Lead AAV capsid variant P2-V1 avoids neutralization by human vitreous samples containing anti-AAV2 neutralizing antibodies Color-coded table displays the dilution range at which either AAV2, P2-V1, or 7m8 vectors were neutralized by human vitreous. Neutralization is defined by a >50% reduction in mCherry transduction as determined by flow cytometry. P2-V1 is not neutralized in nine vitreous samples while 7m8 is not neutralized in two samples. NN, not neutralized.

Figure 7

Transduction of the macaque central retina by P2-V1 and rationally designed variant P2-V1(trpYF+TV) Vectors were intravitreally injected at a dose of 1E−11 vg. GFP fluorescence was imaged by fundoscopy 3 weeks p.i. and, shortly after, retinas were collected for IHC (A). P2-V1 and P2-V1(trpYF+TV) both outperformed AAV2 with regard to foveal ring transduction, with the majority of transduction observed in GC and an occasional cone photoreceptor (A, right). Pixel saturation in the foveal ring (B) and the area of transduction of the foveal ring (C) were quantified for each fundus image. P2-V1 and P2-V1(trpYF+TV) display overall higher transduction that AAV2, although there was variability observed within groups. ∗Fundus image corresponds with adjacent IHC section. OS, outer segments; ONL, outer nuclear layer; INL, inner nuclear layer; GC, ganglion cells.