Evaluating Efficiencies of Dual AAV Approaches for Retinal Targeting

Abstract

Retinal gene therapy has come a long way in the last few decades and the development and improvement of new gene delivery technologies has been exponential. The recent promising results from the first clinical trials for inherited retinal degeneration due to mutations in RPE65 have provided a major breakthrough in the field and have helped cement the use of recombinant adeno-associated viruses (AAV) as the major tool for retinal gene supplementation. One of the key problems of AAV however, is its limited capacity for packaging genomic information to a maximum of around 4.8 kb. Previous studies have demonstrated that homologous recombination and/or inverted terminal repeat (ITR) mediated concatemerization of two overlapping AAV vectors can partially overcome the size limitation and help deliver larger transgenes. The aim of this study was to investigate and compare the use of different AAV dual-vector approaches in the mouse retina using a systematic approach comparing efficiencies in vitro and in vivo using a unique oversized reporter construct. We show that the hybrid approach relying on vector genome concatemerization by highly recombinogenic sequences and ITRs sequence overlap offers the best levels of reconstitution both in vitro and in vivo compared to trans-splicing and overlap strategies. Our data also demonstrate that dose and vector serotype do not affect reconstitution efficiency but a discrepancy between mRNA and protein expression data suggests a bottleneck affecting translation.

Keywords: AAV; dual AAV; gene therapy; oversized AAV; retina; vector reconstitution.

Figure 1

Schematic representation of dual AAV strategies with size of each left and right construct. (A) Oversized monocistronic Rainbow construct showing order of reporter genes separated by furin-2A peptide cleavage sites. (B) Normal sized control construct (CMV.lacZ). (C) Overlap (OL) dual AAV strategy showing split left and right constructs and reconstituted genome structure. (D) Hybrid (HB) dual AAV strategy (top) and the equivalent reconstituted genome structures underneath. (E) Trans-splicing (TS) dual AAV strategy (top) and overview of the possible concatameric permutations for TS vectors with the highlighted combination (green box) that would lead to a correctly reconstituted genome. OL-C, HB-C, and TS-C denominate the structure of the control plasmid used in Figure 2A.

Figure 2

In vitro validation of dual AAV strategies. (A) Comparison of expression levels of β-galactosidade from plasmid transfection of different iterations of reconstituted Rainbow constructs (OL-C, HB-C, and TS-C) and CMV.lacZ coding control plasmid (LacZ). Two tailed Student's t-test was used to determine significance compared to CMV.lacZ (^*p < 0.01). Error bars represent S.E.M. (B) In vitro efficiency comparison of AAV2 dual hybrid (HB), trans-splice (TS) and overlap (OL) to monocistronic CMV.lacZ (LacZ) construct of β-galactosidase protein expression. (C) Dose experiment comparison of AAV2 HB dual constructs at three different doses (MOI of 1E7, 1E8, and 1E9) with each dose relative to CMV.lacZ at the equivalent dose while (D) shows expression levels of the three doses of HB relative to CMV.lacZ control at the 1E7 dose. In vitro data are represented as mean ± S.D. from 1 to 2 assays done in triplicates and values are shown as percentage of control (CMV.lacZ) after background subtraction (L construct only) and corrected for plasmid reconstitution rates (using data from graph A).

Figure 3

In vivo validation of dual AAV strategies. (A) In vivo comparison of dual strategies after subretinal injections using AAV8 or Anc80_L65 relative to AAV8 CMV.lacZ (LacZ) control at a dose of 3E9 vg/eye. (B) mRNA expression data comparing reconstituted genomes from HB at different doses (low: 3E9 vg/eye; high: 1.5E10 vg/eye). The delta-CT presented is normalized to GAPDH, relative to CMV.lacZ (LacZ). (C) Protein expression levels of HB in vivo at different doses (low: 3E9 vg/eye; high: 1.5E10 vg/eye). In vivo data are represented as mean ± S.E.M. from 2 to 3 assays done in duplicate to quadruplicate from 4 to 6 eyes per group.